RFC 3006






Network Working Group                                          B. Davie
Request for Comments: 3006                                 C. Iturralde
Category: Standards Track                                       D. Oran
                                                    Cisco Systems, Inc.
                                                              S. Casner
                                                          Packet Design
                                                          J. Wroclawski
                                                                MIT LCS
                                                          November 2000


       Integrated Services in the Presence of Compressible Flows

Status of this Memo



   This document specifies an Internet standards track protocol for the
   Internet community, and requests discussion and suggestions for
   improvements.  Please refer to the current edition of the "Internet
   Official Protocol Standards" (STD 1) for the standardization state
   and status of this protocol.  Distribution of this memo is unlimited.

Copyright Notice



   Copyright (C) The Internet Society (2000).  All Rights Reserved.

Abstract



   An Integrated Services (int-serv) router performs admission control
   and resource allocation based on the information contained in a TSpec
   (among other things).  As currently defined, TSpecs convey
   information about the data rate (using a token bucket) and range of
   packet sizes of the flow in question.  However, the TSpec may not be
   an accurate representation of the resources needed to support the
   reservation if the router is able to compress the data at the link
   level.  This specification describes an extension to the TSpec which
   enables a sender of potentially compressible data to provide hints to
   int-serv routers about the compressibility they may obtain.  Routers
   which support appropriate compression take advantage of the hint in
   their admission control decisions and resource allocation procedures;
   other routers ignore the hint.  An initial application of this
   approach is to notify routers performing real-time transport protocol
   (RTP) header compression that they may allocate fewer resources to
   RTP flows.








Davie, et al.               Standards Track                     [Page 1]

RFC 3006       Integrated Services in Compressible Flows   November 2000


Table of Contents



   1      Introduction  ...........................................  2
   2      Addition of a Hint to the Sender TSpec  .................  3
   3      Admission Control and Resource Allocation  ..............  4
   4      Object Format  ..........................................  8
   4.1    Hint Numbering  .........................................  9
   5      Backward Compatibility  ................................. 10
   6      Security Considerations  ................................ 10
   7      IANA Considerations  .................................... 11
   8      Acknowledgments  ........................................ 11
   9      References  ............................................. 11
   10     Authors' Addresses  ..................................... 12
   11     Full Copyright Statement ................................ 13

1. Introduction



   In an Integrated Services network, RSVP [RFC 2205] may be used as a
   signalling protocol by which end nodes and network elements exchange
   information about resource requirements, resource availability, and
   the establishment and removal of resource reservations.  The
   Integrated Services architecture currently defines two services,
   Controlled-Load [RFC 2211] and Guaranteed [RFC 2212].  When
   establishing a reservation using either service, RSVP requires a
   variety of information to be provided by the sender(s) and
   receiver(s) for a particular reservation which is used for the
   purposes of admission control and allocation of resources to the
   reservation.  Some of this information is provided by the receiver in
   a FLOWSPEC object; some is provided by the sender in a SENDER_TSPEC
   object [RFC 2210].

   A situation that is not handled well by the current specs arises when
   a router that is making an admission control decision is able to
   perform some sort of compression on the flow for which a reservation
   is requested.  For example, suppose a router is able to perform
   IP/UDP/RTP header compression on one of its interfaces [RFC 2508].
   The bandwidth needed to accommodate a compressible flow on that
   interface would be less than the amount contained in the
   SENDER_TSPEC.  Thus the router might erroneously reject a reservation
   that could in fact have been accommodated.  At the same time, the
   sender is not at liberty to reduce its TSpec to account for the
   compression of the data, since it does not know if the routers along
   the path are in fact able to perform compression.  Furthermore, it is
   probable that only a subset of the routers on the path (e.g., those
   connected to low-speed serial links) will perform compression.






Davie, et al.               Standards Track                     [Page 2]

RFC 3006       Integrated Services in Compressible Flows   November 2000


   This specification describes a mechanism by which the sender can
   provide a hint to network elements regarding the compressibility of
   the data stream that it will generate.  Network elements may use this
   hint as an additional piece of data when making admission control and
   resource allocation decisions.

   This specification is restricted to the case where compression is
   performed only on a link-by-link basis, as with header compression.
   Other cases (e.g., transcoding, audio silence detection) which would
   affect the bandwidth consumed at all downstream nodes are for further
   study.  In these latter cases, it would be necessary to modify a
   sender TSpec as it is passed through a compressing node.  In the
   approach presented here, the sender TSpec that appears on the wire is
   never modified, just as specified in [RFC 2210].

2. Addition of a Hint to the Sender TSpec



   The appropriate place for a `compressibility hint' is the Sender
   TSpec.  The reasons for this choice are:

      -  The sender is the party who knows best what the data will look
         like.

      -  Unlike the Adspec, the Sender TSpec is not modified in transit

      -  From the perspective of RSVP, the Sender TSpec is  a set of
         opaque parameters that are passed to `traffic control'
         (admission control and resource allocation); the
         compressibility hint is just such a parameter.

   An alternative to putting this information in the TSpec would be to
   use an additional object in the RSVP PATH message.  While this could
   be made to work for RSVP, it does not address the issue of how to get
   the same information to an intserv router when mechanisms other than
   RSVP are used to reserve resources.  It would also imply a change to
   RSVP message processing just for the purposes of getting more
   information to entities that are logically not part of RSVP
   (admission control and resource allocation). The inclusion of the
   information in the TSpec seems preferable and more consistent with
   the Integrated Services architecture.

   The contents of the hint are likely to vary depending on the exact
   scenario.  The hint needs to tell the routers that receive it:

      -  the type of compression that is possible on this flow (e.g.
         IP/UDP/RTP);





Davie, et al.               Standards Track                     [Page 3]

RFC 3006       Integrated Services in Compressible Flows   November 2000


      -  enough information to enable a router to determine the likely
         compression ratio that may be achieved.

   In a simple case such as IP/UDP/RTP header compression, it may be
   sufficient to tell the routers nothing more than the fact that
   IP/UDP/RTP data is being sent. Knowing this fact, the maximum packet
   size of the flow (from the TSpec), and the local conditions at the
   router, may be sufficient to allow the router to determine the
   reduction in bandwidth that compression will allow.  In other cases,
   it may be helpful or necessary for the sender to include additional
   quantitative information to assist in the calculation of the
   compression ratio.  To handle these cases, additional parameters
   containing various amounts of information may be added to the sender
   TSpec.  Details of the encoding of these parameters, following the
   approach originally described in [RFC 2210] are described below.

3. Admission Control and Resource Allocation



   Integrated Services routers make admission control and resource
   allocation decisions based on, among other things, information in the
   sender TSpec.  If a router receives a sender TSpec which contains a
   compressibility hint, it may use the hint to calculate a `compressed
   TSpec' which can be used as input to the admission control and
   resource allocation processes in place of the TSpec provided by the
   sender.  To make this concrete, consider the following simple
   example.  A router receives a reservation request for controlled load
   service where:

      -  The Sender TSpec and Receiver TSpec contain identical token
         bucket parameters;

      -  The rate parameter in the token bucket (r) is 48 kbps;

      -  The token bucket depth (b) is 120 bytes;

      -  The maximum packet size (M) in the TSpecs is 120 bytes;

      -  The minimum policed unit (m) is 64 bytes;

      -  The Sender TSpec contains a compressibility hint indicating
         that the data is IP/UDP/RTP;

      -  The compressibility hint includes a compression factor of 70%,
         meaning that IP/UDP/RTP header compression will cause a
         reduction in bandwidth consumed at the link level by a factor
         of 0.7 (the result of compressing 40 bytes of IP/UDP/RTP header
         to 4 bytes on a 120 byte packet)




Davie, et al.               Standards Track                     [Page 4]

RFC 3006       Integrated Services in Compressible Flows   November 2000


      -  The interface on which the reservation is to be installed is
         able to perform IP/UDP/RTP header compression.

   The router may thus conclude that it can scale down the token bucket
   parameters r and b by a factor of 0.7, i.e., to 33.6 kbps and 84
   bytes respectively.  M may be scaled down by the same factor (to 84
   bytes), but a different calculation should be used for m.  If the
   sender actually sends a packet of size m, its header may be
   compressed from 40 bytes to 4, thus reducing the packet to 28 bytes;
   this value should be used for m.

   Note that if the source always sends packets of the same size and
   IP/UDP/RTP always works perfectly, the compression factor is not
   strictly needed.  The router can independently determine that it can
   compress the 40 bytes of IP/UDP/RTP header to 4 bytes (with high
   probability).  To determine the worst-case (smallest) gain provided
   by compression, it can assume that the sender always sends maximum
   sized packets at 48 kbps, i.e., a 120 byte packet every 20
   milliseconds.  The router can conclude that these packets would be
   compressed to 84 bytes, yielding a token bucket rate of 33.6 kbps and
   a token bucket depth of 84 bytes as before.  If the sender is willing
   to allow an independent calculation of compression gain by the
   router, the explicit compression factor may be omitted from the
   TSpec.  Details of the TSpec encoding are provided below.

   To generalize the above discussion, assume that the Sender TSpec
   consists of values (r, b, p, M, m), that the explicit compression
   factor provided by the sender is f percent, and that the number of
   bytes saved by compression is N, independent of packet size.  The
   parameters in the compressed TSpec would be:

     r' = r * f/100
     b' = b * f/100
     p' = p
     M' = M-N
     m' = m-N

   The calculations for r' and b' reflect that fact that f is expressed
   as a percentage and must therefore be divided by 100.  The
   calculations for M' and m' hold only in the case where the
   compression algorithm reduces packets by a certain number of bytes
   independent of content or length of the packet, as is true for header
   compression.  Other compression algorithms may not have this
   property.  In determining the value of N, the router may need to make
   worst case assumptions about the number of bytes that may be removed
   by compression, which depends on such factors as the presence of UDP
   checksums and the linearity of RTP timestamps.




Davie, et al.               Standards Track                     [Page 5]

RFC 3006       Integrated Services in Compressible Flows   November 2000


   All these adjusted values are used in the compressed TSpec.  The
   router's admission control and resource allocation algorithms should
   behave as if the sender TSpec contained those values.  [RFC 2205]
   provides a set of rules by which sender and receiver TSpecs are
   combined to calculate a single `effective' TSpec that is passed to
   admission control.  When a reservation covering multiple senders is
   to be installed, it is necessary to reduce each sender TSpec by its
   appropriate compression factor. The set of sender TSpecs that apply
   to a single reservation on an interface are added together to form
   the effective sender TSpec, which is passed to traffic control.  The
   effective receiver TSpec need not be modified; traffic control takes
   the greatest lower bound of these two TSpecs when making its
   admission control and resource allocation decisions.

   The handling of the receiver RSpec depends on whether controlled load
   or guaranteed service is used.  In the case of controlled load, no
   additional processing of RSpec is needed.  However, a guaranteed
   service RSpec contains a rate term R which does need to be adjusted
   downwards to account for compression.  To determine how R should be
   adjusted, we note that the receiver has chosen R to meet a certain
   delay goal, and that the terms in the delay equation that depend on R
   are b/R and C/R (when the peak rate is large).  The burstsize b in
   this case is the sum of the burstsizes of all the senders for this
   reservation, and each of these numbers has been scaled down by the
   appropriate compression factor.  Thus, R should be scaled down using
   an average compression factor

      f_avg = (b1*f1 + b2*f2 + ... + bn*fn)/(b1 + b2 + ... bn)

   where bk is the burstsize of sender k and fk is the corresponding
   compression factor for this sender.  Note that f_avg, like the
   individual fi's, is a percentage.  Note also that this results in a
   compression factor of f in the case where all senders use the same
   compression factor f.

   To prevent an increase in delay caused by the C/R term when the
   reduced value of R is used for the reservation, it is necessary for
   this hop to `inflate' its value of C by dividing it by (f_avg/100).
   This will cause the contribution to delay made by this hop's C term
   to be what the receiver would expect when it chooses its value of R.

   There are certain risks in adjusting the resource requirements
   downwards for the purposes of admission control and resource
   allocation.  Most compression algorithms are not completely
   deterministic, and thus there is a risk that a flow will turn out to
   be less compressible than had been assumed by admission control.
   This risk is reduced by the use of the explicit compression factor
   provided by the sender, and may be minimized if the router makes



Davie, et al.               Standards Track                     [Page 6]

RFC 3006       Integrated Services in Compressible Flows   November 2000


   worst case assumptions about the amount of compression that may be
   achieved.  This is somewhat analogous to the tradeoff between making
   worst case assumptions when performing admission control or making
   more optimistic assumptions, as in the case of measurement-based
   admission control.  If a flow turns out to be less compressible that
   had been assumed when performing admission control, any extra traffic
   will need to be policed according to normal intserv rules.  For
   example, if the router assumed that the 48 kbps stream above could be
   compressed to 33.6 kbps and it was ultimately possible to compress it
   to 35 kbps, the extra 1.4 kbps would be treated as excess.  The exact
   treatment of such excess is service dependent.

   A similar scenario may arise if  a sender claims that data for a
   certain session is compressible when in fact it is not, or overstates
   the extent of its compressibility.  This might cause the flow to be
   erroneously admitted, and would cause insufficient resources to be
   allocated to it.  To prevent such behavior from adversely affecting
   other reserved flows, any flow that sends a compressibility hint
   should be policed (in any router that has made use of the hint for
   its admission control) on the assumption that it is indeed
   compressible, i.e., using the compressed TSpec.  That is, if the flow
   is found to be less compressible than advertised, the extra traffic
   that must be forwarded  by the router above the compressed TSpec will
   be policed according to intserv rules appropriate for the service.
   Note that services that use the maximum datagram size M for policing
   purposes (e.g. guaranteed service [RFC 2210]) should continue to use
   the uncompressed value of M to allow for the possibility that some
   packets may not be successfully compressed.

   Note that RSVP does not generally require flows to be policed at
   every hop.  To quote [RFC 2205]:

      Some QoS services may require traffic policing at some or all of
      (1) the edge of the network, (2) a merging point for data from
      multiple senders, and/or (3) a branch point where traffic flow
      from upstream may be greater than the downstream reservation being
      requested.  RSVP knows where such points occur and must so
      indicate to the traffic control mechanism.

   For the purposes of policing, a router which makes use of the
   compressibility hint in a sender TSpec should behave as if it is at
   the edge of the network, because it is in a position to receive
   traffic from a sender that, while it passed through policing at the
   real network edge, may still need to be policed if the amount of data
   sent exceeds the amount described by the compressed TSpec.






Davie, et al.               Standards Track                     [Page 7]

RFC 3006       Integrated Services in Compressible Flows   November 2000


4. Object Format



   The compressibility hint may be included in the sender TSpec using
   the encoding rules of Appendix A in [RFC 2210].  The complete sender
   TSpec is as follows:

        31           24 23           16 15            8 7             0
       +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
   1   | 0 (a) |    reserved           |            10 (b)             |
       +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
   2   |    1  (c)     |0| reserved    |             9 (d)             |
       +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
   3   |   127 (e)     |    0 (f)      |             5 (g)             |
       +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
   4   |  Token Bucket Rate [r] (32-bit IEEE floating point number)    |
       +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
   5   |  Token Bucket Size [b] (32-bit IEEE floating point number)    |
       +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
   6   |  Peak Data Rate [p] (32-bit IEEE floating point number)       |
       +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
   7   |  Minimum Policed Unit [m] (32-bit integer)                    |
       +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
   8   |  Maximum Packet Size [M]  (32-bit integer)                    |
       +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
   9   |   126 (h)     |    0 (i)      |             2 (j)             |
       +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
   10  |     Hint (assigned number)                                    |
       +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
   11  |  Compression factor [f] (32-bit integer)                      |
       +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+

        (a) - Message format version number (0)
        (b) - Overall length (10 words not including header)
        (c) - Service header, service number 1 (default/global
              information)
        (d) - Length of service 1 data, 9 words not including header
        (e) - Parameter ID, parameter 127 (Token_Bucket_TSpec)
        (f) - Parameter 127 flags (none set)
        (g) - Parameter 127 length, 5 words not including header
        (h) - Parameter ID, parameter 126 (Compression_Hint)
        (i) - Parameter 126 flags (none set)
        (j) - Parameter 126 length, 2 words not including header

   The difference between this TSpec and the one described in [RFC 2210]
   is that the overall length contained in the first word is increased
   by 3, as is the length of the `service 1 data', and the original
   TSpec parameters are followed by a new parameter, the compressibility
   hint.  This parameter contains the standard parameter header, and an



Davie, et al.               Standards Track                     [Page 8]

RFC 3006       Integrated Services in Compressible Flows   November 2000


   assigned number indicating the type of compression that is possible
   on this data.  Different values of the hint would imply different
   compression algorithms may be applied to the data.  Details of the
   numbering scheme for hints appear below.

   Following the hint value is the compression factor f, expressed as a
   32 bit integer representing the factor as a percentage value.  The
   valid range for this factor is (0,100].  A sender that does not know
   what value to use here or wishes to leave the compression factor
   calculation to the routers' discretion may use the reserved value 0
   to indicate this fact.  Zero is reserved because it is not possible
   to compress a data stream to zero bits per second.  The value 100
   indicates that no compression is expected on this stream.

   In some cases, additional quantitative information about the traffic
   may be required to enable a router to determine the amount of
   compression possible.  In this case, a different encoding of the
   parameter would be required.

   In some cases it may be desirable to include more than one hint in a
   Tspec (e.g., because more than one compression scheme could be
   applied to the data.)  In this case, multiple instances of parameter
   126 may appear in the Tspec and the overall length of the Tspec and
   the length of the Service 1 data would be increased accordingly.

   Note that the Compression_Hint is, like the Token_Bucket_Tspec, not
   specific to a single service, and thus has a parameter value less
   than 128.  It is also included as part of the default/global
   information (service number 1).

4.1. Hint Numbering



   Hints are represented by a 32 bit field, with the high order 16 bits
   being the IP-compression-protocol number as defined in [RFC 1332] and
   [RFC 2509].  The low order 16 bits are a sub-option for the cases
   where the IP-compression-protocol number alone is not sufficient for
   int-serv purposes.  The following hint values are required at the
   time of writing:

      -  hint = 0x002d0000: IP/TCP data that may be compressed according
         to [RFC 1144]

      -  hint = 0x00610000: IP data that may be compressed according to
         [RFC 2507]

      -  hint = 0x00610100:  IP/UDP/RTP data that may be compressed
         according to [RFC 2508]




Davie, et al.               Standards Track                     [Page 9]

RFC 3006       Integrated Services in Compressible Flows   November 2000


5. Backward Compatibility



   It is desirable that an intserv router which receives this new TSpec
   format and does not understand the compressibility hint should
   silently ignore the hint rather than rejecting the entire TSpec (or
   the message containing it) as malformed.  While [RFC 2210] clearly
   specifies the format of TSpecs in a way that they can be parsed even
   when they contain unknown parameters, it does not specify what action
   should be taken when unknown objects are received.  Thus it is quite
   possible that some RSVP implementations will discard PATH messages
   containing a TSpec with the compressibility hint.  In such a case,
   the router should send a PathErr message to the sending host.  The
   message should indicate a malformed TSpec (Error code 21, Sub-code
   04).  The host may conclude that the hint caused the problem and send
   a new PATH without the hint.

   For the purposes of this specification, it would be preferable if
   unknown TSpec parameters could be silently ignored.  In the case
   where a parameter is silently ignored, the node should behave as if
   that parameter were not present, but leave the unknown parameter
   intact in the object that it forwards.  This should be the default
   for unknown parameters of the type described in [RFC 2210].

   It is possible that some future modifications to [RFC 2210] will
   require unknown parameter types to cause an error response.  This
   situation is analogous to RSVP's handling of unknown objects, which
   allows for three different response to an unknown object, based on
   the highest two bits of the Class-Num.  One way to handle this would
   be to divide the parameter space further than already done in [RFC
   2216].  For example, parameter numbers of the form x1xxxxxx could be
   silently ignored if unrecognized, while parameter numbers of the form
   x0xxxxxx could cause an error response if unrecognized.  (The meaning
   of the highest order bit is already fixed by [RFC 2216].)  A third
   possibility exists, which is to remove the unrecognized parameter
   before forwarding, but this does not seem to be useful.

6. Security Considerations



   The extensions defined in this document pose essentially the same
   security risks as those of [RFC 2210].  The risk that a sender will
   falsely declare his data to be compressible is equivalent to the
   sender providing an insufficiently large TSpec and is dealt with in
   the same way.








Davie, et al.               Standards Track                    [Page 10]

RFC 3006       Integrated Services in Compressible Flows   November 2000


7. IANA Considerations



   This specification relies on IANA-assigned numbers for the
   compression scheme hint.  Where possible the existing numbering
   scheme for compression algorithm identification in PPP has been used,
   but it may in the future be necessary for IANA to assign hint numbers
   purely for the purposes of int-serv.

8. Acknowledgments



   Carsten Borman and Mike DiBiasio provided much helpful feedback on
   this document.

9. References



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

   [RFC 1332]  McGregor, G., "The PPP Internet Protocol Control Protocol
               (IPCP)", RFC 1332, May 1992.

   [RFC 2205]  Braden, R., Zhang, L., Berson, S., Herzog, S. and S.
               Jamin, "Resource ReSerVation Protocol (RSVP) -- Version 1
               Functional Specification", RFC 2205, September 1997.

   [RFC 2210]  Wroclawski, J., "The Use of RSVP with IETF Integrated
               Services", RFC 2210, September 1997.

   [RFC 2211]  Wroclawski, J., "Specification of the Controlled-Load
               Network Element Service", RFC 2211, September 1997.

   [RFC 2212]  Shenker, S., Partridge, C. and R. Guerin, "Specification
               of Guaranteed Quality of Service", RFC 2212, September
               1997.

   [RFC 2216]  Shenker, S. and J. Wroclawski, "Network Element Service
               Specification Template", RFC 2216, September 1997.

   [RFC 2507]  Degermark, M., Nordgren, B. and S. Pink,"Header
               Compression for IP", RFC 2507, February 1999.

   [RFC 2508]  Casner, S. and V. Jacobson, "Compressing IP/UDP/RTP
               Headers for Low-Speed Serial Links", RFC 2508, February
               1999.

   [RFC 2509]  Engan, M., Casner, S. and C. Bormann, "IP Header
               Compression over PPP", RFC 2509, February 1999.




Davie, et al.               Standards Track                    [Page 11]

RFC 3006       Integrated Services in Compressible Flows   November 2000


10. Authors' Addresses



   Bruce Davie
   Cisco Systems, Inc.
   250 Apollo Drive
   Chelmsford, MA, 01824

   EMail: bsd@cisco.com


   Carol Iturralde
   Cisco Systems, Inc.
   250 Apollo Drive
   Chelmsford, MA, 01824

   EMail: cei@cisco.com


   Dave Oran
   Cisco Systems, Inc.
   170 Tasman Drive
   San Jose, CA, 95134

   EMail: oran@cisco.com


   Stephen L. Casner
   Packet Design
   66 Willow Place
   Menlo Park, CA 94025

   EMail: casner@acm.org


   John Wroclawski
   MIT Laboratory for Computer Science
   545 Technology Sq.
   Cambridge, MA  02139

   EMail: jtw@lcs.mit.edu











Davie, et al.               Standards Track                    [Page 12]

RFC 3006       Integrated Services in Compressible Flows   November 2000


Full Copyright Statement



   Copyright (C) The Internet Society (2000).  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
   English.

   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
   "AS IS" basis and THE INTERNET SOCIETY AND THE INTERNET ENGINEERING
   TASK FORCE DISCLAIMS ALL WARRANTIES, EXPRESS OR IMPLIED, INCLUDING
   BUT NOT LIMITED TO ANY WARRANTY THAT THE USE OF THE INFORMATION
   HEREIN WILL NOT INFRINGE ANY RIGHTS OR ANY IMPLIED WARRANTIES OF
   MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE.

Acknowledgement



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



















Davie, et al.               Standards Track                    [Page 13]