[Docs] [txt|pdf] [Tracker] [Email] [Diff1] [Diff2] [Nits]

Versions: 00 01 02 03 draft-ietf-rserpool-service

Network Working Group                                          P. Conrad
Internet-Draft                                         Temple University
Expires: May 2, 2003                                              P. Lei
                                                     Cisco Systems, Inc.
                                                        November 1, 2002

              Services Provided By Reliable Server Pooling

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-

   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 to cite them other than as "work in progress."

   The list of current Internet-Drafts can be accessed at http://

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

   This Internet-Draft will expire on May 2, 2003.

Copyright Notice

   Copyright (C) The Internet Society (2002).  All Rights Reserved.


   RSerPool [1] is a framework to provide highly available services
   between clients and servers.  This is achieved by grouping servers
   into pools, each with an identifier and pooling policy.  Three
   classes of entities are defined: Pool Users (clients), Pool Elements
   (servers), and Name Servers.

   This memo defines the services provided by this framework to upper
   layer protocols and applications for Pool Users and Pool Elements.
   It describes the service primitives that the framework provides and
   describes example scenarios.

Conrad & Lei               Expires May 2, 2003                  [Page 1]

Internet-Draft        Services Provided by RSerPool        November 2002

   It also describes the requirements for mapping (or adaption or
   "shim") layers for a variety of transport protocols (SCTP, TCP, or
   others) such that upper layer protocols and applications may use a
   common framework/API to utilize the services provided.

Table of Contents

   1.    Introduction . . . . . . . . . . . . . . . . . . . . . . . .  3
   2.    Conventions Used In This Document  . . . . . . . . . . . . .  4
   3.    Example Application Scenarios  . . . . . . . . . . . . . . .  4
   3.1   Example Scenario for Failover Without RSerPool . . . . . . .  4
   3.2   Example Scenario Using RSerPool Name Services Only . . . . .  5
   3.3   Example Scenario Using Full RSerPool Services  . . . . . . .  7
   4.    Service Primitives . . . . . . . . . . . . . . . . . . . . .  8
   4.1   Initialization . . . . . . . . . . . . . . . . . . . . . . .  8
   4.2   PE Registration Services . . . . . . . . . . . . . . . . . .  8
   4.3   Failover Callback Function . . . . . . . . . . . . . . . . .  9
   4.4   PE Selection Services  . . . . . . . . . . . . . . . . . . . 10
   4.5   Upper Layer/Application Level Acknowledgements . . . . . . . 11
   4.6   RSerPool Managed Data Channel  . . . . . . . . . . . . . . . 11
   5.    Transport Mappings . . . . . . . . . . . . . . . . . . . . . 12
   5.1   Defined Transport Mappings . . . . . . . . . . . . . . . . . 12
   5.2   Transport Mappings Requirements  . . . . . . . . . . . . . . 13
   5.2.1 Mappings: Mandatory Requirements . . . . . . . . . . . . . . 13
   5.2.2 Mappings: Optional Requirements  . . . . . . . . . . . . . . 13
   5.2.3 Mappings: Other Requirements . . . . . . . . . . . . . . . . 14
   6.    Security Considerations  . . . . . . . . . . . . . . . . . . 14
   7.    IANA Considerations  . . . . . . . . . . . . . . . . . . . . 14
   8.    Acknowledgements . . . . . . . . . . . . . . . . . . . . . . 14
         References . . . . . . . . . . . . . . . . . . . . . . . . . 14
         Authors' Addresses . . . . . . . . . . . . . . . . . . . . . 15
         Full Copyright Statement . . . . . . . . . . . . . . . . . . 16

Conrad & Lei               Expires May 2, 2003                  [Page 2]

Internet-Draft        Services Provided by RSerPool        November 2002

1. Introduction

   The Reliable Server Pooling architecture is defined in [1].  The
   architecture provides highly available services by defining three
   classes of entities: pool users (clients), pool elements (servers),
   and name servers.  Pool elements are grouped into server pools and
   can be used by pool users via its pool name (or "handle") and can be
   selected by following the pool's pool element selection policy.

   This memo describes how an upper layer protocol or application for a
   pool user or pool element uses this architecture and associated
   protocols to achieve these goals described in that document.
   Specifically, it describes how the ASAP protocol [5] and transport
   protocols (SCTP, TCP, etc.) can be utilized to realize highly
   available services between pool users and pool elements.

   There are tradeoffs between the amount of application modification
   required, the features and restrictions that the underlying transport
   is required to support, and the richness of the feature set provided
   by RSerPool.  In order to provide support for both existing/legacy
   (non-RSerPool) and new applications, several service primitives are
   defined in which an upper layer protocol can interact with the
   RSerPool framework.  Depending on the number of services utilized,
   the upper layer protocol achieve a range of reliability from simple
   pool element selection to a fully automatic failover capability.

   Utilizing a limited set of RSerPool services provides the capability
   for legacy upper layer software to use a few RSerPool services with
   relatively minor modifications, and allows a broad range of
   underlying transport protocols to be supported.  To achieve a richer
   and more complete failover model, however, a majority of the RSerPool
   services should be used, which places certain requirements and
   restrictions on the transport layers that can be supported.

   Note that regardless of the number of service primitives actually
   utilized by any given upper layer protocol, this document assumes
   that the upper layer protocol/application is operating on a platform
   that has a full running, implmentation of ASAP.

   The following figure illustrates the protocol stacks when using the
   RSerPool framework (Pool Element perspective shown).  Note that the
   mapping layer MAY be a "NULL" layer, if no control channel is
   utilized and/or the data channel is not utilized (e.g.  application
   specific data).

Conrad & Lei               Expires May 2, 2003                  [Page 3]

Internet-Draft        Services Provided by RSerPool        November 2002

     |  Application/Upper Layer Protocol    |
     |            RSerPool API              |
     |(control channel) |  (data channel)   |
     |         ASAP layer        |          |
     +------------------+--------+          |
     |        mapping/adaption layer        |
     |          transport protocol          |

   The purpose of this document is to describe:

   1.  the precise services provided by RSerPool to the upper layer,

   2.  the tradeoffs in choosing which services to utilize,

   3.  how applications must be designed for each of these services,

   4.  how applications written over various transports (SCTP, TCP, and
       others) can be mapped into these services.

2. Conventions Used In This Document

   The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT",
   document are to be interpreted as described in RFC 2119 [2].

3. Example Application Scenarios

   To illustrate the differences among an application without RSerPool,
   an application using limited RserPool services, and an application
   using a full suite of RSerPool services, this section provides an
   informal description of how failover may be handled in each of these

3.1 Example Scenario for Failover Without RSerPool

   Consider a typical client/server application that does not use a
   reliable server pooling framework of any kind.  Typically, the server
   is specified by a DNS name.  At some point, the application
   translates this name to an IP address (via DNS), and subsequently
   makes initial contact with the server to begin a session, via SCTP,
   TCP, UDP, or some other protocol.  If the client loses contact or
   fails to make contact with the server (either due to server failure,

Conrad & Lei               Expires May 2, 2003                  [Page 4]

Internet-Draft        Services Provided by RSerPool        November 2002

   or a failure in the network) the client must either abandon the
   session, or try to contact another server.

   In this scenario, the client must first determine that a failure took
   place.  There are several ways that a client application may
   determine that a server failed, including the following:

   1.  The client may have sent a request to the server, and may time
       out waiting for a response, or may receive a message such as "no
       route to host", "port not available", or "connection refused".

   2.  The client may have sent a request to the server, or may have
       tried to initiate a connection or association and may have
       received a connection/association failure error.

   3.  The client may already have established a connection to server,
       but at some point receives an indication from the transport layer
       that the connection failed.

   Suppose that the client application has a feature by which the user
   can enter the hostname of a secondary server to contact in the event
   of failure.  Once the application determines that a failure took
   place on the primary server, the application can then attempt to
   resolve the hostname of the secondary server, and contact the
   secondary server to establish a session there.  This process can be
   iterated to a tertiary server, and so forth.

   A limitation of this model is that there is no provision (other than
   static client configuration, plus the capabilities of DNS) to
   determine which server to contact initially (other than DNS) or which
   server to contact next in the event of server failure.  (See [3] for
   a discussion of the limitations of using DNS for this purpose.)

3.2 Example Scenario Using RSerPool Name Services Only

   Now consider the same client/server application mentioned in Section
   3.1.  First we describe what the application programmer must do to
   modify the code to use RSerPool name services.  We then describe the
   benefits that these modifications provide.

   For pool user ("client") applications, there are typically only three
   modifications are required along with adding ASAP:

   1.  Instead of specifying the hostnames of primary, secondary,
       tertiary servers, etc., the application user specifies a pool
       handle (or pool name).

   2.  Instead of using a DNS based service (e.g.  the Unix library

Conrad & Lei               Expires May 2, 2003                  [Page 5]

Internet-Draft        Services Provided by RSerPool        November 2002

       function gethostbyname()) to translate from a hostname to an IP
       address, the application will invoke an RSerPool service
       primitive "GetPrimaryServer" that takes as input a pool handle,
       and returns the IP address of the primary server.  The
       application then uses that IP address just as it would have used
       the IP address returned by the DNS in the previous scenario.

   3.  Without the use of additional RSerPool services, failure
       detection is application specific just as in the previous
       scenario.  However, when failure is detected on the primary
       server, instead of invoking DNS translation again on the hostname
       of a secondary server, the application invokes the service
       primitive "GetNextServer", which has a dual meaning.  First it
       indicates to the RSerPool layer the failure of the server
       returned by a previous "GetPrimaryServer" or "GetNextServer"
       call.  Second, it provides the IP address of the next server that
       should be contacted, according to the best information available
       to the RSerPool layer at the present time (e.g.  set of available
       pool elements, pool element policy in effect for the pool, etc.).

   For pool element ("server") applications, two additions in are
   required along with adding ASAP:

   1.  The server should invoke the REGISTER service primitive upon
       startup to add itself into the server pool using an appropriate
       pool handle.  This also includes the address(es) protocol or
       mapping id, port (if required by the mapping), and pooling

   2.  The server should invoke the DEREGISTER service primitive to
       remove itself from the server pool when shutting down.

   When using these RSerPool services, RSerPool provides benefits that
   are limited (as compared to utilizing all services, described in
   Section 3.3), but nevertheless quite useful as compared to not using
   RSerPool at all (as in Section 3.1).  First, the client user need
   only supply a single string, i.e.  the pool handle, rather than a
   list of servers.  Second, the decision as to which server is to be
   used can be determined dynamically by the server selection mechanism
   (i.e.  a "pool policy" performed by ASAP; see [1]).  Finally, when
   failures occur, these are reported to the pool via signaling present
   in ASAP [5]) and ENRP [4], other clients will eventually know (once
   this failure is confirmed by other elements of the RSerPool
   architecture) that this server has failed.

   Utilizing this subset of services is useful for applications built
   over connectionless protocols such as UDP that cannot easily be
   adapted to the transport layer requirements required for full

Conrad & Lei               Expires May 2, 2003                  [Page 6]

Internet-Draft        Services Provided by RSerPool        November 2002

   failover services (see section Section 5) or for an expedient way to
   provide some of the benefits of RSerPool to legacy applications
   (regardless of the transport protocol used).  However, to take full
   advantage of the RSerPool framework, utilization of the full suite of
   services as described in the next section is recommended.

3.3 Example Scenario Using Full RSerPool Services

   Finally, consider the same client/server application as in Section
   3.1, but this time, modified to all RSerPool provided services.  As
   in the Section 3.1, we first describe the modifications needed, then
   we describe the benefits provided.

   When the full suite of RSerPool services are used, all communication
   between the pool user and the pool element is mediated by the
   RSerPool framework, including not only session establishment and
   teardown, but also the sending and receiving of data.  Accordingly,
   it is necessary to modify the application to use the service
   primitives (i.e.  the API) provided by RSerPool, rather than the
   transport layer primitives provided by TCP, SCTP, or whatever
   transport protocol is being used.

   As in the previous case, sessions (rather than connections or
   associations) are established, and the destination endpoint is
   specified as a pool handle rather than as a list of IP addresses with
   a port number.  However, failover from one pool element to another is
   fully automatic, and can be transparent to the application:

      The RSerPool framework control channel provides maintainance
      functions to keep pool element lists, policies, etc.  current.

      Since the application data (e.g.  data channel) is managed by the
      RSerPool framework, any unsent and unacknowledged transport data
      can be automatically re-sent to the newly selected pool element
      upon failover.  This is enhanced by providing the application an
      "upper layer acknowledegment" service.

      The application can provide a callback function (described in
      Section 4.3) that is invoked in the case of a failover.  This
      callback function can execute any application specific failover
      code, such as generating a special message (or sequence of
      messages) that helps the new pool element construct any state
      needed to continue an in-process session.

   Retrofitting an existing application to this mode of RSerPool
   requires more effort on the part of the application programmer than
   retrofitting an application to use just the pool selection services;
   all use of the transport layer's primitives (e.g.  the calls to the

Conrad & Lei               Expires May 2, 2003                  [Page 7]

Internet-Draft        Services Provided by RSerPool        November 2002

   sockets API) must be modified to use the RSerPool primitives (e.g.
   the RSerPool API).  This can be mitigated by making the API for
   RSerPool as close to existing transport APIs as possible.  However,
   failure detection and failover is automated in this case.

   Furthermore, since the primitives provided by RSerPool are similar to
   those of existing transport protocols (and, it is hoped, the APIs
   will be also) for developers of new applications, writing to the
   RSerPool failover mode primitives is not significantly different in
   terms of programmer effort or learning curve than writing the same
   applications over existing transport layer primitives.

4. Service Primitives

   Upper layer protocols and applications may "choose" to use these
   primitive services as needed.  By selecting and using the appropriate
   set of service primitives, a range of failover scenarios may be
   supported.  These service primitives are described in the sub-
   sections that follow.

4.1 Initialization

   [OPEN TBD: what primitive(s) does a PU indicate what mappings can be
   used (are supported), whether automatic rollover, message retrieval
   are desired, etc.  These will likely be in the form of a
   initialization call]

4.2 PE Registration Services

   Pool Elements ("server") must use the following services to add or
   remove themselves from server pools:

      REGISTER, to add the pool element into a server pool using {pool
      handle, mapping mode, protocol or mapping id, port, policy info}
      where mapping mode is defined in Section 5.  A response result
      code is returned.

      DEREGISTER, to remove the pool element from a server pool using
      {pool handle, mapping mode, protocol or mapping id, port, policy
      info} where mapping mode is defined in Section 5.  A response
      result code is returned.

      TBD: if REGISTER also returns an opaque instance id, the
      application can just use that id for DEREGISTER, instead of
      passing in the (same) parameters used in REGISTER.

Conrad & Lei               Expires May 2, 2003                  [Page 8]

Internet-Draft        Services Provided by RSerPool        November 2002

4.3 Failover Callback Function

   The charter of the RSerPool Working Group specifically states that
   transaction failover is out of scope for RSerPool, i.e.  "if a server
   fails during processing of a transaction this transaction may be
   lost.  Some services may provide a way to handle the failure, but
   this is not guaranteed."  Accordingly, the RSerPool framework
   provides a "hook" for applications to provide their own application-
   specific failover mechanism(s).

   Specifically, an application can specify a callback function that is
   invoked whenever a failover has taken place.  This callback function
   is invoked immediately after the new transport layer connection/
   association is established with a new server, and gives the
   application the opportunity to send one or more messages that may
   help the server to resume any transaction or session that was in
   progress when the first server failed.

   As a simple example of how such a callback is useful, consider a file
   transfer service built using RSerPool.  Let us assume that some FTP
   mirroring software is used to maintain mirrored sites, and that the
   actual mirroring is out of scope.  However, we would like to use
   RSerPool to select a server from among the available mirror sites,
   and to failover in the middle of a file transfer if a primary server

   For this example, assume that a simple request/response protocol is
   used, where one request message results in one or more response
   messages.  Each request message contains the filename, and the offset
   desired within the file, (default zero.) Each response message
   contains some portion of the file, along with the offset, length of
   the portion in this message, and the length of the entire file.

   A single request results is sufficient to result in a sequence of
   response messages from the requested offset to the end of the file.
   For simplicity, assume that the response messages are delivered by
   the underlying transport strictly in order (although this requirement
   could be relaxed if a small amount of extra complexity were

   In this protocol, all that is needed for failover is for the
   application to keep track of the number of bytes that it has read
   from the server, and to provide a callback function that reissues the
   request to the new server, replacing the offset with this number.
   When there is no failover, only one request message is sent and the
   minimum number of response messages are returned; in the event of
   failover(s), single new request message is sent for each failover
   that occurs.

Conrad & Lei               Expires May 2, 2003                  [Page 9]

Internet-Draft        Services Provided by RSerPool        November 2002

   While this is a simple example, for more complex application
   requirements, the failover callback could be used in a variety of

      The client might send security credentials for authentication by
      the server, and/or to provide a "key" by which the server could
      locate and setup state by accessing some application-specific (and
      out-of-scope) state sharing mechanism used by the servers.

      The client might keep track of various synchronization points in
      the transaction, and use the failover callback to replay message
      from a recent synchronization point.

      [Open Issue TBD: Are there others to add to this list?]

4.4 PE Selection Services

   When automatic failover is enabled, selection of a new pool element
   according to the pool policy in place is automatically performed by
   the RSerPool framework in case of a detected failure (e.g.  provides
   automatic failover).  No application intervention is required.

   Automatic failover may be enabled by setting the appropriate send
   flag when used in conjuction with data channel services (described in
   Section 4.6) or explicitly during initialization when data channel
   services are not used.

      FAILOVER_INDICATION, delivered by callback, indicates that a
      failover has occurred and that any required application level
      state recovery should be performed.  The newly selected pool
      element handle is provided.

      Business Card services:  when automatic failover is used, the
      exchange of business cards for rendezvous services is
      automatically performed by the RSerPool framework (e.g.  no
      application intervention is required.

   When automatic failover is not enabled, failover detection and
   selection of an alternate PE must be done by the upper layer/
   application.  The following primitives are provided:

      GET_PRIMARY_SERVER, takes as input a pool handle and returns the
      {IP address, transport protocol, transport protocol port} of the
      primary server.

      GET_NEXT_SERVER has a dual meaning.  First, it indicates to the
      RSerPool layer the failure of the server returned by a previous

Conrad & Lei               Expires May 2, 2003                 [Page 10]

Internet-Draft        Services Provided by RSerPool        November 2002

      GET_PRIMARY_SERVER or GET_NEXT_SERVER call.  Second, it provides
      the {IP address, transport protocol, transport protocol port} of
      the next server that should be contacted, according to the best
      information available to the RSerPool layer at the present time.
      The appropriate pool policy for server selection for the pool
      should be used for selecting the next server.

4.5 Upper Layer/Application Level Acknowledgements

   The RSerPool framework provides an upper layer/application level ack
   service.  The upper layer protocol may request that the peer
   acknowledge receipt and successful processing of its sent data,
   providing an additional degree of confidence over transport level
   message retrieval.  When used in conjuction with the data channel
   services (described in Section 4.6), any unacknowledged data will be
   automatically sent to a new pool element in case of failover, if
   desired (e.g.  automatic failover is enabled).  The following service
   primitive is used to acknowledge an upper layer acknowledgement

      ULP_ACK, responds to a received upper layer acknowledgement

4.6 RSerPool Managed Data Channel

   The RSerPool framework provides these services to send and receive
   application layer data, which are used in place of the direct call of
   transport level system functions (e.g.  send/sendto, recv/recvfrom)
   and provides additional functionality to those calls.

      DATA_SEND, to send data to a pool element by using a pool handle,
      specific pool element handle, or by transport address.  An upper
      layer acknowledgement may be requested with this service.
      Appropriate error code(s) are returned.  When sending to a pool
      handle, the specific pool element handle is returned.

      DATA_INDICATION, delivered by callback, to indicate that data has
      been received from a pool element and to pass that data to the
      application layer protocol.  An application layer acknowledgement
      request can be indicated along with the data.

   The application MAY direct that the RSerPool framework multiplex both
   the control and data channels onto the same SCTP association/TCP
   connection/ etc., if desired.

Conrad & Lei               Expires May 2, 2003                 [Page 11]

Internet-Draft        Services Provided by RSerPool        November 2002

5. Transport Mappings

   While SCTP is the preferred transport layer protocol for applications
   built for RSerPool failover mode (for reasons explained shortly), it
   is also possible to use other transport protocols as well (e.g.  TCP)
   if an SCTP implementation is not available on the client and/or
   server.  However, there are certain features present in SCTP that are
   required if the RSerPool framework is to function in failover mode.
   When a transport protocol other than SCTP is used, these features
   must be provided by an "adaption layer" (also called a "shim
   protocol") that sits between the base transport protocol (e.g.  TCP)
   and the RSerPool layer.  We refer to these "adaptation layers" or
   "shim protocols" as "mappings" as the idea is that the requirements
   of the RSerPool framework are "mapped" onto the capabilities of the
   underlying protocol (e.g.  SCTP or TCP).

5.1 Defined Transport Mappings

   In order to support the RSerPool framework over a variety of
   transport protocols and configurations, several mappings are defined
   to provide RSerPool services over a given transport protocol.  Each
   mapping translates the requirements of the RSerPool framework onto
   the capabilities of the transport protocol desired (e.g.  SCTP, TCP,
   etc.).  Initially, three mappings are defined:

      NO_MAPPING (0x00):  With this mapping, no RserPool control channel
      is provided and the application specific communication between a
      pool user and the pool element (e.g.  data channel) is out of
      scope of RSerPool.  However, pool elements can register the
      application specific communication "protocol" and "port", and thus
      can be provided to pool users.

      SCTP (0x01):  SCTP transport is used for the RSerPool control
      channel.  The data channel MAY be multiplexed onto the same SCTP
      association, if desired.  This mapping is the preferred mapping.

      TCP (0x02): TCP transport is used for the RSerPool control
      channel.  The data channel MAY be multiplexed onto the same TCP
      connection, if desired.

   A particular pool element might support any combination of these
   mappings in order to support a variety of pool users with different
   capabilities (i.e.  different mapping support).  In this case, pool
   elements should register each mapping that it supports with its

Conrad & Lei               Expires May 2, 2003                 [Page 12]

Internet-Draft        Services Provided by RSerPool        November 2002

5.2 Transport Mappings Requirements

5.2.1 Mappings: Mandatory Requirements

   These features MUST be present in any mapping of the RSerPool
   framework mode to TCP (or any other transport protocol):

   1.  Message orientation, which facilitates application re-
       synchronization during failover.  Messages must be "framed" in
       order to allow for undelivered message retrieval from the
       transport protocol.

   2.  A heartbeat mechanism to monitor the health of an association or

   3.  A retrieval mechanism to allow an application to retrieve unsent
       or unacknowledged data from the transport layer upon failover.

   4.  A mechanism to transport and differentiate between control
       channel messages (e.g.  ASAP messages) and data channel messages.
       For example in SCTP, the payload protocol identifier (PPID) may
       be used.

   5.  [Open issue TBD: Are there others to be included here?]

5.2.2 Mappings: Optional Requirements

   There are several additional features that are present in SCTP that
   are lacking in TCP.  While these features are not crucial to
   RSerPool, providing them in the mapping layer makes it easier for an
   application layer programmer to write to a single API.  This single
   API can then be mapped over both SCTP and TCP, as well as any other
   transport protocol for which a mapping is provided.  Since these
   features are not essential for RSerPool, they are optional in any
   defined mapping.  However, appropriate error messages or indications
   should be provided when these features are not available.  These
   features include:

   1.  Support for multiple streams

   2.  Support for unordered delivery of messages

   3.  [Open issue TBD: Are there others to be included here?]

Conrad & Lei               Expires May 2, 2003                 [Page 13]

Internet-Draft        Services Provided by RSerPool        November 2002

5.2.3 Mappings: Other Requirements

   There are some features of SCTP that a mapping may not be able to
   provide, because they would require access to transport layer
   internals, or modifications in the transport layer itself.  The
   services provided by the RSerPool layer to the application should
   therefore provide mechanisms for the upper layer to access these
   features when present (e.g.  in SCTP), but also provide appropriate
   error messages or indications that these features are not available
   when they cannot be provided.  These features include:

   1.  Application access to the RTT and RTO estimates

   2.  Application access to the Path MTU value

   3.  [Open issue TBD: Are there others to be included here?]

6. Security Considerations

   [Open Issue TBD: Security issues are not discussed in this memo at
   this time, but will be added in a later version of this draft.]

7. IANA Considerations

   [Open Issue TBD: Will there be an enumeration of the various
   transport layer mappings that must be registered with IANA?]

8. Acknowledgements

   The authors wish to thank Maureen Stillman, Qiaobing Xie, Michael
   Tuexen, Randall Stewart, and many others for their invaluable


   [1]  Ong, L., Shore, M., Stillman, M., Xie, Q., Loughney, J., Tuexen,
        M. and M. Stewart, "Architecture for Reliable Server Pooling",
        draft-ietf-rserpool-arch-03 (work in progress), July 2002.

   [2]  Bradner, S., "Key words for use in RFCs to Indicate Requirement
        Levels", BCP 14, RFC 2119, March 1997.

   [3]  Loughney, J., "Comparison of Protocols for Reliable Server
        Pooling", draft-ietf-rserpool-comp-04 (work in progress), July

   [4]  Stillman, M., Xie, Q. and R. Stewart, "Enpoint Name Resolution

Conrad & Lei               Expires May 2, 2003                 [Page 14]

Internet-Draft        Services Provided by RSerPool        November 2002

        Protocol (ENRP)", draft-ietf-rserpool-enrp-04 (work in
        progress), September 2002.

   [5]  Stillman, M., Xie, Q., Tuexen, M. and R. Stewart, "Aggregate
        Server Access Protocol (ASAP)", draft-ietf-rserpool-asap-04
        (work in progress), July 2002.

Authors' Addresses

   Phillip T. Conrad
   Temple University
   CIS Department
   Room 303, Computer Building (038-24)
   1805 N. Broad St.
   Philadelphia, PA  19122

   Phone: +1 215 204 7910
   EMail: conrad@acm.org
   URI:   http://www.cis.temple.edu/~conrad

   Peter Lei
   Cisco Systems, Inc.
   955 Happfield Dr.
   Arlington Heights, IL  60004

   Phone: +1 847 870 7201
   EMail: peterlei@cisco.com

Conrad & Lei               Expires May 2, 2003                 [Page 15]

Internet-Draft        Services Provided by RSerPool        November 2002

Full Copyright Statement

   Copyright (C) The Internet Society (2002).  All Rights Reserved.

   This document and translations of it may be copied and furnished to
   others, and derivative works that comment on or otherwise explain it
   or assist in its implementation may be prepared, copied, published
   and distributed, in whole or in part, without restriction of any
   kind, provided that the above copyright notice and this paragraph are
   included on all such copies and derivative works.  However, this
   document itself may not be modified in any way, such as by removing
   the copyright notice or references to the Internet Society or other
   Internet organizations, except as needed for the purpose of
   developing Internet standards in which case the procedures for
   copyrights defined in the Internet Standards process must be
   followed, or as required to translate it into languages other than

   The limited permissions granted above are perpetual and will not be
   revoked by the Internet Society or its successors or assigns.

   This document and the information contained herein is provided on an


   Funding for the RFC Editor function is currently provided by the
   Internet Society.

Conrad & Lei               Expires May 2, 2003                 [Page 16]

Html markup produced by rfcmarkup 1.129c, available from https://tools.ietf.org/tools/rfcmarkup/