RFC 3409

Network Working Group                                         K. Svanbro
Request for Comments: 3409                                      Ericsson
Category: Informational                                    December 2002

    Lower Layer Guidelines for Robust RTP/UDP/IP Header Compression

Status of this Memo

   This memo provides information for the Internet community.  It does
   not specify an Internet standard of any kind.  Distribution of this
   memo is unlimited.

Copyright Notice

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


   This document describes lower layer guidelines for robust header
   compression (ROHC) and the requirements ROHC puts on lower layers.
   The purpose of this document is to support the incorporation of
   robust header compression algorithms, as specified in the ROHC
   working group, into different systems such as those specified by
   Third Generation Partnership Project (3GPP), 3GPP Project 2 (3GPP2),
   European Technical Standards Institute (ETSI), etc.  This document
   covers only lower layer guidelines for compression of RTP/UDP/IP and
   UDP/IP headers as specified in [RFC3095].  Both general guidelines
   and guidelines specific for cellular systems are discussed in this

Table of Contents

   1.  Introduction.................................................. 2
   2.  General guidelines............................................ 2
         2.1.  Error detection....................................... 2
         2.2.  Inferred header field information..................... 3
         2.3.  Handling of header size variation..................... 3
         2.4.  Negotiation of header compression parameters.......... 5
         2.5.  Demultiplexing of flows onto logical channels......... 5
         2.6.  Packet type identification............................ 5
         2.7.  Packet duplication.................................... 6
         2.8.  Packet reordering..................................... 6
         2.9.  Feedback packets...................................... 6
   3.  Cellular system specific guidelines........................... 7
         3.1.  Handover procedures................................... 7
         3.2.  Unequal error detection (UED)......................... 8
         3.3.  Unequal error protection (UEP)........................ 9

Svanbro                      Informational                      [Page 1]

RFC 3409         Lower Layer Guidelines for Robust HC      December 2002

   4.  IANA Considerations........................................... 9
   5.  Security Considerations....................................... 9
   6.  References.................................................... 9
   7.  Author's Address..............................................10
   8.  Full Copyright Statement......................................11

1.  Introduction

   Almost all header compression algorithms [RFC1144, RFC2507, RFC2508]
   rely on some functionality from the underlying link layer.  Headers
   (compressed or not) are expected to be delivered without any residual
   bit errors.  IP length fields are inferred from link layer length
   fields.  Packet type identification may be separated from the header
   compression scheme and performed at the underlying link layer.
   [RFC2509], for example, elaborates on how to incorporate IP header
   compression [RFC2507] in PPP [RFC1661].

   It is important to be aware of such assumptions on required
   functionality from underlying layers when incorporating a header
   compression scheme into a system.  The functionality required by a
   specific header compression scheme from lower layers may also be
   needed if incorporation of a header compression scheme is to be
   prepared without knowing the exact details of the final scheme.

   This document describes lower layer guidelines for robust RTP/UDP/IP
   header compression [RFC3095] as specified by the ROHC working group.
   [RFC3095] will from this point be referenced to as ROHC.  These
   guidelines should simplify incorporation of the robust header
   compression algorithms into cellular systems like those standardized
   by 3GPP, 3GPP2, ETSI, etc, and also into specific link layer
   protocols such as PPP.  The document should also enable preparation
   of this incorporation without requiring detailed knowledge about the
   final header compression scheme.  Relevant standardization groups
   standardizing link layers should, aided by this document, include
   required functionality in "their" link layers to support robust
   header compression.

   Hence, this document clarifies the requirements ROHC put on lower
   layers, while the requirements on ROHC may be found in [RFC3096].

2.  General guidelines

2.1.  Error detection

   All current header compression schemes [RFC1144, RFC2507, RFC2508]
   rely on lower layers to detect errors in (compressed) headers.  This
   is usually done with link layer checksums covering at least the

Svanbro                      Informational                      [Page 2]

RFC 3409         Lower Layer Guidelines for Robust HC      December 2002

   compressed header.  However, any error detecting mechanism may fail
   to detect some bit errors, which are usually called residual bit

   As for non-compressed IP packets, lower layers must provide similar
   error detection, at least for ROHC headers.  ROHC has been designed
   not to increase the residual bit error rate (for reasonable residual
   error rates) compared to the case when no header compression is used.
   Headers passed up to the header decompressor should, however, have a
   residual bit error probability close to zero.

   A ROHC decompressor might make use of packets with erroneous headers,
   even if they must be discarded.  It is therefore recommended that
   such invalid packets are passed up to the decompressor instead of
   being discarded by lower layers, but the packet must then be
   accompanied with an error indication.

2.2.  Inferred header field information

   Some fields of the RTP/UDP/IP headers may be classified as inferred,
   that is their values are to be inferred from other values or from an
   underlying link layer.  A ROHC decompressor requires that at least
   the following information can be inferred from any underlying link

   Packet Length (IPv4) / Payload Length (IPv6)

     The received packet (with compressed header) length.

   Length (UDP)

     This field is redundant with the Packet Length (IPv4) or the
     Payload Length (IPv6) field.

   In summary, all these fields relate to the length of the packet the
   compressed header is included in.  These fields may thus be inferred
   by the decompressor if one packet length value is signaled from the
   link layer to the decompressor on a per packet basis.  This packet
   length value should be the length of the received packet including
   the (compressed) header.

2.3.  Handling of header size variations

   It is desirable for many cellular link layer technologies that bit
   rate variations and thus packet size variations are minimized.
   However, there will always be some variation in compressed header
   sizes since there is a trade-off between header size variations and
   compression efficiency, and also due to events in the header flow and

Svanbro                      Informational                      [Page 3]

RFC 3409         Lower Layer Guidelines for Robust HC      December 2002

   on the channel.  Variations in header sizes cause variations in
   packet sizes depending on variations of payload size.  The following
   will only treat header size variations caused by ROHC and not packet
   size variations due to variations of payload size.

   The link layer must in some manner support varying header sizes from
   40 bytes (full RTP/UDP/IPv4 header) or 60 bytes (full RTP/UDP/IPv6)
   down to 1 byte for the minimal compressed header.  It is likely that
   the small compressed headers dominate the flow of headers, and that
   the largest headers are sent rarely, e.g., only a few times in the
   initialization phase of the header compression scheme.

   Header size variations and thus packet size variations depend on
   numerous factors.  Unpredictable changes in the RTP, UDP or IP
   headers may cause compressed headers to momentarily increase in size,
   and header sizes may depend on packet loss rate at lower layers.
   Header size distributions depend also on the mode ROHC operates in.
   However, for e.g., a voice application, carried by RTP/UDP/IPv4, with
   a constant speech frame size and silence suppression, the following
   basic header size changes may be considered as typical:

   In the very beginning of the speech session, the ROHC scheme is
   initialized by sending full headers called IR/DYN.  These are the
   largest headers, with sizes depending basically on the IP-version.
   For IPv4 the size is approximately 40 bytes, and for IPv6
   approximately 60 bytes.  The IR/DYN headers are used typically during
   one round trip time, possible interleaved with compressed headers.
   After that, usually only compressed headers are sent.  Compressed
   headers may vary in size from 1 byte up to several bytes.  The
   smallest compressed headers are used when there is no unpredictable
   changes in header fields, typically during a talk spurt.  In the
   beginning of a talk spurt, compressed header sizes may increase by
   one or a few bytes momentarily.  Apart from increases due to new talk
   spurts, compressed headers may increase in size momentarily due to
   unpredictable changes in header fields.

   ROHC provides some means to limit the amount of produced header
   sizes.  In some cases a larger header than needed may be used to
   limit the number of header sizes used.  Padding octets may also be
   used to fill up to a desired size.  Chapter 6.3 (Implementation
   parameters) in [RFC3095] provides optional implementation parameters
   that make it possible to mandate how a ROHC implementation should
   operate, for instance to mandate how many header sizes that may be

Svanbro                      Informational                      [Page 4]

RFC 3409         Lower Layer Guidelines for Robust HC      December 2002

2.4.  Negotiation of header compression parameters

   ROHC has some parameters that need to be configured in an initial
   setup phase.  Which header compression profiles are allowed may have
   to be determined and also what kind of context identification (CID)
   mechanism to use.

   The lower layers supporting ROHC should thus include mechanisms for
   negotiation of header compression parameters such as CID usage and
   header compression profile support.  In certain environments, it
   might also be desirable to have mechanisms for re-negotiation of
   these parameters.

   The negotiation must also make sure that compressor and decompressor
   use exactly the same profile, i.e. that the set of profiles available
   after negotiation must not include two profile identifiers with the
   same 8-bit LSB value.

   For unidirectional links, this configuration might have to be
   performed out-of-band or a priori, and similar methods could of
   course also be used for bi-directional links if direct negotiation is
   not possible.

2.5.  Demultiplexing of flows onto logical channels

   In some cellular technologies flows are demultiplexed onto radio
   bearers suitable to the particular flows, i.e., onto logically
   separated channels.  For instance, real-time flows such as voice and
   video may be carried on logically separated bearers.  It is
   recommended that this kind of demultiplexing is done in the lower
   layers supporting robust header compression.  By doing so, the need
   for context identification in the header compression scheme is
   reduced.  If there is a one to one mapping between flow and logical
   channel, there is no need at all for context identification at the
   header compression level.

2.6.  Packet type identification

   Header compression schemes like [RFC2507, RFC2508] have relied on the
   underlying link layer to identify different kinds of headers by means
   of packet type identifiers on link layers.  This kind of mechanism is
   not necessarily needed for ROHC since a ROHC packet type identifier
   is included in all compressed ROHC headers.  Only if ROHC packets are
   to be mixed with other packets, such as packets compressed by other
   header compression schemes, must the link layer provide a packet type
   identifier.  In such cases, or if ROHC is used on top of link layers
   already providing packet type identification, one (1) packet type
   identifier must be reserved for identification of ROHC packets. Thus,

Svanbro                      Informational                      [Page 5]

RFC 3409         Lower Layer Guidelines for Robust HC      December 2002

   only one ROHC packet type is needed to mix ROHC and e.g., RFC 2507
   flows, or to support ROHC on links where packet type identifiers are
   already present.

2.7.  Packet duplication

   Exact duplications of one and the same packet may waste transmission
   resources and is in contradiction to compression.  Even so, packet
   duplication may occur for various reasons.  Packet duplication may
   also occur in different places along the path for a packet.

   ROHC can handle packet duplication before the compressor but such
   packet duplications should be avoided for optimal compression
   efficiency.  For correct ROHC operation, lower layers are not allowed
   to duplicate packets on the ROHC compressor-decompressor path.

2.8.  Packet reordering

   Lower layers between compressor and decompressor are assumed not to
   reorder packets, i.e., the decompressor must receive packets in the
   same order as the compressor sends them.  ROHC handles, however,
   reordering before the compression point.  That is, there is no
   assumption that the compressor will only receive packets in sequence.

2.9.  Feedback packets

   ROHC may operate in three different modes; Unidirectional mode (U-
   mode), bidirectional optimistic mode (O-mode) and bidirectional
   reliable mode (R-mode).  A brief description of the modes can be
   found in chapter 4.4 of [RFC3095].

   In U-mode it is not necessary to send any feedback from the
   decompressor to the compressor.  O-mode and R-mode requires however
   that feedback messages from the decompressor to the compressor be
   sent.  Feedback messages consist of small ROHC internal packets
   without any application payload.  It is possible in ROHC to piggy-
   back feedback packets onto regular packets with ROHC compressed
   headers and payload, if there is ROHC type of compression in both the
   forward and reverse direction.  However, this piggy-backing may not
   be desired or possible in some cases.

   To support ROHC O-mode or R-mode operation, lower layers must provide
   transport of feedback packets from decompressor to compressor.  If
   piggybacking of feedback packets is not used, lower layers must be
   able to handle feedback as small stand-alone packets.  For optimal
   compression efficiency, feedback packets from the decompressor should
   be delivered as soon as possible to the compressor.

Svanbro                      Informational                      [Page 6]

RFC 3409         Lower Layer Guidelines for Robust HC      December 2002

3.  Cellular system specific guidelines

   An important group of link layer technologies where robust header
   compression will be needed are future cellular systems, which may
   have a very large number of users in some years.  The need for header
   compression is large in these kinds of systems to achieve spectrum
   efficiency.  Hence, it is important that future cellular systems can
   efficiently incorporate the robust header compression scheme.

3.1.  Handover procedures

   One cellular specific property that may affect header compression is
   mobility and thus, handover (i.e., change of serving base station or
   radio network controller).

   The main characteristics of handovers relevant for robust header
   compression are: the length of the longest packet loss event due to
   handover (i.e., the number of consecutive packet losses), and
   relocation of header compression context when necessary.

   Depending on the location of the header compressor/decompressor in
   the radio access network and the type of handover, handover may or
   may not cause disruptions or packet loss events in the (compressed)
   header flow relevant for the header compression scheme.  For
   instance, if soft handover is used and if the header
   compressor/decompressor reside above the combining point for soft
   handover, there will be no extra packet losses visible to the
   decompressor due to handover.  In hard handovers, where packet loss
   events due to handover is introduced, the length of the longest
   consecutive packet loss is most relevant and thus should be

   To maintain efficient ROHC operation, it should be ensured that
   handover events do not cause significant long events of consecutive
   packet loss.  The term "significant" in this context relates to the
   kind of loss tolerable for the carried real-time application.

   If hard handovers are performed, which may cause significant long
   events of consecutive packet loss, the radio access network should
   notify the compressor when such a handover has started and completed.
   The compressor could then be implemented to take proper actions and
   prevent consequences from such long loss events.

   Cellular systems supporting robust header compression may have
   internal mechanisms for transferring the header compression context
   between nodes where contexts may reside, at or before handover.  If
   no such mechanism for transferring header compression context between
   nodes is available, the contexts may be resynchronized by the header

Svanbro                      Informational                      [Page 7]

RFC 3409         Lower Layer Guidelines for Robust HC      December 2002

   compression scheme itself by means of a context refresh.  The header
   compressor will then perform a new header compression initialization,
   e.g., by sending full headers.  This will, however, introduce an
   increase in the average header size dependent on how often a transfer
   of context is needed.  To reinitialize the context in such cases, the
   lower layers must indicate to the header compressor when a handover
   has occurred, so that it knows when to refresh the context.  Chapter
   6.3 (Implementation parameters) in [RFC3095] provides optional
   implementation parameters that make it possible to trigger e.g., a
   complete context refresh.

3.2.  Unequal error detection (UED)

   Section 3.1 states that ROHC requires error detection from lower
   layers for at least the compressed header.  However, some cellular
   technologies may differentiate the amount of error detection for
   different parts of a packet.  For instance, it could be possible to
   have a stronger error detection for the header part of a packet, if
   the application payload part of the packet is less sensitive to
   errors, e.g., some cellular types of speech codes.

   ROHC does not require UED from lower layers, ROHC requires only an
   error detection mechanism that detects errors in at least the header
   part of the packet.  Thus there is no requirement on lower layers to
   provide separate error detection for the header and payload part of a
   packet.  However, overall performance may be increased if UED is

   For example, if equal error detection is used in the form of one link
   layer checksum covering the entire packet including both header and
   payload part, any bit error will cause the packet to be discarded at
   the ROHC decompressor.  It is not possible to distinguish between
   errors in the header and the payload part of the packet with this
   error detection mechanism and the ROHC decompressor must assume that
   the header is damaged, even if the bit error hit the payload part of
   the packet.  If the header is assumed to be damaged, it is not
   possible to ensure correct decompression and that packet will thus be
   discarded.  If the application is such that it tolerates some errors
   in the payload, it could have been better to deliver that packet to
   the application and let the application judge whether the payload was
   usable or not.  Hence, with an unequal error detection scheme where
   it is possible to separate detection of errors in the header and
   payload part of a packet, more packets may be delivered to
   applications in some cases for the same lower layer error rates.  The
   final benefit depends of course on the cost of UED for the radio
   interface and related protocols.

Svanbro                      Informational                      [Page 8]

RFC 3409         Lower Layer Guidelines for Robust HC      December 2002

3.3.  Unequal error protection (UEP)

   Some cellular technologies can provide different error probabilities
   for different parts of a packet, unequal error protection (UEP).  For
   instance, the lower layers may provide a stronger error protection
   for the header part of a packet compared to the payload part of the

   ROHC does not require UEP.  UEP may be beneficial in some cases to
   reduce the error rate in ROHC headers, but only if it is possible to
   distinguish between errors in header and payload parts of a packet,
   i.e., only if unequal error detection (UED) is used.  The benefit of
   UEP depends of course on the cost of UEP for the radio interface and
   related protocols.

4.  IANA Considerations

   A protocol which follows these guidelines, e.g., [RFC3095], will
   require the IANA to assign various numbers.  This document by itself,
   however, does not require IANA involvement.

5.  Security Considerations

   A protocol which follows these guidelines, e.g., [RFC3095], must be
   able to compress packets containing IPSEC headers according to
   [RFC3096].  There may be other security aspects to consider in such
   protocols.  This document by itself, however, does not add security

6.  References

   [RFC1144]   Jacobson, V., "Compressing TCP/IP Headers for Low-Speed
               Serial Links", RFC 1144, February 1990.

   [RFC1661]   Simpson, W., Ed., "The Point-To-Point Protocol (PPP)",
               STD 51, RFC 1661, July 1994.

   [RFC2507]   Degermark, M., Nordgren, B. and S. Pink, "IP Header
               Compression", RFC 2507, February 1999.

   [RFC2508]   Casner, S. and V. Jacobson, "Compressing IP/UDP/RTP
               Headers for Low-Speed Serial Links", RFC 2508, February

   [RFC2509]   Engan, M., Casner, S. and C. Bormann, "IP Header
               Compression over PPP", RFC 2509, February 1999.

Svanbro                      Informational                      [Page 9]

RFC 3409         Lower Layer Guidelines for Robust HC      December 2002

   [RFC3095]   Borman, C., Burmeister, C., Degermark, M., Fukushima, H.,
               Hannu, H., Jonsson, L-E., Hakenberg, R., Koren, T., Le,
               K., Liu, Z., Martensson, A., Miyazaki, A., Svanbro, K.,
               Wiebke, T., Yoshimura, T. and H. Zheng, "Robust Header
               Compression (ROHC)", RFC 3095, July 2001.

   [RFC3096]   Degermark, M., "Requirements for robust IP/UDP/RTP header
               compression", RFC 3096, July 2001.

7.  Author's Address

   Krister Svanbro
   Box 920
   Ericsson AB
   SE-971 28 Lulea, Sweden

   Phone: +46 920 20 20 77
   Fax:   +46 920 20 20 99
   EMail: krister.svanbro@ericsson.com

Svanbro                      Informational                     [Page 10]

RFC 3409         Lower Layer Guidelines for Robust HC      December 2002

8.  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.

Svanbro                      Informational                     [Page 11]