Network Working Group                        Lars-Erik Jonsson, Ericsson
INTERNET-DRAFT                        Mikael Degermark, Lulea University
Expires: November December 2000                              Hans Hannu, Ericsson
                                               Krister Svanbro, Ericsson
                                                            May 24,
                                                           June 15, 2000

            RObust Checksum-based header COmpression (ROCCO)

                           ROCCO Version: 05 06

Status of this memo

   This document is an Internet-Draft and is in full conformance with
   all provisions of Section 10 of RFC2026.

   Internet-Drafts are working documents of the Internet Engineering
   Task Force (IETF), its areas, and its working groups. Note that other
   groups may also distribute working documents as Internet-Drafts.

   Internet-Drafts are draft documents valid for a maximum of six months
   and may be updated, replaced, or obsoleted by other documents at any
   time. It is inappropriate to use Internet-Drafts as reference
   material or cite them other than as "work in progress".

   The list of current Internet-Drafts can be accessed at

   The list of Internet-Draft Shadow Directories can be accessed at

   This document is a submission of the IETF ROHC WG. Comments should be
   directed to its mailing list,


   Existing header compression schemes do not work well when used over
   links with significant error rates, especially when the round-trip
   time of the link is long. For many bandwidth limited links where
   header compression is essential, such characteristics are common.

   A header compression framework and a highly robust and efficient
   header compression scheme is introduced in this document, adaptable
   to the characteristics of the link over which it is used and also to
   the properties of the packet streams it compresses.

Table of contents

   1.  Introduction..................................................5
   2.  Terminology...................................................7
   3.  Background...................................................10
        3.1.  Header compression fundamentals.......................10
        3.2.  Existing header compression schemes...................10
        3.3.  Requirements on a new header compression scheme.......12
        3.4.  Classification of header fields.......................12
   4.  Header compression framework.................................13
        4.1.  General ROCCO principles..............................14
        4.2.  Data structures.......................................14
        4.3.  Header compression profiles...........................15
        4.4.  Profile negotiation...................................16
        4.5.  Link layer requirements...............................16
   5.  Header compression profiles optimized for voice traffic......17
        5.1.  Usage scenarios, environment and requirements.........17
        5.2.  Profile definitions...................................17
               5.2.1.  List of defined profiles.....................17
               5.2.2.  Additional common profile characteristics....20
        5.3.  Encoding methods used.................................21
               5.3.1.  Least Significant Bits (LSB) encoding........21
               5.3.2.  Least Significant Part (LSP) encoding........21
               5.3.3.  LSB or LSB encoding with extended range......22
        5.4.  Packet formats........................................23
               5.4.1.  Static information packets, initialization...24
               5.4.2.  Dynamic information packets..................25
               5.4.3.  Compressed packets...........................28
               5.4.4.  Extensions to compressed headers.............30
               5.4.5.  Feedback packets.............................35
        5.5.  Encoding of field values..............................37
               5.5.1.  LSP encoding of field values.................37
               5.5.2.  LSB encoding of field values.................37
               5.5.3.  Timer-based timestamp decompression..........38
        5.6.  Header compression CRCs, coverage and polynomials.....39
               5.6.1.  STATIC packet CRC............................39
               5.6.2.  DYNAMIC packet CRC...........................39
               5.6.3.  COMPRESSED packet CRCs.......................39
   6.  Implementation issues........................................40
        6.1. Compressor and decompressor logic, use of feedback.....40
        6.2.  ROCCO over simplex links..............................40
               6.2.1.  Compression slow-start.......................41
               6.2.2.  Periodic refresh.............................41
               6.2.3.  Refresh recommendations......................42
               6.2.4.  Cost and robustness of refreshes.............42
               6.2.5.  Simplex link improvements....................43
        6.3.  Reverse decompression.................................43
        6.4.  Pre-verification of CRCs..............................44
        6.5.  Using "guesses" with LSB and LSP encoding.............45
   7.  Further work.................................................46
        7.1.  Compression of IPv6 extension headers.................46
        7.2.  Efficient compression of CSRC lists...................46
        7.3.  General, media independent profiles...................46
   8.  Implementation status........................................47
   9 . Discussion and conclusions...................................47
   10. Security considerations......................................48
   11. Acknowledgements.............................................49
   12. Intellectual property considerations.........................49
   13. References...................................................50
   14. Authors' addresses...........................................51

   Appendix A.  Detailed classification of header fields............52
        A.1.  General classification................................52
               A.1.1.  IPv6 header fields...........................53
               A.1.2.  IPv4 header fields...........................55
               A.1.3.  UDP header fields............................57
               A.1.4.  RTP header fields............................58
               A.1.5.  Summary for IP/UDP/RTP.......................59
        A.2.  Analysis of change patterns of header fields..........60
               A.2.1.  IPv4 Identification..........................62
               A.2.2.  IP Traffic-Class / Type-Of-Service...........63
               A.2.3.  IP Hop-Limit / Time-To-Live..................63
               A.2.4.  UDP Checksum.................................63
               A.2.5.  RTP CSRC Counter.............................63
               A.2.6.  RTP Marker...................................64
               A.2.7.  RTP Payload Type.............................64
               A.2.8.  RTP Sequence Number..........................64
               A.2.9.  RTP Timestamp................................64
               A.2.10. RTP Contributing Sources (CSRC)..............64
        A.3.  Header compression strategies.........................65
               A.3.1.  Do not send at all...........................65
               A.3.2.  Transmit only initially......................65
               A.3.3.  Transmit initially, update occasionally......65
               A.3.4.  Be prepared to update or send as-is..........66
               A.3.5.  Guarantee continuous robustness..............66
               A.3.6.  Transmit as-is in all packets................66
               A.3.7.  Establish and be prepared to update delta....66

Document history

   The original name of this internet draft was draft-jonsson-robust-hc.
   Since it now has been submitted as a ROHC WG draft, the name has
   changed and then also the numbering. However, the old draft version
   numbering is kept as a version reference for the protocol and this
   one should be referred to as ROCCO 05. 06.

   00  1999-06-22   First release.

   01  1999-09-01   Only small corrections and modifications. Cut-and-
                    paste errors from the 00 draft removed.

   02  1999-10-22   Generalized concept with a number of different
                    profiles. New chapters added describing profile
                    negotiation, implementation status and security.

   03  2000-01-18   LSP encoding and one-octet profiles introduced.
                    Modified and simplified extension formats and
                    small changes in the CONTEXT_REQUEST packets.

   04  2000-03-10   The CONTEXT_UPDATE packet has changed its name to
                    DYNAMIC while also being slightly modified. Both
                    the STATIC and the DYNAMIC packet now include a
                    header compression CRC of 8 bits to ensure
                    reliability of the scheme. Some profiles have been
                    modified, some renumbered, some removed and some
                    added. The CONTEXT_REQUEST packet has been replaced
                    by a more general FEEDBACK packet type with several
                    sub-types including three that leaves much room for
                    implementation features. The profile definitions
                    have been improved in many ways with more details
                    and clarifications. New chapters have been added
                    discussing implementation issues and possible
                    further work.

   05  2000-05-24   New, additional compressed header formats have been
                    added, both for timer-based timestamp decompression
                    and for the cases without that functionality. The
                    extension formats have also been updated. Initial
                    chapters have been reorganized to get a better
                    structure. CID-field has been moved to the beginning
                    of each header. Minor changes have been applied to
                    various parts of the specification.

   06  2000-06-15   Errors in the previous version have been corrected.
                    Since the ROHC group is now creating a final robust
                    header compression scheme, ROCCO has served its
                    purposes and will be terminated. This is therefore
                    the final ROCCO version.

   1.  Introduction

   During the last five years, two communication technologies in
   particular have become commonly used by the general public: cellular
   telephony and the Internet. Cellular telephony has provided its users
   with the revolutionary possibility of always being reachable with
   reasonable service quality no matter where they are. However, until
   now the main service provided has been speech. With the Internet, the
   conditions have been almost the opposite. While flexibility for all
   kinds of usage has been its strength, its focus has been on fixed
   connections and large terminals, and the experienced quality of some
   services (such as Internet telephony) has generally been low.

   Today, IP telephony is gaining momentum thanks to improved technical
   solutions. It seems reasonable to believe that in the years to come,
   IP will become a commonly used way to carry telephony. Some future
   cellular telephony links might also be based on IP and IP telephony.
   Cellular phones may have IP stacks supporting not only audio and
   video, but also web browsing, email, gaming, etc.

   The scenario we are envisioning might then be the one in Figure 1.1,
   where two mobile terminals are communicating with each other. Both
   are connected to base stations over cellular links, and the base
   stations are connected to each other through a wired (or possibly
   wireless) network. Instead of two mobile terminals, there could of
   course be one mobile and one wired terminal, but the case with two
   cellular links is technically more demanding.

   Mobile            Base                      Base            Mobile
   Terminal          Station                   Station         Terminal

         |  ~   ~   ~  \ /                       \ /  ~   ~   ~   ~  |
         |              |                         |                  |
      +--+              |                         |               +--+
      |  |              |                         |               |  |
      |  |              |                         |               |  |
      +--+              |                         |               +--+
                        |                         |

            Cellular              Wired               Cellular
            Link                  Network             Link

        Figure 1.1 : Scenario for IP telephony over cellular links

   It is obvious that the wired network can be IP-based. With the
   cellular links, the situation is less clear. IP could be terminated
   in the fixed network, and special solutions implemented for each
   supported service over the cellular link. However, this would limit
   the flexibility of the services supported. If technically and
   economically feasible, a solution with pure IP all the way from
   terminal to terminal would have certain advantages. However, to make
   IP-all-the-way a viable alternative, a number of problems have to be
   addressed, especially regarding bandwidth efficiency.

   For cellular phone systems, it is of vital importance to use the
   scarce radio resources in an efficient way. A sufficient number of
   users per cell is crucial, otherwise deployment costs will be
   prohibitive [CELL]. The quality of the voice service should also be
   as good as in today's cellular systems. It is likely that even with
   support for new services, lower quality of the voice service is
   acceptable only if costs are significantly reduced.

   A problem with IP over cellular links when used for interactive voice
   conversations is the large header overhead. Speech data for IP
   telephony will most likely be carried by RTP [RTP]. A packet will
   then, in addition to link layer framing, have an IP [IPv4] header (20
   octets), a UDP [UDP] header (8 octets), and an RTP header (12 octets)
   for a total of 40 octets. With IPv6 [IPv6], the IP header is 40
   octets for a total of 60 octets. The size of the payload depends on
   the speech coding and frame sizes used and may be as low as 15-20

   From these numbers, the need for reducing header sizes for efficiency
   reasons is obvious. However, cellular links have characteristics that
   make header compression as defined in [IPHC,CRTP,PPPHC] perform less
   than well. The most important characteristic is the lossy behavior of
   cellular links, where a bit error rate (BER) as high as 1e-3 must be
   accepted to keep the radio resources efficiently utilized [CELL]. In
   severe operating situations, the BER can be as high as 1e-2. The
   other problematic characteristic is the long round-trip time (RTT) of
   the cellular link, which can be as high as 100-200 milliseconds
   [CELL]. A viable header compression scheme for cellular links must be
   able to handle loss on the link between the compression and
   decompression point as well as loss before the compression point.

   Bandwidth is the most costly resource in cellular links. Processing
   power is very cheap in comparison. Implementation or computational
   simplicity of a header compression scheme is therefore of less
   importance than its compression ratio and robustness.

2.  Terminology

   The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT",
   document are to be interpreted as described in RFC 2119.


    Bit Error Rate. Cellular radio links have a rather high BER. In
    this document BER is usually given as a frequency, but one also
    needs to consider the error distribution as bit errors are not
    independent. In our simulations we use a channel with a certain
    BER, and the error distribution is according to a realistic channel

   Cellular links

    Wireless links between mobile terminals and base stations. The BER
    and the RTT are rather high in order to achieve an efficient system

   Compression efficiency

    The performance of a header compression scheme can be described
    with three parameters, compression efficiency, robustness and
    compression reliability. The compression efficiency is determined
    by how much the header sizes are reduced by the compression scheme.

   Compression reliability

    The performance of a header compression scheme can be described
    with three parameters, compression efficiency, robustness and
    compression reliability. The compression reliability is a measure
    for how well the scheme ensures that the decompressed headers are
    not erroneous and the possibility to avoid error propagation from
    the decompressor.


    The context is the state which the compressor uses to compress a
    header and which the decompressor uses to decompress a header. The
    context is basically the uncompressed version of the last header
    sent (compressor) or received (decompressor) over the link, except
    for fields in the header that are included "as-is" in compressed
    headers or can be inferred from, e.g., the size of the link-level
    frame. The context can also contain additional information
    describing the packet stream, for example the typical inter-packet
    increase in sequence numbers or timestamps.

   Context damage

    When the context of the decompressor is not consistent with the
    context of the compressor, header decompression will fail. This
    situation can occur when the context of the decompressor has not
    been initialized properly or when packets have been lost or damaged
    between compressor and decompressor. Packets which cannot be
    decompressed due to inconsistent contexts are said to be lost due
    to context damage.

   Context repair mechanism

    To avoid excessive context damage, a context repair mechanism is
    needed. Context repair mechanisms can be based on explicit requests
    for context updates, periodic updates sent by the compressor, or
    methods for local repair at the decompressor side.


    Frame Error Rate. The FER considered in this document includes the
    frames lost on the channel between compressor and decompressor and
    frames lost due to context damage. FER is here defined to be
    identical to packet loss rate.

   Header compression profile

    A header compression profile is a specification of how to compress
    the headers of a certain kind of packet stream over a certain kind
    of link. Compression profiles provide the details of the header
    compression framework introduced in this document. The profile
    concept makes use of profile identifiers to separate different
    profiles which are used when setting up the compression scheme. All
    variations and parameters of the header compression scheme are
    handled by different profile identifiers, which makes the number of
    profiles rather large. This can act as a deterrent when first
    studying the concept, but is a real strength for several reasons.
    One advantage of this merging of parameters into one is that new
    parameters can be added by the endpoints without affecting the
    negotiation requirements on the link in between. Another benefit of
    the concept is that different combinations of functionality might
    be implemented with different methods, meaning that the scheme can
    be optimized regardless of what functionality is enabled. Finally,
    it should be noted that even if there are a large number of
    profiles, only a small number of them can/will be implemented over
    a specific link (IPv4 and IPv6 profiles will for example probably
    not coexist). Most profiles usable in a certain environment will
    probably also be almost identical from an implementation point of

   Header compression CRC

    A CRC (Cyclic Redundancy Checksum) computed by the compressor and
    included in each compressed header. Its main purpose is to provide
    a way for the decompressor to reliably verify the correctness of
    reconstructed headers. What values the CRC is computed over depends
    on the packet type it is included in; typically it covers most of
    the original header fields.

   Pre-HC links

    Pre-HC links are all links before the header compression point. If
    we consider a path with cellular links as first and last hops, the
    Pre-HC links for the compressor at the last link are the first
    cellular link plus the wired links in between.


    The performance of a header compression scheme can be described
    with three parameters, compression efficiency, robustness and
    compression reliability. A robust scheme tolerates errors on the
    link over which header compression takes place without losing
    additional packets, introducing additional errors, or using more


    Round Trip Time - The time it takes to send a packet back and forth
    over the link.

   Simplex link

    A simplex (or unidirectional) link is a point to point link without
    a return channel. Over simplex links, header compression must rely
    on periodic refreshes since feedback from the decompressor can not
    be sent to the compressor. For simplex links, a header compression
    CRC is mandatory to guarantee correct decompression.

   Spectrum efficiency

    Radio resources are limited and expensive. Therefore they must be
    used efficiently to make the system economically feasible. In
    cellular systems this is achieved by maximizing the number of users
    served within each cell, while the quality of the provided services
    is kept at an acceptable level. A consequence of efficient spectrum
    use is a high BER, even after channel coding with error correction.

   Timestamp delta

    The timestamp delta is the increase in the timestamp value between
    two consecutive packets.

3.  Background

   This chapter provides a background to the subject of header
   compression. The fundamental ideas are described together with
   descriptions of existing header compression schemes, their drawbacks
   and requirements and motivation for new header compression solutions.

3.1.  Header compression fundamentals

   The main reason why header compression can be done at all is the fact
   that there are lots of redundancy between header fields, both within
   the same packet header but especially between consecutive packets
   belonging to the same packet stream. By sending static field
   information only initially and utilize dependencies and
   predictability's for other fields, the header size can be
   significantly reduced for most packets.

   In general, header compression methods maintain a context, which is
   essentially  the uncompressed version of the last header sent over
   the link, at both compressor and decompressor. Compression and
   decompression are done relative to the context. When compressed
   headers carry differences from the previous header, each compressed
   header will update the context of the decompressor. When a packet is
   lost between compressor and decompressor, the context of the
   decompressor will be brought out of sync since it is not updated
   correctly. A header compression method must have a way to repair the
   context, i.e. bring it into sync, after such events.

3.2.  Existing header compression schemes

   The original header compression scheme, CTCP [VJHC], was invented by
   Van Jacobson. CTCP compressed the 40 octet IP+TCP header to 4 octets.

   The CTCP compressor detects transport-level retransmissions and sends
   a header that updates the context completely when they occur. This
   repair mechanism does not require any explicit signaling between
   compressor and decompressor.

   CRTP [CRTP, IPHC] by Casner and Jacobson is a header compression
   scheme that compresses 40 octets of IPv4/UDP/RTP headers to a minimum
   of 2 octets when no UDP checksum is present. If the UDP checksum is
   present, the minimum CRTP header is 4 octets. CRTP cannot use the
   same repair mechanism as CTCP since UDP/RTP does not retransmit.
   Instead, CRTP uses explicit signaling messages from decompressor to
   compressor, called CONTEXT_STATE messages, to indicate that the
   context is out of sync. The link roundtrip time will thus limit the
   speed of this context repair mechanism.

   On lossy links with long roundtrip times, such as most cellular
   links, CRTP does not perform well. Each lost packet over the link
   causes several subsequent packets to be lost since the context is out
   of sync during at least one link roundtrip time. This behavior is
   documented in [CRTPC]. For voice conversations such long loss events
   will degrade the voice quality. Moreover, bandwidth is wasted by the
   large headers sent by CRTP when updating the context. [CRTPC] found
   that CRTP performed much worse than ideally for a lossy cellular
   link. It is clear that CRTP alone is not a viable header compression
   scheme for cellular links.

   To avoid losing headers due to the context being out of sync, CRTP
   decompressors can attempt to repair the context locally by using a
   mechanism known as TWICE. Each CRTP packet contains a counter which
   is incremented by one for each packet sent out by the CRTP
   compressor. If the counter increases by more than one, at least one
   packet was lost over the link. The decompressor then attempts to
   repair the context by guessing how the lost packet(s) would have
   updated it. The guess is then verified by decompressing the packet
   and checking the UDP checksum - if it succeeds, the repair is deemed
   successful and the packet can be forwarded or delivered. TWICE has
   got its name from the observation that when the compressed packet
   stream is regular, the correct guess is to apply the update in the
   current packet twice. [CRTPC] found that even with TWICE, CRTP
   doubled the number of lost packets. TWICE improves CRTP performance
   significantly. However, there are several problems with using TWICE:

   1) It becomes mandatory to use the UDP checksum:

      - the minimal compressed header size increases by 100% to 4

      - most speech codecs developed for cellular links tolerate errors
        in the encoded data. Such codecs will not want to enable the UDP
        checksum, since they want damaged packets to be delivered.

      - errors in the payload will make the UDP checksum fail when the
        guess is correct (and might make it succeed when it is wrong).

   2) Loss in an RTP stream that occurs before the compression point
      will make updates in CRTP headers less regular. Simple-minded
      versions of TWICE will then perform badly. More sophisticated
      versions would need more repair attempts to succeed.

3.3.  Requirements on a new header compression scheme

   The major problem with CRTP is that it is not sufficiently robust
   against packets being damaged between compressor and decompressor. A
   viable header compression scheme must be less fragile. This increased
   robustness must be obtained without increasing the compressed header
   size; a larger header would make IP telephony over cellular links
   economically unattractive.

   A major cause of the bad performance of CRTP over cellular links is
   the long link roundtrip time, during which many packets are lost when
   the context is out of sync. This problem can be attacked directly by
   finding ways to reduce the link roundtrip time. Future generations of
   cellular technologies may indeed achieve lower link roundtrip times.
   However, these will probably always be rather high [CELL]. The
   benefits in terms of lower loss and smaller bandwidth demands if the
   context can be repaired locally will be present even if the link
   roundtrip time is decreased. A reliable way to detect a successful
   context repair is then needed.

   One might argue that a better way to solve the problem is to improve
   the cellular link so that packet loss is less likely to occur. It
   would of course be nice if the links were almost error free, but such
   a system would not be able to support a sufficiently large number of
   users per cell and would thus be economically unfeasible [CELL].

   One might also argue that the speech codecs should be able to deal
   with the kind of packet loss induced by CRTP, in particular since the
   speech codecs probably must be able to deal with packet loss anyway
   if the RTP stream crosses the Internet. While the latter is true, the
   kind of loss induced by CRTP is difficult to deal with. It is usually
   not possible to hide a loss event where well over 100 ms worth of
   sound is completely lost. If such loss occurs frequently at both ends
   of the path, the speech quality will suffer.

3.4.  Classification of header fields

   As mentioned earlier, header compression is possible due to the fact
   that there are much redundancy between header field values within
   packets, but especially between consecutive packets. To utilize these
   properties for header compression, it is important to understand the
   behavior of the various header fields. To do this, all header fields
   have been classified in detail in appendix A. The fields are first
   classified on a high level and then some of them are studied more in
   detail. Finally, the appendix concludes with recommendations about
   how the various fields should be handled by header compression
   algorithms. The main conclusion that can be drawn is that most of the
   header fields can easily be compressed away since they are never or
   seldom changing. Only 5 fields with a total size of about 10 octets
   are rather difficult to compress and must be handled in a
   sophisticated way by the compression scheme. Those fields are:

    - IPv4 Identification (16 bits)
    - UDP Checksum (16 bits)
    - RTP Marker (1 bit)
    - RTP Sequence Number (16 bits)
    - RTP Timestamp (32 bits)

   It is rather obvious that these field then will have a large impact
   on how a header compression scheme is designed. However, all details
   about this should be found in Appendix A.

4.  Header compression framework

   A solution is proposed for efficient, robust and reliable header
   compression, ROCCO. The scheme is heavily geared towards local repair
   of the context in combination with robust encoding of header fields.
   What is needed is a reliable way to detect when a repair attempt has
   succeeded, plus possibly hints to the decompressor about how the
   header fields have changed.

   The key element of the proposed header compression scheme is that
   compressed headers should carry a CRC computed over the header before
   compression. This provides a reliable way to detect whether
   decompression and context repair have succeeded. In addition to the
   CRC, the header could contain codes and additional information as
   needed for decompression.

   A completely general solution cannot achieve compression rates high
   enough to make IP telephony over cellular economically feasible. On
   the other hand, the solution needs to be extendable so that other
   kinds of packet streams can also be compressed well. Therefore, we
   envision a scheme where the basic framework is supplemented with a
   set of compression profiles, where each compression profile provides
   the exact details on how a packet stream is to be compressed and
   decompressed. The use of compression profiles allows the basic
   framework to be adapted to the properties of packet streams as well
   as various link properties.

   The ROCCO header compression framework does not state any exact
   details about how the compression is to be performed, what formats
   the headers should have, etc. This is left to the compression
   profiles to define. The framework instead describes general
   principles for how to do header compression "the ROCCO way". The
   header compression profile concept is presented describing what a
   profile is, what to consider when designing a profile and what every
   profile must or should define. The concept also exactly states the
   rules regarding negotiation of compression profiles.

4.1.  General ROCCO principles

   ROCCO header compression is based on the principle of decompressor
   context repair attempts relying on a header compression CRC included
   in compressed headers. Profiles will define various packet types and
   all of them do not have to carry a header compression CRC. In
   general, if the CRC is present it is RECOMMENDED to calculate it over
   the uncompressed header, but profiles MAY define the coverage
   differently for some packet types.

   Distinguishing packet streams and packet types is necessary, but some
   of that information may be available from the underlying technology.
   To avoid wasting precious header bits, it is left to the compression
   profile to decide how to distinguish packet types and packet streams.
   This allows efficient use of header bits overall.

   If each packet stream has its own logical channel, it is not
   necessary to have any additional information for distinguishing
   between streams. Otherwise there MUST be slightly different profiles
   defined with support for various numbers of concurrent packet

   The link layer could carry explicit information about packet types,
   but that would not lead to an efficient use of bits, since different
   profiles could use different number of packet types. If the packet
   type distinguishing mechanism is included in the header compression
   profile instead, the profile could optimize the bit usage of that
   mechanism to its own packet types. However, it is up to the profile
   designer to choose if this mechanism is included in the profile or
   required from the link layer.

   A compression profile MAY define headers which do not have a
   corresponding original packet. Such packets would be internal header
   compression packets, and would not be delivered further from the
   decompressor. An example would be to initially send non-changing
   fields of a packet stream as a separate packet. Another example would
   be to send packets to update the RTP timestamp field even when no RTP
   packets arrive, in order to decrease the increment in the RTP
   timestamp field when a packet does arrive.

   The profiles defined in this document SHOULD be considered as
   examples for how profiles are supposed to be defined and described.

4.2.  Data structures

   The compression scheme needs to maintain a context for compression
   and decompression of a packet stream. The context must be kept
   updated at both compressor and decompressor. The context is
   essentially the header of the last packet transmitted, and includes
   all static header fields plus some fields that change more or less
   frequently. If the compression profile used is designed to handle a
   certain amount of packet loss on the link, both compressor and
   decompressor will typically keep information about earlier packets;
   packets that arrived before the current packet. Finally, there may be
   packet stream related information such as field deltas (e.g. RTP
   timestamp) or a list of which CSRC items have occurred in the packet

4.3.  Header compression profiles

   The details on how a packet stream is to be compressed and
   decompressed are determined by a compression profile. Over a link a
   number of profiles can be active, either within the same or within
   different logical channels. How to determine what profile to use for
   a certain packet stream is not defined in this document, but the
   usage MUST be negotiated between compressor and decompressor as
   described in the subsequent chapter.

   One way to select a suitable profile is to have a common channel over
   which a general-purpose header compression profile is used. When the
   packet stream characteristics are identified, it is switched to
   another channel where a suitable compression profile is used.

   Profiles can be defined to compress one packet stream only, in which
   case the link layer must be able to separate packet streams. Profiles
   can also be defined for compression of more than one packet stream,
   in which case the profile must provide a way to distinguish between
   the packet streams.

   Important parameters to consider when designing a profile are:

     - what kind of packet streams to compress (IPv6, IPv4, UDP,
       UDP/RTP, TCP) and if UDP, whether the UDP checksum is supported.

     - the rate and pattern of loss of the channel.

     - the pattern of change of the changing fields.

     - the expected rate and pattern of loss and reordering before the
       compression point.

     - if included in the profile, the number of concurrent packet
       streams supported through context identifiers.

     - what support there is from the link layer, such as the number of
       packet streams supported, and if it can indicate packet types.

   When these things have been considered, the specifics of the profile
   can be determined. The profile MUST specify:

     - the exact semantics of the various packet types and how the
       desired functionality is supported.

     - the size of, polynomials for, and what the Header Compression
       CRC covers for all packet types.

     - the information needed in the contexts for compression and
       decompression, including history information and properties of
       the packet stream.

     - procedures for compression and decompression.

     - how compression is initiated (which packets are used and how).

     - description of context repair mechanisms.

   Chapter 5 defines compression profiles optimized for conversational
   voice traffic over cellular radio links.

4.4.  Profile negotiation

   To initiate ROCCO header compression, compressor and decompressor
   must be able to negotiate which header compression profile to use. A
   header compression profile is identified by a 16 bit profile
   identifier, and underlying link layers MUST provide a way to
   negotiate this.

4.5.  Link layer requirements

   This chapter lists general ROCCO requirements on an underlying link
   layer. Profiles could also state additional requirements, but these
   MUST be provided for ALL ROCCO profiles. See also [LLG].


     Framing, which makes it possible to separate different packets, is
     the most important link layer functionality.


     Most link layers can indicate the length of the packet, and this
     information has therefore been removed from the packet headers.
     This means that it now MUST be given by the link layer.

   Profile negotiation

     In addition to the packet handling mechanisms above, the link
     layer MUST also provide a way to carry on the negotiation of
     header compression profiles, described in chapter 5.4.

5.  Header compression profiles optimized for voice traffic

   This section exemplifies how the framework outlined in chapter 4
   could be instantiated by defining profiles optimized for header
   compression of packet streams carrying voice data. A number of
   profiles are defined providing support for both IPv6 and IPv4 in
   combination with various functionality.

5.1.  Usage scenarios, environment and requirements

   These profiles are optimized for voice traffic over error prone
   links. The profiles are designed to successfully handle loss of
   several consecutive packets over the link, without introducing any
   additional loss. Packet type identification is included in all
   profiles, which means that such functionality SHOULD NOT be provided
   by the link layer used. Regarding packet stream separation, various
   profiles are defined supporting different numbers of concurrent
   As an error prone link with similar characteristics is expected at
   the other end of the connection (see Figure 1.1), the profiles are
   also designed to handle some consecutive lost packet before the
   compression point without increasing the size of the compressed
   header. The profiles are also in general designed to handle
   reordering of single packets before the compression point without
   increasing the size of compressed headers.

5.2.  Profile definitions

   This document defines a number of different header compression
   profiles. The definitions are built up of the requirements on and
   capabilities of each profile in combination with information about
   which mechanisms are used to implement the desired behavior.

5.2.1.  List of defined profiles

   All defined profiles are listed in Table 5.1 together with their
   characteristics, capabilities and pointers to implementation details
   that may differ from profile to profile. For more information about
   the profile concept see chapter 4.3 and "Header compression profile"
   in the Terminology chapter.

   The first seven columns state requirements on and capabilities of the
   profiles. The meaning of the columns are:

   Nr     This is the identification number for each profile. These
          numbers are used when negotiating profiles in the header
          compression setup phase. The numbers are preliminary and will
          perhaps change in future versions of this profile

   IPv    This is the IP version for which the profile is designed.
          Possible values for this column are 6 and 4.

   CID    This column gives the number of concurrent packet streams
          that are supported by the header compression profile through
          context identifiers (CIDs).

   Chk    This column indicates whether the profile supports packet
          streams with the UDP checksum (E)nabled or D(isabled).

   ID     For profiles supporting IPv4, this column indicates which
          behavior of the IPv4 Identification field the profile is
          optimized for. Possible values in this column are:

          (S)EQUENTIAL           These profiles can handle all kind of
                                 Identification assignment methods but
                                 will be less efficient than RANDOM
                                 profiles if the assignment truly is
                                 random. If the value is sequentially
                                 assigned, no extra overhead is added
                                 for Identification.

          (R)ANDOM               These profiles are recommended if it is
                                 known that random assignment is used.
                                 The Identification field will be
                                 included "as-is" which means that the
                                 header size will increase by two

   TbT    Timer-based Timestamp decompression. Requires a timer at the
          decompressor side to estimate timestamp jumps. Compressor
          never sends more than a few bits of timestamp LSB with these
          profiles. Can be (E)nabled or (D)isabled (see chapter 5.5.3).

   S      S gives the minimal header Size for the profile.

   The next five columns indicate how each profile is implemented. This
   includes header formats for STATIC (STA, see chapter 5.4.1), DYNAMIC
   (DYN, see chapter 5.4.2) and COMPRESSED (COM, see chapter 5.4.3)
   packets, and also what EXTENSION (EXT, see chapter 5.4.4) formats are
   used with the COMPRESSED packets. The CRC column tells the coverage
   of the header compression CRC: uncompressed (H)eader or the same
   coverage as for the UDP (C)hecksum (see chapter 5.6).

    +-----+---+-----+---+---+---+  +---+  +---+---+-------+---+---+
    |  N  | I |  C  | C | I | T |  | S |  | S | D |   C   | E | C |
    |  r  | P |  I  | h | D | b |  |   |  | T | Y |   O   | X | R |
    |     | v |  D  | k |   | T |  |   |  | A | N |   M   | T | C |
    +=====+===+=====+===+===+===+  +===+  +===+===+=======+===+===+
    |  1  | 6 |  1  | E | - | E |  | 2 |  | 1 | 1 |   1   | C A | C |
    +-----+---+-----+---+---+---+  +---+  +---+---+-------+---+---+
    |  2  | 6 |  1  | E | - | D |  | 2 |  | 1 | 1 |   1   | A | C |
    +-----+---+-----+---+---+---+  +---+  +---+---+-------+---+---+
    |  3  | 6 | 256 | E | - | E |  | 3 |  | 2 | 2 |   2   | C A | C |
    +-----+---+-----+---+---+---+  +---+  +---+---+-------+---+---+
    |  4  | 6 | 256 | E | - | D |  | 3 |  | 2 | 2 |   2   | A | C |
    +-----+---+-----+---+---+---+  +---+  +---+---+-------+---+---+
    |  5  | 4 |  1  | D | S | E |  | 1 |  | 3 | 3 | 5, 9  | D | H |
    +-----+---+-----+---+---+---+  +---+  +---+---+-------+---+---+
    |  6  | 4 |  1  | D | S | E |  | 2 |  | 3 | 3 |   1   | D | H |
    +-----+---+-----+---+---+---+  +---+  +---+---+-------+---+---+
    |  7  | 4 |  1  | D | S | D |  | 1 |  | 3 | 3 | 5, 13 | B | H |
    +-----+---+-----+---+---+---+  +---+  +---+---+-------+---+---+
    |  8  | 4 |  1  | D | S | D |  | 2 |  | 3 | 3 |   1   | B | H |
    +-----+---+-----+---+---+---+  +---+  +---+---+-------+---+---+
    |  9  | 4 |  1  | D | R | E |  | 3 |  | 3 | 3 | 7, 11 | D C | H |
    +-----+---+-----+---+---+---+  +---+  +---+---+-------+---+---+
    | 10  | 4 |  1  | D | R | E |  | 4 |  | 3 | 3 |   3   | D C | H |
    +-----+---+-----+---+---+---+  +---+  +---+---+-------+---+---+
    | 11  | 4 |  1  | D | R | D |  | 3 |  | 3 | 3 | 7, 15 | B A | H |
    +-----+---+-----+---+---+---+  +---+  +---+---+-------+---+---+
    | 12  | 4 |  1  | D | R | D |  | 4 |  | 3 | 3 |   3   | B A | H |
    +-----+---+-----+---+---+---+  +---+  +---+---+-------+---+---+
    | 13  | 4 |  1  | E | S | E |  | 2 |  | 3 | 5 |   1   | D | C |
    +-----+---+-----+---+---+---+  +---+  +---+---+-------+---+---+
    | 14  | 4 |  1  | E | S | D |  | 2 |  | 3 | 5 |   1   | B | C |
    +-----+---+-----+---+---+---+  +---+  +---+---+-------+---+---+
    | 15  | 4 |  1  | E | R | E |  | 4 |  | 3 | 5 |   3   | D C | C |
    +-----+---+-----+---+---+---+  +---+  +---+---+-------+---+---+
    | 16  | 4 |  1  | E | R | D |  | 4 |  | 3 | 5 |   3   | B A | C |
    +-----+---+-----+---+---+---+  +---+  +---+---+-------+---+---+
    | 17  | 4 | 256 | D | S | E |  | 2 |  | 4 | 4 | 6, 10 | D | H |
    +-----+---+-----+---+---+---+  +---+  +---+---+-------+---+---+
    | 18  | 4 | 256 | D | S | E |  | 3 |  | 4 | 4 |   2   | D | H |
    +-----+---+-----+---+---+---+  +---+  +---+---+-------+---+---+
    | 19  | 4 | 256 | D | S | D |  | 2 |  | 4 | 4 | 6, 14 | B | H |
    +-----+---+-----+---+---+---+  +---+  +---+---+-------+---+---+
    | 20  | 4 | 256 | D | S | D |  | 3 |  | 4 | 4 |   2   | B | H |
    +-----+---+-----+---+---+---+  +---+  +---+---+-------+---+---+
    | 21  | 4 | 256 | D | R | E |  | 4 |  | 4 | 4 | 8, 12 | D C | H |
    +-----+---+-----+---+---+---+  +---+  +---+---+-------+---+---+
    | 22  | 4 | 256 | D | R | E |  | 5 |  | 4 | 4 |   4   | D C | H |
    +-----+---+-----+---+---+---+  +---+  +---+---+-------+---+---+
    | 23  | 4 | 256 | D | R | D |  | 4 |  | 4 | 4 | 8, 16 | B A | H |
    +-----+---+-----+---+---+---+  +---+  +---+---+-------+---+---+
    +-----+---+-----+---+---+---+  +---+  +---+---+---+---+---+
    |  N  | I |  C  | C | I | T |  | S |  | S | D | C | E | C |
    |  r  | P |  I  | h | D | b |  |   |  | T | Y | O | X | R |
    |     | v |  D  | k |   | T |  |   |  | A | N | M | T | C |
    +=====+===+=====+===+===+===+  +===+  +===+===+===+===+===+
    | 24  | 4 | 256 | D | R | D |  | 5 |  | 4 | 4 | 4 | B A | H |
    +-----+---+-----+---+---+---+  +---+  +---+---+---+---+---+
    | 25  | 4 | 256 | E | S | E |  | 3 |  | 4 | 6 | 2 | D | C |
    +-----+---+-----+---+---+---+  +---+  +---+---+---+---+---+
    | 26  | 4 | 256 | E | S | D |  | 3 |  | 4 | 6 | 2 | B | C |
    +-----+---+-----+---+---+---+  +---+  +---+---+---+---+---+
    | 27  | 4 | 256 | E | R | E |  | 5 |  | 4 | 6 | 4 | D C | C |
    +-----+---+-----+---+---+---+  +---+  +---+---+---+---+---+
    | 28  | 4 | 256 | E | R | D |  | 5 |  | 4 | 6 | 4 | B A | C |
    +-----+---+-----+---+---+---+  +---+  +---+---+---+---+---+

                   Table 5.1 : List of defined profiles

5.2.2.  Additional common profile characteristics

   In addition to what was stated in the left part of Table 5.1, the
   following applies to all the profiles defined in this document:

   Packet stream characteristics

     These profiles are designed for packet streams carrying
     conversational voice data.

   Link layer requirements

     Except for the general link layer requirements (framing, length &
     profile negotiation) stated in chapter 4.5, these profiles also
     require a reliable link layer CRC covering at least the header part
     of the packet. The CRC SHOULD ensure that packets with errors in
     the header part are never delivered.

   Pre link characteristics

     With these profiles, several consecutive packet losses before the
     header compression point are handled without introducing additional
     header overhead. Packet reordering on pre links is expected to be
     uncommon but is handled, very efficiently when limited.

   Link Characteristics

     The link over which header compression is performed is expected to
     have a loss characteristic that very seldom leads to loss of many
     consecutive packets. These profiles can without extra
     reconstruction attempts handle loss of up to 25 consecutive packets
     over the link without losing context, and even if loss on pre links
     decreases this robustness, it should be more than sufficient for
     all realistic scenarios. The round-trip time of the link is
     expected to be between 100 and 200 milliseconds.

5.3.  Encoding methods used

   The analysis of header field changes in appendix A excluded changes
   due to loss and/or reordering before the header compression point.
   Such changes will have an impact on the regularity of the RTP
   sequence number, the RTP timestamp value and, for IPv4, the IP ID
   value. However, as described in A.2, both the RTP timestamp and the
   IP ID value (if sequentially assigned) are expected to follow the RTP
   sequence number for most packets. The most important task is then to
   communicate RTP sequence number information in an efficient way. The
   profiles defined in this document make use of two different methods
   of handling the sequence number field and also other fields, LSB
   encoding and LSP encoding. The methods are interpreted in different
   ways for different fields, and this chapter therefore describes these
   methods in a general way. How these two methods are applied to fields
   in different compressed headers is described in chapter 5.6.

5.3.1.  Least Significant Bits (LSB) encoding

   A commonly used method for updating fields whose values are always
   subject to small changes (usually positive) is Least Significant Bits
   (LSB) encoding. For example, an increase of up to 16 could be handled
   with only 4 bits with LSB encoding (if decreases are not expected).
   This method is used for many different fields by the ROCCO profiles
   defined in this document, especially when information such as
   timestamps is sent in EXTENDED COMPRESSED headers. If a field is
   labeled "<fieldname> LSB", it means that the field contains only the
   least significant bits of the corresponding original field.

5.3.2.  Least Significant Part (LSP) encoding

   One restriction with LSB encoding is that whole bits are needed,
   meaning that only 2, 4, 8, 16, 32, ... code-points could be used. In
   some cases, especially when several mechanisms are integrated for
   efficiency reasons, it would be desirable to have a method that could
   make use of any number of available code-points. To signal one
   special event one could either use one single bit or, if the event is
   not to be signaled in parallel with other information, as one bit
   pattern for several bits. That would leave more bit patterns for
   other usage.

   Assume that we have 4 special events to signal and 5 bits available.
   Taking 2 bits for these events, then there would be 3 bits (8 code-
   points) left for other usage. If we instead reserved 4 of the code-
   points represented by all 5 bits, there would instead be 32-4=28
   code-points left for other usage. The only disadvantage would be that
   the bits cannot be used for both purposes at the same time.

   What would be desirable is to do LSB encoding of code-points instead
   of whole bits. Therefore the method called Least Significant Part
   (LSP) encoding has been introduced. LSP encoding of size (number of
   code-points) M for a value N is defined as:

     LSP:M(N) = N modulo M

   An example showing the LSP encoding and decoding of a counter S(n)
   with M code-points is used below to illustrate the LSP principle.
   S'(n) is the decoded value corresponding to the original S(n) value.
   With S'(n-1), we denote the last correctly decompressed value.

     Input sequence: S(n)
     Encoded sequence: LSP:M(S(n)) = S(n) modulo M
     Decoded sequence: S'(n) = S'(n-1) - LSP:M(S'(n-1)) + LSP:M(S(n)) =
                               S'(n-1) - S(n-1) modulo M + S(n) modulo M

   To handle modulo wrap-around, an additional verification is inserted.
   If the decoded value S'(n) is smaller than S'(n-1)-R, S'(n) is
   increased with M (reordering of order R is then handled with this

   When applying LSP encoding, there are thus two parameters that must
   be set:

      M - The number of code-points to use (modulo value)
      R - The reordering order to handle

   A similar mechanism as for modulo wrap-around should also be used to
   handle full-field wrap-around.

5.3.3.  LSB or LSP encoding with extended range

   If needed, it could be good to extend the range covered by the LSB or
   LSP encoding. For the LSB case, it is simple to send only the next
   more significant bits. For the LSP, what must be done is to rewrite
   the definition of LSP so that it defines an additional parameter.

   The LSP definition from previous chapter can instead be expressed as:

     LSP:M(N) = N - INT:M(N)*M      [ INT:M(N) = (N - LSP:M(N)) / M) ]

   And in that case, INT:M(N) is the integer part left after division.
   If additional bits can be transmitted to increase the range covered,
   this can be done by sending the least significant bits (LSB) of this
   integer part, INT:M(N). The example from previous chapter will then
   change into:

     Input sequence: S(n)
     Encoded sequence: LSP:M(S(n)) = S(n) modulo M
                       INT:M(S(n)) = (S(n)-LSP:M(S(n))) / M
     Decoded sequence: S'(n) = S'(n-1) - LSP:M(S'(n-1)) * M
                                       + LSP:M(S(n)) * M
                                       - LSB(INT:M(S'(n-1))) * M
                                       + LSB(INT:M(S(n))) * M

5.4.  Packet formats

   The profiles defined in this document make use of four different

   To identify packet types, 4 bit patterns for the initial 5 bits of
   the first octet (not including a potential CID field) in all packets
   are reserved. These patterns are:

     STATIC       11111
     DYNAMIC      1110*    (both 11100 and 11101 are reserved for this)
     FEEDBACK     11110

   The other 28 (32-4) bit patterns indicate a COMPRESSED packet format
   and the usage of these patterns are explained further on.

   This section defines the header formats of the four ordinary packet
   types STATIC, DYNAMIC, COMPRESSED and FEEDBACK together with
   descriptions of when and how to use them. A subsections is also
   dedicated to the EXTENSION formats of COMPRESSED headers.

5.4.1.  Static information packets, initialization

   The STATIC packet type is a packet containing no payload but only the
   header fields that are expected to be constant throughout the
   lifetime of the packet stream (classified as STATIC in appendix A). A
   packet of this kind MUST be sent once as the first packet from
   compressor to decompressor and also when requested by the
   decompressor (see chapter 5.4.5). The packet formats are shown below
   for IPv6 and IPv4, respectively. Note that some fields are only
   present in some of the STATIC packet types.

   IPv6 (45-46 octets): STATIC1, STATIC2:

          0   1   2   3   4   5   6   7
        :   Context Identifier  (CID)   :   only present in STATIC2
        | 1   1   1   1   1 | - | - | - |
        |                               |
        +          Flow Label           +
        |                               |
        +               +---+---+---+---+
        |               | - | - | P | E |
        |                               |
        /        Source Address         /   16 octets
        |                               |
        |                               |
        /      Destination Address      /   16 octets
        |                               |
        |                               |
        +          Source Port          +
        |                               |
        |                               |
        +       Destination Port        +
        |                               |
        |                               |
        /             SSRC              /   4 octets
        |                               |
        |    Header Compression CRC     |   see chapter 5.6.1.
   IPv4 (18-19 octets): STATIC3, STATIC4:

          0   1   2   3   4   5   6   7
        :   Context Identifier  (CID)   :   only present in STATIC4
        | 1   1   1   1   1 | F | P | E |
        |                               |
        /        Source Address         /   4 octets
        |                               |
        |                               |
        /      Destination Address      /   4 octets
        |                               |
        |                               |
        +          Source Port          +
        |                               |
        |                               |
        +       Destination Port        +
        |                               |
        |                               |
        /             SSRC              /   4 octets
        |                               |
        |    Header Compression CRC     |   see chapter 5.6.1.

   All fields except for the initial five bits, the padding (-) and the
   Header Compression CRC are the ordinary IP, UDP and RTP fields
   (F=IPv4 May Fragment, P=RTP Padding, E=RTP Extension).

   The number of STATIC packets sent on each occasion should be limited.
   If the decompressor receives DYNAMIC or COMPRESSED headers without
   having received a STATIC packet, the decompressor MUST send a

5.4.2. Dynamic information packets

   The DYNAMIC packet type has a header containing all changing header
   fields in their original, uncompressed form, and carries a payload
   just like ordinary COMPRESSED packets. This packet type is used after
   the initial STATIC packet to set up the decompressor context for the
   first time, and also whenever the header field information cannot be
   encoded in EXTENDED_COMPRESSED packets. DYNAMIC packets could be used
   due to significant field changes or upon INVALID_CONTEXT_FEEBACK.

   All fields except for the initial four bits, the Timestamp Delta, and
   the Header Compression CRC are ordinary IP, UDP and RTP fields. The
   Timestamp Delta is the current delta between RTP timestamps in
   consecutive RTP packets. Initially this value SHOULD be set to 160.

   The packet formats are shown below for IPv6 and IPv4, respectively.
   Note that some fields are only present in some of the DYNAMIC packet

   IPv6 (13-16 octets + CSRC List of 0-60 octets): DYNAMIC1, DYNAMIC2:

          0   1   2   3   4   5   6   7
        :   Context Identifier  (CID)   :  only in DYNAMIC2
        | 1   1   1   0 |  CSRC Counter |
        |         Traffic Class         |
        |           Hop Limit           |
        |                               |
        +         UDP Checksum          +
        |                               |
        | M |       Payload Type        |
        |                               |
        +        Sequence Number        +
        |                               |
        |                               |
        /           Timestamp           /  4 octets
        |                               |
        :                               :
        :           CSRC List           :  0-15 x 4 octets
        :                               :
        |                               |
        +        Timestamp Delta        +
        |                               |
        |    Header Compression CRC     |  see chapter 5.6.2.
        /            Payload            /
   IPv4 (15-18 octets + CSRC List of 0-60 octets): DYNAMIC3, DYNAMIC4,
                                                   DYNAMIC5, DYNAMIC6:

          0   1   2   3   4   5   6   7
        :   Context Identifier  (CID)   :  only in DYNAMIC4 and DYNAMIC6
        | 1   1   1   0 |  CSRC Counter |
        |        Type Of Service        |
        |                               |
        +        Identification         +
        |                               |
        |         Time To Live          |
        :                               :
        +         UDP Checksum          +  only in DYNAMIC5 and DYNAMIC6
        :                               :
        | M |       Payload Type        |
        |                               |
        +        Sequence Number        +
        |                               |
        |                               |
        /           Timestamp           /  4 octets
        |                               |
        :                               :
        :           CSRC List           :  0-15 x 4 octets
        :                               :
        |                               |
        +            TS Delta           +
        |                               |
        |    Header Compression CRC     |  see chapter 5.6.2.
        /            Payload            /

   Each time a DYNAMIC packet is sent, several subsequent packets SHOULD
   also be DYNAMIC packets to ensure a successful update even when
   packets are lost. Context updates both with DYNAMIC and COMPRESSED
   packets could also be acknowledged with CONTEXT_UPDATED_FEEDBACK.

5.4.3.  Compressed packets

   The COMPRESSED packet type is the most commonly used packet and is
   designed to handle ordinary changes as efficiently as possible.

   When changes are regular, all information is carried in the base
   header. When desired, it is possible to send additional information
   in extensions to the COMPRESSED base-header.

   The COMPRESSED base-header formats are shown below. Note that some
   fields are only present in some of the COMPRESSED packet types.

   Defines packet types: COMPRESSED1..COMPRESSED4:

        0   1   2   3   4   5   6   7
      :   Context Identifier  (CID)   : only in COMPRESSED type 2 and 4
      |   Sequence LSP#   |           |    # see chapter 5.5.2
      +---+---+---+---+---+       +---+
      |   Header Compression CRC* | X |    * see chapter 5.6.3
      :                               :
      +        Identification         + only in COMPRESSED type 3 and 4
      :                               :
      :                               :
      /           Extension           / only present if X=1
      :                               :
      /            Payload            /

   Defines packet types: COMPRESSED5..COMPRESSED8:

        0   1   2   3   4   5   6   7
      :   Context Identifier  (CID)   : only in COMPRESSED 6 and 8
      | 0 | Sequence LSB# |    CRC*   |    # see chapter 5.5.1
      +---+---+---+---+---+---+---+---+    * see chapter 5.6.3
      :                               :
      +        Identification         + only in COMPRESSED 7 and 8
      :                               :
      /            Payload            /
   Defines packet types: COMPRESSED9..COMPRESSED12:

        0   1   2   3   4   5   6   7
      :   Context Identifier  (CID)   : only in COMPRESSED 10 and 12
      | 1   0 |          CRC*         |    * see chapter 5.6.3
      | Sequence LSB# |  STS LSB# | X |    # see chapter 5.5.1
      :                               :
      +        Identification         + only in COMPRESSED 11 and 12
      :                               :
      :                               :
      /           Extension           / only present if X=1
      :                               :
      /            Payload            /

   Defines packet types: COMPRESSED13..COMPRESSED16:

        0   1   2   3   4   5   6   7
      :   Context Identifier  (CID)   : only in COMPRESSED 14 and 16
      | 1   0 | M |      STS LSB#     |    # see chapter 5.5.1
      | Sequence LSB# |    CRC*   | X |    * see chapter 5.6.3
      :                               :
      +        Identification         + only in COMPRESSED 15 and 16
      :                               :
      :                               :
      /           Extension           / only present if X=1
      :                               :
      /            Payload            /

   The coverage of the Header Compression CRC is described in chapter
   5.6.3. In that chapter, the CRC polynomials to use are also defined.

   The interpretations of the Sequence and STS (Scaled TimeStamp) fields
   for different profiles are given in section 5.5.

5.4.4.  Extensions to compressed headers

   Less regular changes in the header fields or updates of decompressor
   contexts require an extension in addition to the base header. When
   there is an extension present in the COMPRESSED packet, this is
   indicated by the extension bit (X) being set. Extensions are of
   variable size depending on the information needed to be transmitted.
   However, the first three extension bits are used as an extension Type
   field for all extension formats. The extension can carry an M-bit, a
   t-bit, a SEQ EXT LSB field (called SEQ*), a (S)TS (EXT) LSB field
   (called TS*), an ID LSB field and a bit mask for additional fields.
   The M-bit is the RTP marker bit and the (S)TS (EXT) LSB is timestamp
   information sent with the least significant bits (the most
   significant bits are then expected to be unchanged compared to
   context). The timestamp information could either be the LSB of the
   (S)caled (T)ime(S)tamp value (if indicated with the t-bit unset) or
   the LSB of the absolute timestamp value. For profiles with a
   timestamp field in the compressed base header, the timestamp
   information is sent as an extended range to that field. The SEQ EXT
   LSB is extended range for the RTP sequence number. How extended range
   works is described in chapter 5.5.1 and 5.5.2. The t-bit is sent when
   timestamp is not scaled, otherwise it is always scaled with the
   timestamp delta. The ID LSB is the LSB of the IP Identification
   value. Various bit mask patterns are possible and can consist of
   S,H,C,D,T and I. The interpretations of these bits are given at the
   end of this chapter.

   The guiding principle for choosing the extension type is to find the
   smallest header type that can communicate the information needed.

   For the profiles defined in this document, four different extension
   sets are used, called A, B, C and D. Set A and C are for IPv6 and do
   not handle the IPv4 identification field, which set B and D do. Set A
   and B handle timestamp information which set C and D do not. All
   possible extensions are shown below with indications of which sets
   and types the extensions correspond to. For instance, B3 means that
   in extension set B, the extension is used with type value 3
   (indicated in the type field).

   The defined extension types are shown below:

                   0             7
              - - +-+-+-+-+-+-+-+-+
    A0, B0,       |0 0 0|   SEQ*  |
    C0, D0    - - +-+-+-+-+-+-+-+-+

                   0             7
              - - +-+-+-+-+-+-+-+-+
    A1, B1        |0 0 1|M|  TS*  |
              - - +-+-+-+-+-+-+-+-+
                   0             7 8             5
              - - +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
    A2            |0 1 0|M|          TS*          |
              - - +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+

                                                 1 1             2
                   0             7 8             5 6             3
              - - +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
    A3            |0 1 1|M|t|                 TS*                 |
              - - +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+

                   0             7
              - - +-+-+-+-+-+-+-+-+ - -
    A4            |1 0 0|M|H|D|T|t|
              - - +-+-+-+-+-+-+-+-+ - -

                   0             7 8             5
              - - +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ - -
    A5            |1 0 1|M|C|H|S|D|      TS*      |
              - - +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ - -

                                                 1 1             2
                   0             7 8             5 6             3
              - - +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ - -
    A6            |1 1 0|M|C|H|S|D|t|             TS*             |
              - - +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ - -

                   0             7 8             5
              - - +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ - -
    A7            |1 1 1|M|C|H|S|D|T|t|    SEQ*   |
              - - +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ - -

                   0             7 8             5
              - - +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
    B2            |0 1 0|M|      TS*      |  SEQ* |
              - - +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+

                   0             7 8             5
              - - +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
    B3            |0 1 1|M|  TS*  |     ID LSB    |
              - - +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
                   0             7
              - - +-+-+-+-+-+-+-+-+
    B4, D4        |1 0 0|M| ID LSB|
              - - +-+-+-+-+-+-+-+-+

                   0             7
              - - +-+-+-+-+-+-+-+-+ - -
    B5            |1 0 1|M|H|D|T|I|
              - - +-+-+-+-+-+-+-+-+ - -

                                                 1               2
                   0             7 8             5               3
              - - +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
    B6            |1 1 0|M|t|         TS*         |     ID LSB    |
              - - +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+

                   0             7 8             5
              - - +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ - -
    B7            |1 1 1|M|C|H|S|D|T|I|t|   SEQ*  |
              - - +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ - -

                   0             7 8             5
              - - +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
    C1, D1        |0 0 1|M|         SEQ*          |
              - - +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+

                   0             7 8             5
              - - +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ - -
    C2, D2        |0 1 0|M|C|H|S|       SEQ*      |
              - - +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ - -

                   0             7
              - - +-+-+-+-+-+-+-+-+ - -
    C3, D3        |0 1 1|M|C|H|S|-|
              - - +-+-+-+-+-+-+-+-+ - -

                   0             7 8             5
              - - +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
    D5            |1 0 1|M|        ID LSB         |
              - - +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
                                                 1 1             2
                   0             7 8             5 6             3
              - - +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ - -
    D6            |1 1 0|M|C|H|S|-|              ID               |
              - - +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ - -

   Bit masks indicating additional fields have the following meaning:

   C - Traffic (C)lass / Type Of Service
   H - (H)op Limit / Time To Live
   S - Contributing (S)ources - CSRC
   D - Timestamp (D)elta
   T - (T)imestamp LSB
   I - (I)dentification LSB

   If any of these fields are included, they will appear in the order as
   listed above and the format of the fields will be as described below.

   C - Traffic Class / Type Of Service

       The field contains the value of the original IP header field.

                8 bits
       - - +-+-+-+-+-+-+-+-+ - -
           |   TC / TOS    |
       - - +-+-+-+-+-+-+-+-+ - -

   H - Hop Limit / Time To Live

       The field contains the value of the original IP header field.

                8 bits
       - - +-+-+-+-+-+-+-+-+ - -
           |   HL / TTL    |
       - - +-+-+-+-+-+-+-+-+ - -

   S - Contributing Sources

       The CSRC field is built up of:

       - a counter of the number of CSRC items present (4 bits)
       - an unused field (4 bits)
       - the CSRC items, 1 to 15 (4-60 octets)

                1 octet    +     4 to 60 octets
       - - +-+-+-+-+-+-+-+-+-+-+-+-+-+~~~+-+-+-+-+-+ - -
           | Count | Unused|      CSRC Items       |
       - - +-+-+-+-+-+-+-+-+-+-+-+-+-+~~~+-+-+-+-+-+ - -
   D - Timestamp Delta

       The Timestamp Delta field is a one-octet field. We want to
       communicate Timestamp Delta values corresponding to 80 ms.
       Therefore, the Timestamp Delta value communicated is not the
       actual number of samples, but the number of samples divided by
       8. Thus, only Timestamp Delta values evenly divisible by 8 can
       be communicated in the Timestamp Delta field of an extension. On
       the other hand, the maximum value is 255*8 = 2040 (255 ms at
       8000 Hz). Delta values larger than 2040 or delta values not
       evenly divisible by 8 must be communicated in a DYNAMIC packet.

                 8 bits
       - - +-+-+-+-+-+-+-+-+ - -
           |Timestamp Delta|
       - - +-+-+-+-+-+-+-+-+ - -

       Note that when the Timestamp Delta is changed, Timestamp LSB
       field MUST also be included not downscaled with the delta.

   T - Timestamp LSB

       The field contains the 16 least significant bits of the RTP
       timestamp, scaled if t-bit not set. May be sent as extended
       range for some profiles.

                        16 bits
       - - +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ - -
           |              TS*              |
       - - +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ - -

   I - Identification

       The field contains the IP Identification.

                        16 bits
       - - +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
           |              ID               |
       - - +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+

   When information of any kind is sent in an extension, the
   corresponding information SHOULD also be sent in some subsequent
   packets (either as Extensions or in DYNAMIC packets).

5.4.5.  Feedback packets

   Feedback packets are used by the decompressor to provide various
   types of feedback to the compressor. That could include active
   feedback to assure error free performance or passive feedback (in
   case of invalidated context) to request a context update from the
   compressor. The feedback mechanisms defined here leave a lot to the
   implementation regarding how to use feedback. The general feedback
   packet format is shown below:

                                   0   1   2   3   4   5   6   7
    FEEDBACK (GENERAL)           :   Context Identifier  (CID)   :
                                 | 1   1   1   1   0 |    Type   |

   Note that The CID field is present only for profiles using STATIC
   packet format 2 or 4, which are profiles supporting multiple packet
   streams. The Type field tells what kind of feedback the packet
   corresponds to and the feedback types defined are the following:

                                   0   1   2   3   4   5   6   7
    STATIC_FAILURE_FEEDBACK      :   Context Identifier  (CID)   :
                                 | 1   1   1   1   0 | 0   0   0 |

   The STATIC_FAILURE_FEEDBACK packet tells the compressor that the
   static part of the decompressor context is invalid, and that an
   update of that part is required. Reasons for sending such feedback
   could be that no STATIC packet has been received at all, or that
   decompression has failed even when DYNAMIC packets are decompressed.

                                   0   1   2   3   4   5   6   7
    INVALID_CONTEXT_FEEDBACK     :   Context Identifier  (CID)   :
                                 | 1   1   1   1   0 | 0   0   1 |
                                 |   Last Sequence Number LSB    |

   The INVALID_CONTEXT_FEEDBACK packet SHOULD be sent to signal an
   invalid decompressor context, indicated by failing decompression of
   COMPRESSED packets.

                                   0   1   2   3   4   5   6   7
    NO_PACKETS_FEEDBACK          :   Context Identifier  (CID)   :
                                 | 1   1   1   1   0 | 0   1   0 |
                                 |   Last Sequence Number LSB    |

   The NO_PACKET_FEEDBACK packet can be used by the decompressor to
   signal that packets have not been received for some time. It is not
   always possible for the decompressor to notice such events, and it is
   therefore up to the implementers to decide whether and when to use
   this feedback packet.

                                   0   1   2   3   4   5   6   7
    LONGEST_LOSS_FEEDBACK        :   Context Identifier  (CID)   :
                                 | 1   1   1   1   0 | 0   1   1 |
                                 |   Last Sequence Number LSB    |
                                 |    Length of longest loss     |

   The LONGEST_LOSS_FEEDBACK packet can be used by the decompressor to
   inform the compressor about the length of the longest loss event that
   has occurred on the link between compressor and decompressor. It is
   not always possible for the decompressor to provide this information,
   and it is therefore up to the implementers to decide whether and when
   to use this feedback packet.

                                   0   1   2   3   4   5   6   7
    CONTEXT_UPDATED_FEEDBACK     :   Context Identifier  (CID)   :
                                 | 1   1   1   1   0 | 1   0   0 |
                                 |   Last Sequence Number LSB    |

   The CONTEXT_UPDATED_FEEDBACK packet can be used to signal that an
   update of some header fields has been correctly received, either in a
   DYNAMIC packet or in an EXTENDED_COMPRESSED packet. It is optional to
   use this active feedback mechanism and the compressor MUST NOT expect
   such packets initially. First after reception of one such packet, the
   compressor can expect to get this feedback from the decompressor.

5.5.  Encoding of field values

   The source increases the RTP sequence number by one for each packet
   sent. However, due to losses and reordering before the compression
   point, the changes seen by the compressor may vary. This would
   especially be the case if we consider the scenario in Figure 1.1
   where there are cellular links at both ends of the path. That is one
   reason why sequence number changes need special treatment, but
   another reason is that both timestamps and IP identification for most
   packets can be recreated with a combination of history and sequence
   number knowledge. The profiles defined in this document handle the
   sequence number, timestamp and identification values with LSB
   encoding, except for some profiles that use LSP encoding for the
   sequence number. For timestamp, some profiles also use the principle
   with timer-based decompression. This chapter describes how the
   different encoding methods are applied to the various field values.

5.5.1.  LSB encoding of field values

   LSB encoding is used for sequence number, timestamp and
   identification encoding as described in chapter 5.3.1. The sequence
   numbers, included in all compressed headers, can be sent with
   extended range in extension headers. This is also the case with the
   timestamp value for profiles not using timer-based TS reconstruction
   (see 5.2.1 and 5.5.3). With timer-based timestamp decompression, the
   amount of timestamp LSB that is sent is always limited to the size of
   the field in the compressed header. Note that in most headers, the
   timestamp value is sent as STS LSB (scaled timestamp LSB), which
   means that it is the least significant bits of the timestamp, scaled
   down with the timestamp delta (STS LSB = LSB of [TS / TS Delta]).

5.5.2.  LSP encoding of field values

   LSP, as described in chapter 5.3.2, is used for sequence numbers in
   the "Sequence LSP" field of COMPRESSED1..COMPRESSED4 headers. For
   those headers, there are 28 code-points left for sequence information
   because 4 are reserved for packet type identification. An LSP of size
   28 is therefore used with the following encoding:

     CODE(n) = LSP:28(n)

   The sequence range can be extended with extra bits in extension
   headers, as described in chapter 5.3.3. The "SEQ EXT LSB" field must
   for the case of extended LSP consist of the LSB of the integer

   The reordering parameter for LSP MUST be set to 2 meaning that first
   and second order reordering can be handled by the encoding.

5.5.3.  Timer-based timestamp decompression

   The RTP timestamp field is one of the header fields that may change
   dynamically on a per packet basis. For audio services, the timestamp
   value can be inferred from the encoded RTP sequence number value
   during talk spurts. When the encoded sequence number is incremented
   by N, the timestamp value is incremented by N * Timestamp-Delta-
   value. However, when a talk spurt has faded into silence and a new
   talk spurt starts, the timestamp value will take a leap compared to
   the sequence number. To communicate this leap in the timestamp value,
   some additional action has to be taken. In chapter 5.4.4, extension
   headers are defined that can transfer this leap in the timestamp
   value. That increases, however, the average header size. This chapter
   describes an optional method used by some profiles (see the TbT
   column of table 5.1) to reconstruct the timestamp value, requiring
   only a fixed number of added bits for timestamp leaps. The method
   makes use of timers or a local wall clock at the decompressor.

   To initialize the header compression and the timer-based timestamp
   reconstruction, the absolute value of the timestamp together with the
   sequence number must be transferred from compressor to decompressor
   at the beginning of the compression session. A default timestamp
   delta is also transferred. This is done through the transmission of a
   DYNAMIC header. For speech codecs with 8 kHz sampling frequency and
   20 ms frame sizes, for example, the timestamp delta will be 8000*0.02
   = 160. The decompressor then knows that the timestamp will increase
   by 160 for each packet containing 20 ms of speech. Hence, by using a
   local clock and by measuring packet arrival times, the decompressor
   can estimate the timestamp change compared to the previous packet.
   If, for example, a speech period has been succeeded by a silence
   period at the time T0 and a new speech period starts at the time
   T0+dT, it can be assumed that the timestamp has changed by:

      round(dT/(time for one speech frame)) * (timestamp delta)

   The packet time interval (or codec frame size in time) may be
   determined through the a priori knowledge that most speech codecs
   have constant frame sizes of 10, 20 or 30 ms, or through measurements
   on packet arrival times.

   The decompressor can then get an estimate of the timestamp change,
   add this change to the previous value and replace the least
   significant bits with those received in the compressed header. This
   should give the correct timestamp value.

   It is very important to verify the correctness of a timer-based
   timestamp decompression. However, this is automatically done in ROCCO
   with the header compression CRC verification.

5.6.  Header compression CRCs, coverage and polynomials

   This chapter contains a description of how to calculate the different
   CRCs used by the ROCCO profiles defined in this document.

5.6.1. STATIC packet CRC

   The CRC in the STATIC header is calculated over the whole STATIC
   packet except for the header compression CRC itself. Therefore, the
   header compression CRC field MUST be set to 0 before the CRC

   The CRC polynomial to be used in STATIC packets is:

     C(x) = 1 + x + x^2 + x^8

5.6.2.  DYNAMIC packet CRC

   The CRC in the DYNAMIC packet is calculated over the original
   IP/UDP/RTP header. Before the calculation of the CRC, the IPv4 header
   checksum and the UDP checksum have to be set to 0. This makes it
   possible to recalculate the checksums after the decompression.
   Calculation over the full IP/UDP/RTP headers ensures that the
   decompressed IP/UDP/RTP header is a correct header.

   The CRC polynomial to be used in DYNAMIC packets is:

     C(x) = 1 + x + x^2 + x^8

5.6.3.  COMPRESSED packet CRCs


   The header compression CRC in COMPRESSED header types 1 to 4 is
   calculated over the same headers as the CRC in the DYNAMIC packet,
   except for profiles which use replacement of the UDP checksum, I.e.
   except for profiles 1-4 and 13-16. In profiles 1-4 and 13-16, the
   header compression CRC also covers the payload covered by the UDP

   The polynomial to be used is:

     C(x) = 1 + x + x^4 + x^5 + x^9 + x^10

   In COMPRESSED header types 5 to 8 and 14 to 16 the header compression
   CRC is calculated over the same headers as the CRC in the DYNAMIC
   packet, but with a different polynomial:

     C(x) = 1 + x + x^3


   In COMPRESSED header types 10 to 12 the header compression CRC is
   calculated over the same headers as the CRC in the DYNAMIC packet,
   but with a different polynomial:

     C(x) = 1 + x + x^3 + x^4 + x^6

6.  Implementation issues

   The profiles defined in this document specifies mechanisms for the
   protocol, while much of the usage of these mechanisms is left to the
   implementers to decide upon. This chapter is aimed to give
   guidelines, ideas and suggestions for implementing the scheme.

6.1.  Compressor and decompressor logic, use of feedback

   How to send and respond to the various kinds of FEEDBACK packets is
   not defined in this document, but left to the implementers to decide.
   However, it is recommended to reduce both the number of requests and
   the number of corresponding updating packets to a suitable level.
   Also it is recommended to use COMPRESSED packets with EXTENSIONS
   instead of DYNAMIC packets to update an invalid context, when

   More guidelines on this issue will be included here at a later date.

6.2.  ROCCO over simplex links

   This chapter contains a discussion about how ROCCO can be used over
   simplex links.

   Previous chapters assumed that the decompressor has the possibility
   of sending requests to the compressor. This is true for many systems
   but there are several important scenarios where it is not possible to
   send information from the decompressor back to the compressor. The
   most important case may be when the packet are broadcasted in some
   way. It can also, for example, be the communication from a satellite.
   Over a simplex link the decompressor does not have the possibility of
   sending information back to the compressor. The compressor does not
   know when the decompressor needs a STATIC or DYNAMIC packet. If
   STATIC and DYNAMIC packets are sent at regular intervals it is
   possible for the decompressor to recover a lost context. A slow-start
   mechanism and a periodic refresh guarantee that the decompressor can
   recover a lost synchronization fast.

   The ROCCO scheme is especially suited for simplex links, since it is
   possible for the decompressor to continue with new reconstruction
   attempts, verified with the header checksum, until the next
   STATIC/DYNAMIC packet arrives. It is then possible for the
   decompressor to recover the context before the next STATIC/DYNAMIC
   packet arrives.

   When ROCCO is used over simplex links it is RECOMMENDED that only one
   DYNAMIC packet be sent at a time and not several as stated in
   previous chapters.

6.2.1.  Compression slow-start

   When a field in the STATIC or DYNAMIC packet has changed or if we are
   at the beginning of a ROCCO session, it is necessary to send the
   STATIC/DYNAMIC packet to the decompressor. To ensure that the new
   information reaches the decompressor as fast as possible even if
   packets are lost over the link, a slow-start mechanism is used. After
   the first two packets (STATIC and DYNAMIC), STATIC and DYNAMIC
   packets (read refresh) are sent with an exponentially increasing
   period until a new change occurs. The following figure shows how the
   slow-start works:

   Change in STATIC and/or DYNAMIC packet

   Sent packets:
   . Packet with compressed header
   | STATIC packet followed by a DYNAMIC packet

6.2.2.  Periodic refresh

   To prevent the period between two refreshes from increasing too much,
   an upper limit on the interval between refreshes is set (MAX_PERIOD).
   This is used to avoid losing too many packets if the decompressor has
   lost its context. If the MAX_PERIOD between two refreshes is reached
   a new refresh (STATIC/DYNAMIC) has to be sent.

   To avoid long time periods between two refreshes an upper limit on
   the time between two packets is used (MAX_TIME). This ensures that
   the time between two refreshes does not exceed the MAX_TIME even if
   no MAX_PERIOD packets are sent. If the MAX_TIME between two refreshes
   is reached, a new refresh (STATIC/DYNAMIC) has to be sent.

6.2.3.  Refresh recommendations

   It is recommended that the MAX_PERIOD not exceed 256 packets and that
   the maximum time between two refreshes (MAX_TIME) not exceed 5

6.2.4. Cost and robustness of refreshes

   If we assume that STATIC/DYNAMIC packets are sent every f'th packet,
   the average header size is:

      ----- + C

   S = STATIC header size
   U = DYNAMIC header size
   C = COMPRESSED header size

   The increase in average header size compared with the COMPRESSED
   header size is:


   If we assume that we use ROCCO profile number 8 and that a refresh is
   sent every 256th packet, the increase in average header size is
   (18+15-2)/256=0.12 octets. The average header size for ROCCO profile
   number 4 over duplex links is, with realistic BER, 2.15 octets
   [PERF]. This results in an average header size of 2.15+0.12=2.27

   The difference in robustness of ROCCO between simplex links and
   duplex links is very small. The reason for this is that the ROCCO
   decompressor very seldom loses its context. This results in that
   FEEDBACK packets are almost never needed, as proven in simulations.
   For example: In profile number 8 it is possible to lose up to 24
   consecutive packets without losing the context in the decompressor.
   The probability that 25 consecutive packets are lost is very small
   even if channels with high bit error rates are used. This indicates
   that a ROCCO scheme implemented over simplex links is almost as
   robust as the duplex ROCCO scheme.

6.2.5.  Simplex link improvements

   DYNAMIC information can be sent in two different ways: either as an
   ordinary DYNAMIC packet or as extensions to COMPRESSED headers. If
   the information is sent in extensions to COMPRESSED headers, it is
   possible to reduce the average header size, since a COMPRESSED header
   with extension is smaller than a DYNAMIC header. If COMPRESSED
   headers are used for transmission of DYNAMIC information the
   following is important:

     - All fields in the DYNAMIC packet that have changed since the last
       3-4 refreshes have to be transmitted in the extension.
     - DYNAMIC and STATIC packets still have to be sent at regular
       intervals to ensure that it is possible to recover a lost context
       even if COMPRESSED extension refreshes have failed.

6.3.  Reverse decompression

   This chapter describes an optional decompressor operation to reduce
   discarded packets due to an invalid context.

   Once a context becomes invalid (e.g., in the case when more
   consecutive packet losses than expected has occurred), subsequent
   compressed packets cannot be decompressed correctly with normal
   decompression operation. This decompression operation aims at
   decompressing these packets with a later recovered context. The
   decompressor stores them until the context is validated. After the
   context is updated, the decompressor tries to recover the stored
   packets in the reverse order from the packet updating the context.
   Each time the stored packet is decompressed, its correctness is
   verified using the header compression CRC, which is transmitted in
   each compressed header. Correctly decompressed packets are
   transferred to upper layers in the original order.

   Note that this reverse decompression introduces buffering while
   waiting for the context to be validated and thereby introduces
   additional delay. Thus, it should be used only when some amount of
   delay could be accepted. For example, for video packets belonging to
   the same video frame, the delay of packet arrival time does not cause
   presentation time delay. Delay-insensitive streaming applications can
   also be tolerant to such delay.

   The following illustrates the decompression procedure in some detail:

   1. The decompressor stores compressed packets that cannot be
      decompressed correctly due to an invalid context.

   2. When the decompressor has received a context updating packet and
      the context has been validated, it starts to recover the stored
      packets in reverse order. Decompression is carried out followed
      by the last decompressed packet to its previous packet as if the
      two packets were reordered. After that, decompressor checks the
      correctness of the reconstructed header using the header
      compression CRC.

   3. If the header compression CRC indicates a successful
      decompression, the decompressor stores the complete packet and
      tries to decompress its previous packet. In this way, the stored
      packets are recovered from correctly decompressed packets until
      no compressed packets are left. For each packet, the decompressor
      checks the correctness of the decompressed headers using header
      compression CRC.

   4. If the header compression CRC indicates an incorrectly
      decompressed packet, the reverse decompression attempt must be
      terminated and all remaining packets must be discarded.

   5. Finally, the decompressor forwards all the correctly decompressed
      packets to upper layers in the original order.

6.4.  Pre-verification of CRCs

   For reasons of compression efficiency, it is desirable to keep the
   size of the header compression CRC as small as possible. However, if
   the size of the CRC is decreased, the reliability is also decreased
   and erroneous headers could be generated and passed on from the
   decompressor. It would then be desirable to find a method of
   increasing the strength of the CRC without making it larger.

   There is one property of the ROCCO CRC and its usage that can be used
   to achieve this goal. The CRCs that will occur at the decompressor
   will in most cases follow a pattern well known also to the
   compressor. There are two factors that will affect which CRCs are
   generated and in which order they will occur. If the decompressor
   makes several reconstruction attempts, the first factor affecting the
   CRCs will be the order and properties of the assumptions made for
   each reconstruction attempt. The attempts are in general:

   1:st attempt:    No loss is assumed
   2:nd attempt:    Loss of the preceding packet is assumed
   3:rd attempt:    Loss of the two preceding packets is assumed
   4:th attempt:    Loss of the three preceding packets is assumed

   The other factor that will affect the CRCs generated is what has
   really happened to preceding packets, that is, if no loss has
   occurred or if one or several preceding packets have been lost
   between compressor and decompressor.

   Since the compressor knows how the decompressor performs the
   reconstruction attempts, the compressor can PRE-CALCULATE and VERIFY
   the most probable CRC situations. If a CRC is found not to detect an
   erroneous header, then a different packet type with a larger CRC
   (such as the "normal" COMPRESSED packet) should be used instead or
   additional information could be sent (by using EXTENDED_COMPRESSED or
   DYNAMIC packets). To ensure reliability, the important thing is that
   the CRC must fail if the header is not correctly reconstructed.
   Combining the two factors described above gives a list of the most
   probable CRCs that MUST fail.

   - If ONE packet WAS lost, attempt one (no loss) MUST fail
   - If TWO packets WERE lost, attempt one (no loss) MUST fail
   - If TWO packets WERE lost, attempt two (one lost) MUST fail
   - If THREE packets WERE lost, attempt one (no loss) MUST fail
   - If THREE packets WERE lost, attempt two (one lost) MUST fail
   - If THREE packets WERE lost, attempt three (two lost) MUST fail
   - etc.

   By doing PRE-CALCULATIONS of the six CRCs that would be the result of
   the events listed above, the CRC can be kept strong enough, even with
   a reduced size, because CRCs likely to fail will be avoided.

6.5.  Using "guesses" with LSB and LSP encoding

   ROCCO profiles using LSP encoding can handle 25 consecutive packet
   losses without invalidating the context. LSB or LSP encoding is also
   used for other fields and the range handled is then much larger.
   However, for all LSP or LSB decoding, the range can be extended with
   multiples by making reconstruction attempts (also called "guesses").
   The limiting factors are implementation complexity and time. The
   following example shows how this can be done:

   In chapter 5.3.2, LSP encoding is described. When an LSP encoded
   value for M code-points is decoded to a value S'(n), the original
   header can be reconstructed. If the CRC verification fails, a new
   reconstruction attempt could be made with S'(n)+M as the sequence
   number. If M was a multiple of 2 (LSB encoding), this would be the
   same as changing the value of the lowest MSB bit (i.e. the lowest bit
   NOT transmitted in LSB). More attempts could then be made increasing
   the sequence number by M for each attempt.

7.  Further work

   The ROCCO scheme, including the compression profiles optimized for
   conversational voice defined in this document, has been iterated and
   optimized for almost a year, and most of the desired functionality is
   today supported. However, much work remains before all details are
   settled. In addition to the tuning efforts, there are still new
   issues that should be investigated and implemented. This chapter
   elaborates on some ideas that might be sensible to apply to the

7.1.  Compression of IPv6 extension headers

   The ROCCO compression profiles defined in this document currently do
   not support compression of IPv6 extension headers, which is an
   undesirable limitation. Therefore, it is necessary to investigate
   what is really needed from the compression scheme regarding
   compression of extensions, and also to further develop the current
   and future compression profiles including the desired extension

7.2.  Efficient compression of CSRC lists

   The compression profiles defined in this document do support
   transmission of CSRC items, but this could probably be done in a much
   more efficient manner. Improved solutions for the CSRC compression
   would be preferable because if CSRC lists occur, the headers will be
   significantly expanded due to the size of the CSRC items.

7.3.  General, media independent profiles

   This document defines header compression profiles optimized for
   packet streams carrying conversational voice. Independently of packet
   stream characteristics, these profiles will successfully compress and
   decompress the headers of all IP/UDP/RTP packets. However, the
   compression may not be done in an optimal way. Therefore, general
   profiles should be designed that is optimized to handle
   uncharacterized or intermixed RTP packet streams as efficient as

8.  Implementation status

   The ROCCO algorithm, as defined in previous versions of this Internet
   draft, has been implemented in a testbed environment for realtime IP
   traffic over wireless channels. In the testbed it is possible to
   listen to the effects of header compression in conjunction with
   packet losses. The currently implemented profiles are optimized for
   voice traffic only. A first rough estimate of the CPU utilization
   showed that ROCCO used only slightly more computational power than
   CRTP. On the other hand, with ROCCO the audio quality is
   significantly better. Figure 8.1 shows a block diagram of the testbed

      +---------+   +---------------+   +----------+   +------------+
      | Speech  |-->| RTP/UDP/IP    |-->| Wired IP |-->| Header     |-->
      | Encoder |   | Encapsulation |   | Network  |   | Compressor |
      +---------+   +---------------+   +----------+   +------------+

               +----------+   +--------------+   +---------+
            ~~>| Cellular |~~>| Header De-   |-->| Speech  |
               | Link     |   | compressor   |   | Decoder |
               +----------+   +--------------+   +---------+

             Figure 8.1 : Block diagram of testbed environment

   The implementation has made some impact on the ROCCO protocol,
   realized in this document. Continuously updated information about
   implementation status can be obtained from the ROCCO homepage:
   See also [PERF] for simulated performance results.

9.  Discussion and conclusions

   This document has presented ROCCO, a robust header compression
   protocol framework adaptable to various usage and requirements. In
   addition to the general framework, realizations of the scheme
   optimized for conversational voice packet streams have been presented
   together with performance results for these realizations.

   ROCCO uses CRCs to make local decompressor repairs of the context
   possible. Together with robust encoding methods for header fields,
   the usage of CRCs has made the scheme very robust and capable of
   coping with many consecutive packet losses (up to 25). One other
   important advantage with the CRC approach is that it makes the scheme
   reliable, meaning that it has a very low probability of incorrect
   header reconstruction and error propagation from the decompressor.

   ROCCO defines a concept with header compression profiles, which is an
   abstraction that merges all scheme parameters into one. There are two
   fundamental advantages with the profile concept. First of all, only
   one scheme parameter must be negotiated by the link layer between the
   compression and the decompression points and this requirement does
   not change even if new internal header compression parameters are
   later added with new realizations of the scheme. The second advantage
   is the possibility of optimizing the scheme completely for all
   situations and functionality requirements by using different

   Thanks to the profile concept, it has been possible to compress the
   headers down to a minimal size of 1 octet, the average header size
   being only 1.25 octets. As shown in [PERF], this is achieved without
   introducing any loss of packets due to invalid header compression
   context even over links with bit error rates as high as 1e-2. The
   probability of packet loss due to invalid header compression context
   is practically eliminated thanks to the robustness of ROCCO, even at
   the high BERs (1e-3 to 1e-2) a cellular system may produce.

   With the profiles defined today in this document and in a separate
   document for conversational video [ROVID], ROCCO can efficiently
   compress IP/UDP/RTP packet streams for both conversational voice and
   video. There are profiles defined for both IPv4 and IPv6, support for
   various numbers of concurrent packet streams, enabled or disabled UDP
   checksum, etc. Optimizations have been done to efficiently take care
   of the IPv4 Identification field and make it more compressible if
   knowledge about the sender's assignment policy can be obtained.
   Finally, it is also possible to tune the compression scheme to the
   characteristics of the channel it is used over.

   ROCCO has been evolved and improved during one year. Experiences from
   implementations have been taken into account in the process of
   improving the scheme. However, even if many suggestions for efficient
   implementations are included, there is still room for implementers to
   find even more efficient realizations.

   Hence, ROCCO provides at the present date a powerful toolbox for
   achieving efficient and robust header compression in various types of
   scenarios and over various types of links. ROCCO has also been proven
   to be suitable for cellular environments.

10.  Security considerations

   Because encryption eliminates the redundancy that header compression
   schemes try to exploit, there is some inducement to forego encryption
   in order to achieve operation over low-bandwidth links. However, for
   those cases where encryption of data (and not headers) is sufficient,
   RTP does specify an alternative encryption method in which only the
   RTP payload is encrypted and the headers are left in the clear. That
   would still allow header compression to be applied.

   A malfunctioning or malicious header compressor could cause the
   header decompressor to reconstitute packets that do not match the
   original packets but still have valid IP, UDP and RTP headers and
   possibly even valid UDP checksums. Such corruption may be detected
   with end-to-end authentication and integrity mechanisms which will
   not be affected by the compression. Further, this header compression
   scheme provides an internal checksum for verification of re-
   constructed headers. This reduces the probability of producing
   decompressed headers not matching the original ones without this
   being noticed.

   Denial-of-service attacks are possible if an intruder can introduce
   (for example) bogus STATIC, DYNAMIC or FEEDBACK packets onto the link
   and thereby cause compression efficiency to be reduced. However, an
   intruder having the ability to inject arbitrary packets at the link
   layer in this manner raises additional security issues that dwarf
   those related to the use of header compression.

11.  Acknowledgements

   When designing this protocol, earlier header compression ideas
   described in [CJHC], [IPHC] and [CRTP] have been important sources of

   Thanks to Takeshi Yoshimura at NTT DoCoMo for providing the reverse
   decompression chapter (chapter 6.3). Thanks also to Anton Martensson
   for many valuable draft contributions and to Andreas Jonsson (Lulea
   University), who made a great job supporting this work in his study
   of header field change patterns. Thanks also to all others who have
   given comments.

12.  Intellectual property considerations

   This proposal in is conformity with RFC 2002. 2026.

   Telefonaktiebolaget LM Ericsson and its subsidiaries, in accordance
   with corporate policy, will for submissions rightfully made by its
   employees which are adopted or recommended as a standard by the IETF
   offer patent licensing as follows:

   If part(s) of a submission by Ericsson employees is (are) included in
   a standard and Ericsson has patents and/or patent application(s) that
   are essential to implementation of such included part(s) in said
   standard, Ericsson is prepared to grant - on the basis of reciprocity
   (grant-back) - a license on such included part(s) on reasonable, non-
   discriminatory terms and conditions.

   For the avoidance of doubt this general patent licensing undertaking
   applies to this proposal.

13.  References

   [UDP]    Jon Postel, "User Datagram Protocol", RFC 768, August 1980.

   [IPv4]   Jon Postel, "Internet Protocol", RFC 791, September 1981.

   [IPv6]   Steven Deering, Robert Hinden, "Internet Protocol, Version 6
            (IPv6) Specification", RFC 2460, December 1998.

   [RTP]    Henning Schulzrinne, Stephen Casner, Ron Frederick, Van
            Jacobson, "RTP: A Transport Protocol for Real-Time
            Applications", RFC 1889, January 1996.

   [HDLC]   William Simpson, "PPP in HDLC-like framing", RFC 1662, 1994.

   [VJHC]   Van Jacobson, "Compressing TCP/IP Headers for Low-Speed
            Serial Links", RFC 1144, February 1990.

   [IPHC]   Mikael Degermark, Bjorn Nordgren, Stephen Pink, "IP Header
            Compression", RFC 2507, February 1999.

   [CRTP]   Steven Casner, Van Jacobson, "Compressing IP/UDP/RTP Headers
            for Low-Speed Serial Links", RFC 2508, February 1999.

   [PPPHC]  Mathias Engan, Steven Casner, Carsten Bormann, "IP Header
            Compression over PPP", RFC 2509, February 1999.

   [CRTPC]  Mikael Degermark, Hans Hannu, Lars-Erik Jonsson, Krister
            Svanbro, "CRTP over cellular radio links", Internet Draft
            (work in progress), December 1999.

   [CELL]   Lars Westberg, Morgan Lindqvist, "Realtime traffic over
            cellular access networks", Internet Draft
           (work in progress), May 2000.

   [ROVID] Anton Martensson, Torbjorn Einarsson, Lars-Erik Jonsson,
           "ROCCO Conversational Video Profiles", Internet Draft
           (work in progress), May 2000.

   [PERF]  Hans Hannu, Krister Svanbro, Lars-Erik Jonsson, "ROCCO
           Performance Evaluation", Internet Draft (work in progress),
           May 2000.

   [LLG]   Krister Svanbro, "Lower Layer Guidelines for Robust Header
           Compression", Internet Draft (work in progress), May 2000.
   [WCDMA] "Universal Mobile Telecommunications System (UMTS);
           Selection procedures for the choice of radio transmission
           technologies of the UMTS (UMTS 30.03 version 3.1.0)".
           ETSI TR 101 112 V3.0.1, November 1997.

14.  Authors' addresses

   Lars-Erik Jonsson              Tel: +46 920 20 21 07
   Ericsson Erisoft AB            Fax: +46 920 20 20 99
   Box 920                        Mobile: +46 70 554 82 71
   SE-971 28 Lulea, Sweden        EMail:

   Mikael Degermark               Tel: +46 920 911 88
   Dept of CS & EE                Fax: +46 920 728 01
   Lulea University of Technology Moblie: +46 70 833 89 33
   SE-971 87 Lulea                EMail:

   Hans Hannu                     Tel: +46 920 20 21 84
   Ericsson Erisoft AB            Fax: +46 920 20 20 99
   Box 920                        Mobile: +46 70 559 90 15
   SE-971 28 Lulea, Sweden        EMail:

   Krister Svanbro                Tel: +46 920 20 20 77
   Ericsson Erisoft AB            Fax: +46 920 20 20 99
   Box 920                        Mobile: +46 70 531 25 08
   SE-971 28 Lulea, Sweden        EMail:

Appendix A.  Detailed classification of header fields

   Header compression is possible due to the fact that most header
   fields do not vary randomly from packet to packet. Many of the fields
   exhibit static behavior or changes in a more or less predictable way.
   When designing a header compression scheme, it is of fundamental
   importance to understand the behavior of the fields in detail.

   In this appendix, all IP, UDP and RTP header fields are classified
   and analyzed in two steps. First, we have a general classification in
   A.1 where the fields are classified based on stable knowledge and
   assumptions. The general classification does not take into account
   the change characteristics of changing fields because those will vary
   more or less depending on the implementation and on the application
   used. A less stable but more detailed analysis considering the change
   characteristics is then done in A.2. Finally, A.3 summarizes this
   appendix with conclusions about how the various header fields should
   be handled by the header compression scheme to optimize compression
   and functionality.

A.1.  General classification

   On a general level, the header fields are separated into 5 classes:

   INFERRED       These fields contain values that can be inferred from
                  other values, for example the size of the frame
                  carrying the packet, and thus does not have to be
                  handled at all by the compression scheme.

   STATIC         These fields are expected to be constant throughout
                  the lifetime of the packet stream. Static information
                  must in some way be communicated once.

   STATIC-DEF     STATIC fields whose values define a packet stream.
                  They are in general handled as STATIC.

   STATIC-KNOWN   These STATIC fields are expected to have well-known
                  values and therefore do not need to be communicated
                  at all.

   CHANGING       These fields are expected to vary in some way, either
                  randomly, within a limited value set or range, or in
                  some other manner.

   In this section, each of the IP, UDP and RTP header fields is
   assigned to one of these classes. For all fields except those
   classified as CHANGING, the motives for the classification are also
   stated. CHANGING fields are in A.2 further examined and classified
   based on their expected change behavior.

A.1.1.  IPv6 header fields

    | Field               | Size (bits) |    Class       |
    | Version             |      4      |  STATIC-KNOWN  |
    | Traffic Class       |      8      |    CHANGING    |
    | Flow Label          |     20      |   STATIC-DEF   |
    | Payload Length      |     16      |    INFERRED    |
    | Next Header         |      8      |  STATIC-KNOWN  |
    | Hop Limit           |      8      |    CHANGING    |
    | Source Address      |    128      |   STATIC-DEF   |
    | Destination Address |    128      |   STATIC-DEF   |


     The version field states which IP version the packet is based on.
     Packets with different values in this field must be handled by
     different IP stacks. For header compression, different compression
     profiles must also be used. When compressor and decompressor have
     negotiated which profile to use, the IP version is also known to
     both parties. The field is therefore classified as STATIC-KNOWN.

   Flow Label

     This field may be used to identify packets belonging to a specific
     packet stream. If not used, the value should be set to zero.
     Otherwise, all packets belonging to the same stream must have the
     same value in this field, it being one of the fields defining the
     stream. The field is therefore classified as STATIC-DEF.

   Payload Length

     Information about the packet length (and then also payload length)
     is expected to be provided by the link layer. The field is
     therefore classified as INFERRED.

   Next Header

     This field is expected to have the same value in all packets of a
     packet stream. As for the version number, a certain compression
     profile can only handle a specific next header which means that
     this value is known when profile has been negotiated. The field is
     therefore classified as STATIC-KNOWN.

   Source and Destination addresses

     These fields are part of the definition of a stream and must thus
     be constant for all packets in the stream. The fields are therefore
     classified as STATIC-DEF.

   Summarizing the bits corresponding to the classes gives:

    | Class        | Size (octets)|
    | INFERRED     |       2      |
    | STATIC-DEF   |     34.5     |
    | STATIC-KNOWN |      1.5     |
    | CHANGING     |       2      |

A.1.2.  IPv4 header fields

    | Field               | Size (bits) |     Class      |
    | Version             |      4      |  STATIC-KNOWN  |
    | Header Length       |      4      |  STATIC-KNOWN  |
    | Type Of Service     |      8      |    CHANGING    |
    | Packet Length       |     16      |    INFERRED    |
    | Identification      |     16      |    CHANGING    |
    | Reserved flag       |      1      |  STATIC-KNOWN  |
    | May Fragment flag   |      1      |     STATIC     |
    | Last Fragment flag  |      1      |  STATIC-KNOWN  |
    | Fragment Offset     |     13      |  STATIC-KNOWN  |
    | Time To Live        |      8      |    CHANGING    |
    | Protocol            |      8      |  STATIC-KNOWN  |
    | Header Checksum     |     16      |    INFERRED    |
    | Source Address      |     32      |   STATIC-DEF   |
    | Destination Address |     32      |   STATIC-DEF   |


     The version field states which IP version the packet is based on
     and packets with different values in this field must be handled by
     different IP stacks. For header compression, different compression
     profiles must also be used. When compressor and decompressor has
     negotiated which profile to use, the IP version is also well known
     to both parties. The field is therefore classified as STATIC-KNOWN.

   Header Length

     As long as there are no options present in the IP header, the
     header length is constant and well known. If there are options, the
     fields would be STATIC, but we assume no options. The field is
     therefore classified as STATIC-KNOWN.

   Packet Length

     Information about the packet length is expected to be provided by
     the link layer. The field is therefore classified as INFERRED.


     The Reserved flag must be set to zero and is therefore classified
     as STATIC-KNOWN. The May Fragment flag will be constant for all
     packets in a stream and is therefore classified as STATIC. Finally,
     the Last Fragment bit is expected to be zero because fragmentation
     is NOT expected, due to the small packet size expected. The Last
     Fragment bit is therefore classified as STATIC-KNOWN.

   Fragment Offset

     With the assumption that no fragmentation occurs, the fragment
     offset is always zero. The field is therefore classified as STATIC-


     This field is expected to have the same value in all packets of a
     packet stream. As for the version number, a certain compression
     profile can only handle a specific next header which means that
     this value is well known when profile has been negotiated. The
     field is therefore classified as STATIC-KNOWN.

   Header Checksum

     The header checksum protects individual hops from processing a
     corrupted header. When almost all IP header information is
     compressed away, there is no need to have this additional checksum;
     instead it can be regenerate at the decompressor side. The field is
     therefore classified as INFERRED.

   Source and Destination addresses

     These fields are part of the definition of a stream and must thus
     be constant for all packets in the stream. The fields are therefore
     classified as STATIC-DEF.

   Summarizing the bits corresponding to the classes gives:

    | Class        | Size (octets)|
    | INFERRED     |      4       |
    | STATIC       |    1 bit     |
    | STATIC-DEF   |      8       |
    | STATIC-KNOWN |   3 +7 bits  |
    | CHANGING     |      4       |

A.1.3.  UDP header fields

    | Field            | Size (bits) |    Class    |
    | Source Port      |     16      | STATIC-DEF  |
    | Destination Port |     16      | STATIC-DEF  |
    | Length           |     16      |  INFERRED   |
    | Checksum         |     16      |  CHANGING   |

   Source and Destination ports

     These fields are part of the definition of a stream and must thus
     be constant for all packets in the stream. The fields are therefore
     classified as STATIC-DEF.


     This field is redundant and is therefore classified as INFERRED.

   Summarizing the bits corresponding to the classes gives:

    | Class      | Size (octets)|
    | INFERRED   |       2      |
    | STATIC-DEF |       4      |
    | CHANGING   |       2      |

A.1.4.  RTP header fields

    | Field           | Size (bits) |     Class      |
    | Version         |      2      |  STATIC-KNOWN  |
    | Padding         |      1      |     STATIC     |
    | Extension       |      1      |     STATIC     |
    | CSRC Counter    |      4      |    CHANGING    |
    | Marker          |      1      |    CHANGING    |
    | Payload Type    |      7      |    CHANGING    |
    | Sequence Number |     16      |    CHANGING    |
    | Timestamp       |     32      |    CHANGING    |
    | SSRC            |     32      |   STATIC-DEF   |
    | CSRC            |   0(-480)   |    CHANGING    |


     There exists only one working RTP version and that is version 2.
     The field is therefore classified as STATIC-KNOWN.


     The use of this field depends on the application, but when payload
     padding is used it is likely to be present in all packets. The
     field is therefore classified as STATIC.


     If RTP extensions is used by the application, it is likely to be an
     extension present in all packets (but use of extensions is very
     uncommon). However, for safety's sake this field is classified as


     This field is part of the definition of a stream and must thus be
     constant for all packets in the stream. The field is therefore
     classified as STATIC-DEF.

   Summarizing the bits corresponding to the classes gives:

    | Class        | Size (octets)|
    | STATIC       |    2 bits    |
    | STATIC-DEF   |      4       |
    | STATIC-KNOWN |    2 bits    |
    | CHANGING     |  7.5(-67.5)  |

A.1.5.  Summary for IP/UDP/RTP

   If we summarize this for IP/UDP/RTP we get:

    | Class \ IP ver | IPv6 (octets)| IPv4 (octets)|
    | INFERRED       |       4      |       6      |
    | STATIC         |    2 bits    |    3 bits    |
    | STATIC-DEF     |     42.5     |      16      |
    | STATIC-KNOWN   |   1 +6 bits  |   4 +1 bit   |
    | CHANGING       |  11.5(-71.5) |  13.5(-73.5) |
    | Total          |   60(-120)   |   40(-100)   |

A.2.  Analysis of change patterns of header fields

   To design suitable mechanisms for efficient compression of all header
   fields, their change patterns must be analyzed. For this reason, an
   extended classification is done based on the general classification
   in A.1, considering the fields which were labeled CHANGING in that
   classification. Different applications will use the fields in
   different ways, which may affect their behavior. When this is the
   case, typical behavior for conversational audio and video will be

   The CHANGING fields are separated into five different subclasses:

   STATIC                These are fields that were classified as
                         CHANGING on a general basis, but are classified
                         as STATIC here due to certain additional

   SEMISTATIC            These fields are STATIC most of the time.
                         However, occasionally the value changes but
                         reverts to its original value after a known
                         number of packets.

   RARELY-CHANGING (RC)  These are fields that change their values
                         occasionally and then keep their new values.

   ALTERNATING           These fields alternate between a small number
                         of different values.

   IRREGULAR             These, finally, are the fields for which no
                         useful change pattern can be identified.

   To further expand the classification possibilities without increasing
   complexity, the classification can be done either according to the
   values of the field and/or according to the values of the deltas for
   the field.

   When the classification is done, other details are also stated
   regarding possible additional knowledge about the field values and/or
   field deltas, according to the classification. For fields classified
   as STATIC or SEMISTATIC, the case could be that the value of the
   field is not only STATIC but also well KNOWN a priori (two states for
   SEMISTATIC fields). For fields with non-irregular change behavior, it
   could be known that changes usually are within a LIMITED range
   compared to the maximal change for the field. For other fields, the
   values are completely UNKNOWN.

   Table A.1 classifies all the CHANGING fields based on their expected
   change patterns, especially for conversational audio and video.

    |         Field          | Value/Delta |    Class    |  Knowledge  |
    |             Sequential |    Delta    |    STATIC   |    KNOWN    |
    |             -----------+-------------+-------------+-------------+
    | IPv4 Id:    Seq. jump  |    Delta    |      RC     |   LIMITED   |
    |             -----------+-------------+-------------+-------------+
    |             Random     |    Value    |  IRREGULAR  |   UNKNOWN   |
    | IP TOS / Tr. Class     |    Value    |      RC     |   UNKNOWN   |
    | IP TTL / Hop Limit     |    Value    | ALTERNATING |   LIMITED   |
    |               Disabled |    Value    |    STATIC   |    KNOWN    |
    | UDP Checksum: ---------+-------------+-------------+-------------+
    |               Enabled  |    Value    |  IRREGULAR  |   UNKNOWN   |
    |                 No mix |    Value    |    STATIC   |    KNOWN    |
    | RTP CSRC Count: -------+-------------+-------------+-------------+
    |                 Mixed  |    Value    |      RC     |   LIMITED   |
    | RTP Marker             |    Value    |  SEMISTATIC | KNOWN/KNOWN |
    | RTP Payload Type       |    Value    |      RC     |   UNKNOWN   |
    | RTP Sequence Number    |    Delta    |    STATIC   |    KNOWN    |
    | RTP Timestamp          |    Delta    |      RC     |   LIMITED   |
    |                 No mix |      -      |      -      |      -      |
    | RTP CSRC List:  -------+-------------+-------------+-------------+
    |                 Mixed  |    Value    |      RC     |   UNKNOWN   |

           Table A.1 : Classification of CHANGING header fields

   The following subsections discuss the various header fields in
   detail. Note that table A.1 and the discussions below do not consider
   changes caused by loss or reordering before the compression point.

A.2.1.  IPv4 Identification

   The Identification field (IP ID) of the IPv4 header is there to
   identify which fragments constitute a datagram when reassembling
   fragmented datagrams. The IPv4 specification does not specify exactly
   how this field is to be assigned values, only that each packet should
   get an IP ID that is unique for the source-destination pair and
   protocol for the time the datagram (or any of its fragments) could be
   alive in the network. This means that assignment of IP ID values can
   be done in various ways, which we have separated into three classes.


      This assignment policy keeps a separate counter for each outgoing
      packet stream and thus the IP ID value will increment by one for
      each packet in the stream. Therefore, the delta value of the
      field is constant and well known a priori. When RTP is used on
      top of UDP and IP, the IP ID value follows the RTP sequence
      number. This assignment policy is the most desirable for header
      compression purposes but its usage is not as common as it should
      be. The reason is that it can be realized only if UDP and IP are
      implemented together so that UDP, which separates packet streams
      by the port identification, can make IP use separate ID counters
      for each packet stream.

   Sequential jump

      This is the most common assignment policy in today's IP stacks.
      The difference from the sequential method is that only one
      counter is used for all connections. When the sender is running
      more than one packet stream simultaneously, the IP ID can
      increase by more than one. The IP ID values will be much more
      predictable and require less bits to transfer than random values,
      and the packet-to-packet increment (determined by the number of
      active outgoing packet streams and sending frequencies) will
      usually be limited.


      Some IP stacks assign IP ID values using a pseudo-random number
      generator. There is thus no correlation between the ID values of
      subsequent datagrams. Therefore there is no way to predict the IP
      ID value for the next datagram. For header compression purposes,
      this means that the IP ID field needs to be sent uncompressed
      with each datagram, resulting in two extra octets of header. IP
      stacks in cellular terminals SHOULD NOT use this IP ID assignment

   It should be noted that the ID is an IPv4 mechanism and is therefore
   not needed at all in IPv6 profiles. For IPv4 the ID could be handled
   in three different ways. Firstly, we have the inefficient but
   reliable solution where the ID field is sent as-is in all packets,
   increasing the compressed headers with two octets. This is the best
   way to handle the ID field if the sender uses random assignment of
   the ID field. Secondly, there can be solutions with more flexible
   mechanisms requiring less bits for the ID handling as long as
   sequential jump assignment is used. Such solutions will probably
   require even more bits if random assignment is used by the sender.
   Knowledge about the sender's assignment policy could therefore be
   useful when choosing between the two solutions above. Finally, even
   for IPv4, header compression could be designed without any additional
   information for the ID field included in compressed headers. To use
   such schemes, it must be known that the sender makes use of the pure
   sequential assignment policy for the ID field. That might not be
   possible to know, which implies that the applicability of such
   solutions is very uncertain. However, designers of IPv4 stacks for
   cellular terminals SHOULD use the sequential policy.

A.2.2.  IP Traffic-Class / Type-Of-Service

   The Traffic-Class (IPv6) or Type-Of-Service (IPv4) field is expected
   to be constant during the lifetime of a packet stream or to change
   relatively seldom.

A.2.3.  IP Hop-Limit / Time-To-Live

   The Hop-Limit (IPv6) or Time-To-Live (IPv4) field is expected to be
   constant during the lifetime of a packet stream or to alternate
   between a limited number of values due to route changes.

A.2.4.  UDP Checksum

   The UDP checksum is optional. If disabled, its value is constantly
   zero and could be compressed away. If enabled, its value depends on
   the payload, which for compression purposes is equivalent to it
   changing randomly with every packet.

A.2.5.  RTP CSRC Counter

   This is a counter indicating the number of CSRC items present in the
   CSRC list. This number is expected to be almost constant on a packet-
   to-packet basis and change by small amount. As long as no RTP mixer
   is used, the value of this field is zero.

A.2.6.  RTP Marker

   For audio the marker bit should be set only in the first packet of a
   talkspurt while for video it should be set in the last packet of
   every picture. This means that in both cases the RTP marker is
   classified as SEMISTATIC with well-known values for both states.

A.2.7.  RTP Payload Type

   Changes of the RTP payload type within a packet stream are expected
   to be rare. Applications could adapt to congestion by changing
   payload type and/or frame sizes, but that is not expected to happen

A.2.8.  RTP Sequence Number

   The RTP sequence number will be incremented by one for each packet

A.2.9.  RTP Timestamp

   In the audio case:

      As long as there are no pauses in the audio stream, the RTP
      timestamp will be incremented by a constant delta, corresponding
      to the number of samples in the speech frame. It will thus mostly
      follow the RTP sequence number. When there has been a silent
      period and a new talkspurt begins, the timestamp will jump in
      proportion to the length of the silent period. However, the
      increment will probably be within a relatively limited range.

   In the video case:

      The timestamp change between two consecutive packets will either
      be zero or increase by a multiple of a fixed value corresponding
      to the picture clock frequency. The timestamp can also decrease
      by a multiple of the fixed value if B-pictures are used. The
      delta interval, expressed as a multiple of the picture clock
      frequency, is in most cases very limited.

A.2.10.  RTP Contributing Sources (CSRC)

   The participants in a session, which are identified by the CSRC
   fields, are expected to be almost the same on a packet-to-packet
   basis with relatively few additions or removals. As long as RTP
   mixers are not used, no CSRC fields are present at all.

A.3.  Header compression strategies

   This section elaborates on what has been done in previous sections.
   Based in the classifications, recommendations are given on how to
   handle the various fields in the header compression process. Seven
   different actions are possible and these are listed together with the
   fields to which each action applies.

A.3.1.  Do not send at all

   The fields that have well known values a priori do not have to be
   sent at all. These are:

   - IP Version
   - IPv6 Payload Length
   - IPv6 Next Header
   - IPv4 Header Length
   - IPv4 Reserved Flag
   - IPv4 Last Fragment Flag
   - IPv4 Fragment Offset
   - IPv4 Protocol
   - UDP Checksum (if disabled)
   - RTP Version

A.3.2.  Transmit only initially

   The fields that are constant throughout the lifetime of the packet
   stream have to be transmitted and correctly delivered to the
   decompressor only once. These are:

   - IP Source Address
   - IP Destination Address
   - IPv6 Flow Label
   - IPv4 May Fragment Flag
   - UDP Source Port
   - UDP Destination Port
   - RTP Padding Flag
   - RTP Extension Flag

A.3.3.  Transmit initially, but be prepared to update occasionally

   The fields that are changing only occasionally must be transmitted
   initially but there must also be a way to update these fields with
   new values if they change. These fields are:

   - IPv6 Traffic Class
   - IPv6 Hop Limit
   - IPv4 Type Of Service (TOS)
   - IPv4 Time To Live (TTL)
   - RTP CSRC Counter
   - RTP Payload Type
   - RTP CSRC List

A.3.4.  Be prepared to update or send as-is frequently

   For fields that normally are either constant or whose values can be
   deduced from some other field but frequently diverge from that
   behavior, there must be an efficient way to update the field value or
   send it as-is in some packets. Those fields are:

   - IPv4 Identification (if not sequentially assigned)
   - RTP Marker
   - RTP Timestamp

A.3.5.  Guarantee continuous robustness

   Fields that behave like a counter with a fixed delta for ALL packets,
   the only requirement on the transmission encoding is that packet
   losses between compressor and decompressor must be tolerable. If more
   than one such field exists, all these can be communicated together.
   Such fields can also be used to interpret the values for fields
   listed in the previous section. Fields that have this counter
   behavior are:

   - IPv4 Identification (if sequentially assigned)
   - RTP Sequence Number

A.3.6.  Transmit as-is in all packets

   Fields that have completely random values for each packet must be
   included as-is in all compressed headers. Those fields are:

   - IPv4 Identification (if randomly assigned)
   - UDP Checksum (if enabled)

A.3.7.  Establish and be prepared to update delta

   Finally, there is a field that is usually increasing by a fixed delta
   and is correlated to another field. For this field it would make
   sense to make that delta part of the context state. The delta must
   then be possible to initiate and update in the same way as the fields
   listed in A.3.3. The field to which this applies is:

   - RTP Timestamp
   This Internet-Draft expires November 24, December 15, 2000.