draft-ietf-tsvwg-udp-guidelines-01.txt   draft-ietf-tsvwg-udp-guidelines-02.txt 
Transport Area Working Group L. Eggert Transport Area Working Group L. Eggert
Internet-Draft Nokia Internet-Draft Nokia
Intended status: Best Current G. Fairhurst Intended status: Best Current G. Fairhurst
Practice University of Aberdeen Practice University of Aberdeen
Expires: December 13, 2007 June 11, 2007 Expires: January 10, 2008 July 9, 2007
UDP Usage Guidelines for Application Designers UDP Usage Guidelines for Application Designers
draft-ietf-tsvwg-udp-guidelines-01 draft-ietf-tsvwg-udp-guidelines-02
Status of this Memo Status of this Memo
By submitting this Internet-Draft, each author represents that any By submitting this Internet-Draft, each author represents that any
applicable patent or other IPR claims of which he or she is aware applicable patent or other IPR claims of which he or she is aware
have been or will be disclosed, and any of which he or she becomes have been or will be disclosed, and any of which he or she becomes
aware will be disclosed, in accordance with Section 6 of BCP 79. aware will be disclosed, in accordance with Section 6 of BCP 79.
Internet-Drafts are working documents of the Internet Engineering Internet-Drafts are working documents of the Internet Engineering
Task Force (IETF), its areas, and its working groups. Note that Task Force (IETF), its areas, and its working groups. Note that
skipping to change at page 1, line 35 skipping to change at page 1, line 35
and may be updated, replaced, or obsoleted by other documents at any and may be updated, replaced, or obsoleted by other documents at any
time. It is inappropriate to use Internet-Drafts as reference time. It is inappropriate to use Internet-Drafts as reference
material or to cite them other than as "work in progress." material or to cite them other than as "work in progress."
The list of current Internet-Drafts can be accessed at The list of current Internet-Drafts can be accessed at
http://www.ietf.org/ietf/1id-abstracts.txt. http://www.ietf.org/ietf/1id-abstracts.txt.
The list of Internet-Draft Shadow Directories can be accessed at The list of Internet-Draft Shadow Directories can be accessed at
http://www.ietf.org/shadow.html. http://www.ietf.org/shadow.html.
This Internet-Draft will expire on December 13, 2007. This Internet-Draft will expire on January 10, 2008.
Copyright Notice Copyright Notice
Copyright (C) The IETF Trust (2007). Copyright (C) The IETF Trust (2007).
Abstract Abstract
The User Datagram Protocol (UDP) provides a minimal, message-passing The User Datagram Protocol (UDP) provides a minimal, message-passing
transport that has no inherent congestion control mechanisms. transport that has no inherent congestion control mechanisms.
Because congestion control is critical to the stable operation of the Because congestion control is critical to the stable operation of the
Internet, applications and upper-layer protocols that choose to use Internet, applications and upper-layer protocols that choose to use
UDP as an Internet transport must employ mechanisms to prevent UDP as an Internet transport must employ mechanisms to prevent
congestion collapse and establish some degree of fairness with congestion collapse and establish some degree of fairness with
concurrent traffic. This document provides guidelines on the use of concurrent traffic. This document provides guidelines on the use of
UDP for the designers of such applications and upper-layer protocols UDP for the designers of such applications and upper-layer protocols
that cover congestion-control and other topics, including message that cover congestion-control and other topics, including message
sizes and reliability. sizes, reliability, checksums and middlebox traversal.
Table of Contents Table of Contents
1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . 3 1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . 3
2. Terminology . . . . . . . . . . . . . . . . . . . . . . . . . 4 2. Terminology . . . . . . . . . . . . . . . . . . . . . . . . . 4
3. UDP Usage Guidelines . . . . . . . . . . . . . . . . . . . . . 4 3. UDP Usage Guidelines . . . . . . . . . . . . . . . . . . . . . 4
3.1. Congestion Control Guidelines . . . . . . . . . . . . . . 5 3.1. Congestion Control Guidelines . . . . . . . . . . . . . . 5
3.2. Message Size Guidelines . . . . . . . . . . . . . . . . . 7 3.2. Message Size Guidelines . . . . . . . . . . . . . . . . . 7
3.3. Reliability Guidelines . . . . . . . . . . . . . . . . . . 7 3.3. Reliability Guidelines . . . . . . . . . . . . . . . . . . 8
3.4. Checksum Guidelines . . . . . . . . . . . . . . . . . . . 8 3.4. Checksum Guidelines . . . . . . . . . . . . . . . . . . . 8
3.5. Middlebox Traversal Guidelines . . . . . . . . . . . . . . 9 3.5. Middlebox Traversal Guidelines . . . . . . . . . . . . . . 10
4. Security Considerations . . . . . . . . . . . . . . . . . . . 11 4. Security Considerations . . . . . . . . . . . . . . . . . . . 11
5. IANA Considerations . . . . . . . . . . . . . . . . . . . . . 11 5. IANA Considerations . . . . . . . . . . . . . . . . . . . . . 11
6. Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . 11 6. Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . 11
7. References . . . . . . . . . . . . . . . . . . . . . . . . . . 11 7. References . . . . . . . . . . . . . . . . . . . . . . . . . . 11
7.1. Normative References . . . . . . . . . . . . . . . . . . . 11 7.1. Normative References . . . . . . . . . . . . . . . . . . . 11
7.2. Informative References . . . . . . . . . . . . . . . . . . 12 7.2. Informative References . . . . . . . . . . . . . . . . . . 12
Authors' Addresses . . . . . . . . . . . . . . . . . . . . . . . . 14 Authors' Addresses . . . . . . . . . . . . . . . . . . . . . . . . 14
Intellectual Property and Copyright Statements . . . . . . . . . . 15 Intellectual Property and Copyright Statements . . . . . . . . . . 15
1. Introduction 1. Introduction
The User Datagram Protocol (UDP) [RFC0768] provides a minimal, The User Datagram Protocol (UDP) [RFC0768] provides a minimal,
unreliable, message-passing transport to applications and upper-layer unreliable, best-effort, message-passing transport to applications
protocols (both simply called "applications" in the remainder of this and upper-layer protocols (both simply called "applications" in the
document). Compared to other transport protocols, UDP and its UDP- remainder of this document). Compared to other transport protocols,
Lite variant [RFC3828] are unique in that they do not establish end- UDP and its UDP-Lite variant [RFC3828] are unique in that they do not
to-end connections between communicating end systems. UDP establish end-to-end connections between communicating end systems.
communication consequently does not incur connection establishment UDP communication consequently does not incur connection
and teardown overheads and there is no associated end system state. establishment and teardown overheads and there is no associated end
Because of these characteristics, UDP can offer a very efficient system state. Because of these characteristics, UDP can offer a very
communication transport to some applications. efficient communication transport to some applications.
A second unique characteristic of UDP is that it provides no inherent A second unique characteristic of UDP is that it provides no inherent
congestion control mechanisms. [RFC2914] describes the best current congestion control mechanisms. [RFC2914] describes the best current
practice for congestion control in the Internet. It identifies two practice for congestion control in the Internet. It identifies two
major reasons why congestion control mechanisms are critical for the major reasons why congestion control mechanisms are critical for the
stable operation of the Internet: stable operation of the Internet:
1. The prevention of congestion collapse, i.e., a state where an 1. The prevention of congestion collapse, i.e., a state where an
increase in network load results in a decrease in useful work increase in network load results in a decrease in useful work
done by the network. done by the network.
skipping to change at page 3, line 50 skipping to change at page 3, line 50
application that generates five 1500-byte UDP packets in one second application that generates five 1500-byte UDP packets in one second
can already exceed the capacity of a 56 Kb/s path. For applications can already exceed the capacity of a 56 Kb/s path. For applications
that can operate at higher, potentially unbounded data rates, that can operate at higher, potentially unbounded data rates,
congestion control becomes vital to prevent congestion collapse and congestion control becomes vital to prevent congestion collapse and
establish some degree of fairness. Section 3 describes a number of establish some degree of fairness. Section 3 describes a number of
simple guidelines for the designers of such applications. simple guidelines for the designers of such applications.
A UDP message is carried in a single IP packet and is hence limited A UDP message is carried in a single IP packet and is hence limited
to a maximum payload of 65,487 bytes. The transmission of large IP to a maximum payload of 65,487 bytes. The transmission of large IP
packets frequently requires IP fragmentation, which decreases packets frequently requires IP fragmentation, which decreases
communication reliability and efficiency and should be avoided communication reliability and efficiency and should be avoided. One
[I-D.heffner-frag-harmful]. Some of the guidelines in Section 3 reason for this decrease in reliability is because many NATs and
firewalls do not forward IP fragments; other reasons are documented
in [I-D.heffner-frag-harmful]. Some of the guidelines in Section 3
describe how applications should determine appropriate message sizes. describe how applications should determine appropriate message sizes.
This document provides guidelines to designers of applications that This document provides guidelines to designers of applications that
use UDP for unicast transmission. A special class of applications use UDP for unicast transmission. A special class of applications
uses UDP for IP multicast transmissions. Congestion control, flow uses UDP for IP multicast transmissions. Congestion control, flow
control or reliability for multicast transmissions is more difficult control or reliability for multicast transmissions is more difficult
to establish than for unicast transmissions, because a single sender to establish than for unicast transmissions, because a single sender
may transmit to multiple receivers across potentially very may transmit to multiple receivers across potentially very
heterogeneous paths at the same time. Designing multicast heterogeneous paths at the same time. Designing multicast
applications requires expertise that goes beyond the simple applications requires expertise that goes beyond the simple
skipping to change at page 4, line 45 skipping to change at page 4, line 47
If used correctly, congestion-controlled transport protocols are not If used correctly, congestion-controlled transport protocols are not
as "heavyweight" as often claimed. For example, TCP with SYN cookies as "heavyweight" as often claimed. For example, TCP with SYN cookies
[I-D.ietf-tcpm-syn-flood], which are available on many platforms, [I-D.ietf-tcpm-syn-flood], which are available on many platforms,
does not require a server to maintain per-connection state until the does not require a server to maintain per-connection state until the
connection is established. TCP also requires the end that closes a connection is established. TCP also requires the end that closes a
connection to maintain the TIME-WAIT state that prevents delayed connection to maintain the TIME-WAIT state that prevents delayed
segments from one connection instance to interfere with a later one. segments from one connection instance to interfere with a later one.
Applications that are aware of this behavior can shift maintenance of Applications that are aware of this behavior can shift maintenance of
the TIME-WAIT state to conserve resources. Finally, TCP's built-in the TIME-WAIT state to conserve resources. Finally, TCP's built-in
capacity-probing and PMTU awareness results in efficient data capacity-probing and awareness of the maximum transmission unit
transmission that quickly compensates for the initial connection supported by the path (PMTU) results in efficient data transmission
setup delay, for transfers that exchange more than a few packets. that quickly compensates for the initial connection setup delay, for
transfers that exchange more than a few packets.
3.1. Congestion Control Guidelines 3.1. Congestion Control Guidelines
If an application or upper-layer protocol chooses not to use a If an application or upper-layer protocol chooses not to use a
congestion-controlled transport protocol, it SHOULD control the rate congestion-controlled transport protocol, it SHOULD control the rate
at which it sends UDP messages to a destination host. It is at which it sends UDP messages to a destination host. It is
important to stress that an application SHOULD perform congestion important to stress that an application SHOULD perform congestion
control over all UDP traffic it sends to a destination, independent control over all UDP traffic it sends to a destination, independent
of how it generates this traffic. For example, an application that of how it generates this traffic. For example, an application that
forks multiple worker processes or otherwise uses multiple sockets to forks multiple worker processes or otherwise uses multiple sockets to
skipping to change at page 5, line 28 skipping to change at page 5, line 29
It is important to note that congestion control should not be viewed It is important to note that congestion control should not be viewed
as an add-on to a finished application. Many of the mechanisms as an add-on to a finished application. Many of the mechanisms
discussed in the guidelines below require application support to discussed in the guidelines below require application support to
operate correctly. Application designers need to consider congestion operate correctly. Application designers need to consider congestion
control throughout the design of their application, similar to how control throughout the design of their application, similar to how
they consider security aspects throughout the design process. they consider security aspects throughout the design process.
3.1.1. Bulk Transfer Applications 3.1.1. Bulk Transfer Applications
Applications that perform bulk transmission of data to a peer over Applications that perform bulk transmission of data to a peer over
UDP SHOULD consider implementing TCP-Friendly Rate Control (TFRC) UDP SHOULD implement TCP-Friendly Rate Control (TFRC) [RFC3448],
[RFC3448], window-based, TCP-like congestion control, or otherwise window-based, TCP-like congestion control, or otherwise ensure that
ensure that the application complies with the congestion control the application complies with the congestion control principles.
principles.
TFRC has been designed to provide both congestion control and TFRC has been designed to provide both congestion control and
fairness in a way that is compatible with the IETF's other transport fairness in a way that is compatible with the IETF's other transport
protocols. TFRC is currently being updated protocols. TFRC is currently being updated
[I-D.ietf-dccp-rfc3448bis], and application designers SHOULD always [I-D.ietf-dccp-rfc3448bis], and application designers SHOULD always
evaluate whether the latest published specification fits their needs. evaluate whether the latest published specification fits their needs.
If an application implements TFRC, it need not follow the remaining If an application implements TFRC, it need not follow the remaining
guidelines in Section 3.1, but SHOULD still follow the guidelines on guidelines in Section 3.1, because TFRC already addresses them, but
message sizes in Section 3.2 and reliability in Section 3.2. SHOULD still follow the remaining guidelines in the subsequent
subsections of Section 3.
Bulk transfer applications that choose not to implement TFRC or TCP- Bulk transfer applications that choose not to implement TFRC or TCP-
like windowing SHOULD implement a congestion control scheme that like windowing SHOULD implement a congestion control scheme that
results in bandwidth use that competes fairly with TCP within an results in bandwidth use that competes fairly with TCP within an
order of magnitude. [RFC3551] suggests that applications SHOULD order of magnitude. [RFC3551] suggests that applications SHOULD
monitor the packet loss rate to ensure that it is within acceptable monitor the packet loss rate to ensure that it is within acceptable
parameters. Packet loss is considered acceptable if a TCP flow parameters. Packet loss is considered acceptable if a TCP flow
across the same network path under the same network conditions would across the same network path under the same network conditions would
achieve an average throughput, measured on a reasonable timescale, achieve an average throughput, measured on a reasonable timescale,
that is not less than that of the UDP flow. The comparison to TCP that is not less than that of the UDP flow. The comparison to TCP
skipping to change at page 6, line 15 skipping to change at page 6, line 17
Finally, some bulk transfer applications chose not to implement any Finally, some bulk transfer applications chose not to implement any
congestion control mechanism and instead rely on transmitting across congestion control mechanism and instead rely on transmitting across
reserved path capacity. This might be an acceptable choice for a reserved path capacity. This might be an acceptable choice for a
subset of restricted networking environments, but is by no means a subset of restricted networking environments, but is by no means a
safe practice for operation in the Internet. When the UDP traffic of safe practice for operation in the Internet. When the UDP traffic of
such applications leaks out on unprovisioned paths, results are such applications leaks out on unprovisioned paths, results are
detrimental. detrimental.
3.1.2. Low Data-Volume Applications 3.1.2. Low Data-Volume Applications
When applications that exchange only a small number of messages with
a destination at any time implement TFRC or one of the other
congestion control schemes in Section 3.1.1, the network sees little
benefit, because those mechanisms perform congestion control in a way
that is only effective for longer transmissions.
Applications that exchange only a small number of messages with a Applications that exchange only a small number of messages with a
destination at any time may not benefit from implementing TFRC or one destination at any time applications SHOULD still control their
of the other congestion control schemes in Section 3.1.1. Such transmission behavior by not sending more than one UDP message per
applications SHOULD still control their transmission behavior by not round-trip time (RTT) to a destination. Similar to the
sending more than one UDP message per round-trip time (RTT) to a recommendation in [RFC1536], an application SHOULD maintain an
destination. Similar to the recommendation in [RFC1536], an estimate of the RTT for any destination it communicates with.
application SHOULD maintain an estimate of the RTT for any Applications SHOULD implement the algorithm specified in [RFC2988] to
destination it communicates with. Applications SHOULD implement the compute a smoothed RTT (SRTT) estimate. A lost response from the
algorithm specified in [RFC2988] to compute a smoothed RTT (SRTT) peer SHOULD be treated as a very large RTT sample, instead of being
estimate. A lost response from the peer SHOULD be treated as a very ignored, in order to cause a sufficiently large (exponential) back-
large RTT sample, instead of being ignored, in order to cause a off. When implementing this scheme, applications need to choose a
sufficiently large (exponential) back-off. When implementing this sensible initial value for the RTT. This value SHOULD generally be
scheme, applications need to choose a sensible initial value for the as conservative as possible for the given application. TCP uses an
RTT. This value SHOULD generally be as conservative as possible for initial value of 3 seconds [RFC2988], which is also RECOMMENDED as an
the given application. TCP uses an initial value of 3 seconds initial value for UDP applications. SIP [RFC3261] and GIST
[RFC2988], which is also RECOMMENDED as an initial value for UDP [I-D.ietf-nsis-ntlp] use an initial value of 500 ms, and initial
applications. SIP [RFC3261] and GIST [I-D.ietf-nsis-ntlp] use an timeouts that are shorter than this are likely problematic in many
initial value of 500 ms, and initial timeouts that are shorter than cases. It is also important to note that the initial timeout is not
this are likely problematic in many cases. It is also important to the maximum possible timeout - the RECOMMENDED algorithm in [RFC2988]
note that the initial timeout is not the maximum possible timeout - yields timeout values after a series of losses that are much longer
the RECOMMENDED algorithm in [RFC2988] yields timeout values after a than the initial value.
series of losses that are much longer than the initial value.
Some applications cannot maintain a reliable RTT estimate for a Some applications cannot maintain a reliable RTT estimate for a
destination. The first case is applications that exchange too few destination. The first case is applications that exchange too few
messages with a peer to establish a statistically accurate RTT messages with a peer to establish a statistically accurate RTT
estimate. Such applications MAY use a fixed transmission interval estimate. Such applications MAY use a fixed transmission interval
that is exponentially backed-off during loss. TCP uses an initial that is exponentially backed-off during loss. TCP uses an initial
value of 3 seconds [RFC2988], which is also RECOMMENDED as an initial value of 3 seconds [RFC2988], which is also RECOMMENDED as an initial
value for UDP applications. SIP [RFC3261] and GIST value for UDP applications. SIP [RFC3261] and GIST
[I-D.ietf-nsis-ntlp] use an interval of 500 ms, and shorter values [I-D.ietf-nsis-ntlp] use an interval of 500 ms, and shorter values
are likely problematic in many cases. As in the previous case, note are likely problematic in many cases. As in the previous case, note
skipping to change at page 7, line 32 skipping to change at page 7, line 40
Because IP fragmentation lowers the efficiency and reliability of Because IP fragmentation lowers the efficiency and reliability of
Internet communication [I-D.heffner-frag-harmful], an application Internet communication [I-D.heffner-frag-harmful], an application
SHOULD NOT send UDP messages that result in IP packets that exceed SHOULD NOT send UDP messages that result in IP packets that exceed
the MTU of the path to the destination. Consequently, an application the MTU of the path to the destination. Consequently, an application
SHOULD either use the path MTU information provided by the IP layer SHOULD either use the path MTU information provided by the IP layer
or implement path MTU discovery itself [RFC1191][RFC1981][RFC4821] to or implement path MTU discovery itself [RFC1191][RFC1981][RFC4821] to
determine whether the path to a destination will support its desired determine whether the path to a destination will support its desired
message size without fragmentation. message size without fragmentation.
Applications that choose not to do so SHOULD NOT send UDP messages Applications that choose not adapt the packet size SHOULD NOT send
that exceed the minimum PMTU. The minimum PMTU depends on the IP UDP messages that exceed the minimum PMTU. The minimum PMTU depends
version used for transmission, and is the lesser of 576 bytes and the on the IP version used for transmission, and is the lesser of 576
first-hop MTU for IPv4 [RFC1122] and 1280 bytes for IPv6 [RFC2460]. bytes and the first-hop MTU for IPv4 [RFC1122] and 1280 bytes for
To determine an appropriate UDP payload size, applications must IPv6 [RFC2460]. To determine an appropriate UDP payload size,
subtract IP header and option lengths as well as the length of the applications must subtract IP header and option lengths as well as
UDP header from the PMTU size. Transmission of minimum-sized the length of the UDP header from the PMTU size. Transmission of
messages is inefficient over paths that support a larger PMTU, which minimum-sized messages is inefficient over paths that support a
is a second reason to implement PMTU discovery. larger PMTU, which is a second reason to implement PMTU discovery.
Applications that do not send messages that exceed the minimum PMTU Applications that do not send messages that exceed the minimum PMTU
of IPv4 or IPv6 need not implement any of the above mechanisms. of IPv4 or IPv6 need not implement any of the above mechanisms.
3.3. Reliability Guidelines 3.3. Reliability Guidelines
Application designers are generally aware that UDP does not provide Application designers are generally aware that UDP does not provide
any reliability. Often, this is a main reason to consider UDP as a any reliability. Often, this is a main reason to consider UDP as a
transport. Applications that do require reliable message delivery transport. Applications that do require reliable message delivery
SHOULD implement an appropriate mechanism themselves. SHOULD implement an appropriate mechanism themselves.
UDP also does not protect against message duplication, i.e., an UDP also does not protect against message duplication, i.e., an
application may receive multiple copies of the same message. application may receive multiple copies of the same message.
Application designers SHOULD consider whether their application Application designers SHOULD consider whether their application
handles message duplication gracefully, and may need to implement handles message duplication gracefully, and may need to implement
mechanisms to detect duplicates. Even if message reception triggers mechanisms to detect duplicates. Even if message reception triggers
idempotent operations, applications may want to suppress duplicate idempotent operations, applications may want to suppress duplicate
messages to reduce load. messages to reduce load.
Finally, UDP messages may be reordered in the network and arrive at Finally, UDP messages may be reordered in the network and arrive at
the receiver in an order different from the send order. Applications the receiver in an order different from the transmission order.
that require ordered delivery SHOULD reestablish message ordering Applications that require ordered delivery SHOULD reestablish message
themselves. ordering themselves.
3.4. Checksum Guidelines 3.4. Checksum Guidelines
The UDP header includes a 16-bit ones' complement checksum that The UDP header includes an optional, 16-bit ones' complement checksum
provides an integrity check. The UDP checksum provides assurance that provides an integrity check. The UDP checksum provides
that the payload was not corrupted in transit. It also verifies that assurance that the payload was not corrupted in transit. It also
the datagram was delivered to the intended end point, because it verifies that the datagram was delivered to the intended end point,
covers the IP addresses, port numbers and protocol number, and it because it covers the IP addresses, port numbers and protocol number,
verifies that the datagram is not truncated or padded, because it and it verifies that the datagram is not truncated or padded, because
covers the size field. It therefore protects an application against it covers the size field. It therefore protects an application
receiving corrupted payload data in place of, or in addition to, the against receiving corrupted payload data in place of, or in addition
data that was sent. to, the data that was sent.
Applications SHOULD enable UDP checksums, although [RFC0793] permits Applications SHOULD enable UDP checksums, although [RFC0793] permits
the option to disable their use. Applications that choose to disable the option to disable their use. Applications that choose to disable
UDP checksums when transmitting over IPv4 therefore MUST NOT make UDP checksums when transmitting over IPv4 therefore MUST NOT make
assumptions regarding the correctness of received data and MUST assumptions regarding the correctness of received data and MUST
behave correctly when a packet is received that was originally sent behave correctly when a packet is received that was originally sent
to a different end point or is otherwise corrupted. The use of the to a different end point or is otherwise corrupted. The use of the
UDP checksum is MANDATORY when applications transmit UDP over IPv6 UDP checksum is MANDATORY when applications transmit UDP over IPv6
[RFC2460] and applications consequently MUST NOT disable their use. [RFC2460] and applications consequently MUST NOT disable their use.
(The IPv6 header does not have a separate checksum field to protect (The IPv6 header does not have a separate checksum field to protect
the IP addressing information.) the IP addressing information.)
The UDP checksum provides relatively weak protection from a coding The UDP checksum provides relatively weak protection from a coding
point of view [RFC3819] and, where appropriate, applications point of view [RFC3819] and, where data integrity is important,
developers SHOULD provide additional checks, e.g., through a CRC application developers SHOULD provide additional checks, e.g.,
included with the data to verify the integrity of an entire object/ through a CRC included with the data to verify the integrity of an
file sent over UDP service. entire object/file sent over UDP service.
3.4.1. UDP-Lite 3.4.1. UDP-Lite
A special class of applications derive benefit from having partially A special class of applications derive benefit from having partially
damaged payloads delivered rather than discarded when using paths damaged payloads delivered rather than discarded when using paths
that include error-prone links. Such applications can tolerate that include error-prone links. Such applications can tolerate
payload corruption and MAY choose to use the Lightweight User payload corruption and MAY choose to use the Lightweight User
Datagram Protocol (UDP-Lite) [RFC3828] variant of UDP instead of Datagram Protocol (UDP-Lite) [RFC3828] variant of UDP instead of
basic UDP. Applications that choose to use UDP-Lite instead of UDP basic UDP. Applications that choose to use UDP-Lite instead of UDP
MUST still follow the congestion control and other guidelines MUST still follow the congestion control and other guidelines
skipping to change at page 9, line 51 skipping to change at page 10, line 9
and/or encryption are used, this can result in UDP-Lite packets being and/or encryption are used, this can result in UDP-Lite packets being
treated the same as UDP packets, i.e., result in packet loss. Use of treated the same as UDP packets, i.e., result in packet loss. Use of
IP fragmentation can also prevent special treatment for UDP-Lite IP fragmentation can also prevent special treatment for UDP-Lite
packets, and is another reason why applications SHOULD avoid IP packets, and is another reason why applications SHOULD avoid IP
fragmentation Section 3.2. fragmentation Section 3.2.
3.5. Middlebox Traversal Guidelines 3.5. Middlebox Traversal Guidelines
Network address translators (NATs) and firewalls are examples of Network address translators (NATs) and firewalls are examples of
intermediary devices ("middleboxes") that can exist along an end-to- intermediary devices ("middleboxes") that can exist along an end-to-
end path. A middlebox typically perform a function that requires it end path. A middlebox typically performs a function that requires it
to maintain per-flow state. For connection-oriented protocols, such to maintain per-flow state. For connection-oriented protocols, such
as TCP, middleboxes snoop and parse the connection-management traffic as TCP, middleboxes snoop and parse the connection-management traffic
and create and destroy per-flow state accordingly. For a and create and destroy per-flow state accordingly. For a
connectionless protocol such as UDP, this approach is not possible. connectionless protocol such as UDP, this approach is not possible.
Consequently, middleboxes may create per-flow state when they see a Consequently, middleboxes may create per-flow state when they see a
packet that indicates a new flow, and destroy the state after some packet that indicates a new flow, and destroy the state after some
period of time during which no packets belonging to the same flow period of time during which no packets belonging to the same flow
have arrived. have arrived.
Depending on the specific function that the middlebox performs, this Depending on the specific function that the middlebox performs, this
skipping to change at page 10, line 44 skipping to change at page 10, line 50
communication failures and implement mechanisms to re-establish their communication failures and implement mechanisms to re-establish their
UDP sessions. UDP sessions.
Applications MAY in addition send periodic keep-alive messages to Applications MAY in addition send periodic keep-alive messages to
attempt to refresh middlebox state. Unfortunately, no common timeout attempt to refresh middlebox state. Unfortunately, no common timeout
has been specified for per-flow UDP state for arbitrary middleboxes. has been specified for per-flow UDP state for arbitrary middleboxes.
For NATs, [RFC4787] requires a state timeout of 2 minutes or longer, For NATs, [RFC4787] requires a state timeout of 2 minutes or longer,
and it is likely that other types of middleboxes use timeouts of and it is likely that other types of middleboxes use timeouts of
similar timescales. Consequently, if applications choose to send similar timescales. Consequently, if applications choose to send
periodic keep-alives, they SHOULD NOT send them more frequently than periodic keep-alives, they SHOULD NOT send them more frequently than
once every two minutes. once every two minutes. (Not that some deployed middleboxes use a
shorter timeout value than 2 minutes, violating [RFC4787].)
It is important to note that sending keep-alives is not a substitute It is important to note that sending keep-alives is not a substitute
for implementing a robust connection handling. Like all UDP for implementing a robust connection handling. Like all UDP
messages, keep-alives can be delayed or dropped, causing middlebox messages, keep-alives can be delayed or dropped, causing middlebox
state to time out. In addition, the congestion control guidelines in state to time out. In addition, the congestion control guidelines in
Section 3.1 cover all UDP transmissions by an application, including Section 3.1 cover all UDP transmissions by an application, including
the transmission of middlebox keep-alives. Congestion control may the transmission of middlebox keep-alives. Congestion control may
thus lead to delays or temporary suspension of keep-alive thus lead to delays or temporary suspension of keep-alive
transmission. transmission.
4. Security Considerations 4. Security Considerations
skipping to change at page 11, line 21 skipping to change at page 11, line 27
guidelines to designers of UDP-based applications to congestion- guidelines to designers of UDP-based applications to congestion-
control their transmissions. As such, it does not raise any control their transmissions. As such, it does not raise any
additional security concerns. additional security concerns.
5. IANA Considerations 5. IANA Considerations
This document raises no IANA considerations. This document raises no IANA considerations.
6. Acknowledgments 6. Acknowledgments
Thanks to Mark Allman, Sally Floyd, Joerg Ott, Colin Perkins, Pasi Thanks to Mark Allman, Sally Floyd, Philip Matthews, Joerg Ott, Colin
Sarolahti and Magnus Westerlund for their comments on this document. Perkins, Pasi Sarolahti and Magnus Westerlund for their comments on
this document.
The middlebox traversal guidelines in Section 3.5 incorporate ideas The middlebox traversal guidelines in Section 3.5 incorporate ideas
from Section 5 of [I-D.ford-behave-app] by Bryan Ford, Pyda Srisuresh from Section 5 of [I-D.ford-behave-app] by Bryan Ford, Pyda Srisuresh
and Dan Kegel. and Dan Kegel.
7. References 7. References
7.1. Normative References 7.1. Normative References
[RFC0768] Postel, J., "User Datagram Protocol", STD 6, RFC 768, [RFC0768] Postel, J., "User Datagram Protocol", STD 6, RFC 768,
skipping to change at page 12, line 39 skipping to change at page 12, line 47
[RFC4821] Mathis, M. and J. Heffner, "Packetization Layer Path MTU [RFC4821] Mathis, M. and J. Heffner, "Packetization Layer Path MTU
Discovery", RFC 4821, March 2007. Discovery", RFC 4821, March 2007.
7.2. Informative References 7.2. Informative References
[I-D.floyd-dccp-ccid4] [I-D.floyd-dccp-ccid4]
Floyd, S. and E. Kohler, "Profile for Datagram Congestion Floyd, S. and E. Kohler, "Profile for Datagram Congestion
Control Protocol (DCCP) Congestion ID 4: TCP-Friendly Control Protocol (DCCP) Congestion ID 4: TCP-Friendly
Rate Control for Small Packets (TFRC-SP)", Rate Control for Small Packets (TFRC-SP)",
draft-floyd-dccp-ccid4-00 (work in progress), draft-floyd-dccp-ccid4-01 (work in progress), July 2007.
November 2006.
[I-D.ford-behave-app] [I-D.ford-behave-app]
Ford, B., "Application Design Guidelines for Traversal Ford, B., "Application Design Guidelines for Traversal
through Network Address Translators", through Network Address Translators",
draft-ford-behave-app-05 (work in progress), March 2007. draft-ford-behave-app-05 (work in progress), March 2007.
[I-D.heffner-frag-harmful] [I-D.heffner-frag-harmful]
Heffner, J., "IPv4 Reassembly Errors at High Data Rates", Heffner, J., "IPv4 Reassembly Errors at High Data Rates",
draft-heffner-frag-harmful-05 (work in progress), draft-heffner-frag-harmful-05 (work in progress),
May 2007. May 2007.
skipping to change at page 13, line 17 skipping to change at page 13, line 22
Specification", draft-ietf-dccp-rfc3448bis-01 (work in Specification", draft-ietf-dccp-rfc3448bis-01 (work in
progress), March 2007. progress), March 2007.
[I-D.ietf-nsis-ntlp] [I-D.ietf-nsis-ntlp]
Schulzrinne, H. and R. Hancock, "GIST: General Internet Schulzrinne, H. and R. Hancock, "GIST: General Internet
Signalling Transport", draft-ietf-nsis-ntlp-13 (work in Signalling Transport", draft-ietf-nsis-ntlp-13 (work in
progress), April 2007. progress), April 2007.
[I-D.ietf-tcpm-syn-flood] [I-D.ietf-tcpm-syn-flood]
Eddy, W., "TCP SYN Flooding Attacks and Common Eddy, W., "TCP SYN Flooding Attacks and Common
Mitigations", draft-ietf-tcpm-syn-flood-04 (work in Mitigations", draft-ietf-tcpm-syn-flood-05 (work in
progress), May 2007. progress), May 2007.
[RFC1122] Braden, R., "Requirements for Internet Hosts - [RFC1122] Braden, R., "Requirements for Internet Hosts -
Communication Layers", STD 3, RFC 1122, October 1989. Communication Layers", STD 3, RFC 1122, October 1989.
[RFC1536] Kumar, A., Postel, J., Neuman, C., Danzig, P., and S. [RFC1536] Kumar, A., Postel, J., Neuman, C., Danzig, P., and S.
Miller, "Common DNS Implementation Errors and Suggested Miller, "Common DNS Implementation Errors and Suggested
Fixes", RFC 1536, October 1993. Fixes", RFC 1536, October 1993.
[RFC2309] Braden, B., Clark, D., Crowcroft, J., Davie, B., Deering, [RFC2309] Braden, B., Clark, D., Crowcroft, J., Davie, B., Deering,
 End of changes. 22 change blocks. 
80 lines changed or deleted 88 lines changed or added

This html diff was produced by rfcdiff 1.33. The latest version is available from http://tools.ietf.org/tools/rfcdiff/