Robust Header Compression enables IP Telephony to Mobile Phones
by Mikael Degermark
Since early 1999, researchers at SICS and Luleå University of Technology have been working with Ericsson Research on developing a robust header compression scheme suitable for IP telephony over cellular systems such as GSM or UMTS. The research has resulted in a novel header compression scheme called ROCCO, RObust Checksum-based header COmpression, with excellent performance.
To make IP telephony over cellular radio systems an economically viable alternative to circuit-switched voice, it is an absolute requirement that the 40-octet IP/UDP/RTP headers of IP telephony packets are reduced in size to conserve bandwidth and expensive radio spectrum. Links in cellular radio systems can have bit error rates of 10% before channel coding and 1/1000 after channel coding. Add a link roundtrip time of 200 milliseconds and all earlier schemes for header compression perform badly. This has been a major obstacle to using the Internet Protocol, IP, for transmitting telephony traffic to mobile phones.
Since early 1999, researchers at SICS and Luleå University of Technology have been working with Ericsson Research on developing a robust header compression scheme suitable for IP telephony over cellular systems such as GSM or UMTS. The research has resulted in a novel header compression scheme called ROCCO, RObust Checksum-based header COmpression, with excellent performance. ROCCO is now being standardized in the IETF, Internet Engineering Task Force, the protocol development and standardization body of the Internet.
With IP telephony gaining momentum and cellular telephony having several hundred million users, it seems inevitable that some future wireless telephony systems will be based on IP technology. What we today know as cellular phones may in addition to telephony and video have IP stacks, web browsers, email clients, networked games, etc. If based on IP, the telephony service will be considerably more flexible than today. The project initially concentrated on the problem of providing a good IP solution for speech, but it is understood that applications for video, games, etc, will also have to be supported. It is vital for cellular phone systems to use the radio resources efficiently in order to support a sufficient number of users per cell. Only then can deployment costs be kept low enough. It is also important to provide sufficiently high quality voice and video. In particular the voice service should be as good as what users expect from the cellular phone systems of today. A lower quality may only be accepted if costs are significantly reduced.
The radio channels used in cellular systems have very high bit error rates (BER) due to shadow fading, multipath fading, and continuous mobility. The radio signals of one user will interfere with the radio signals of other users, so with the desired number of users per cell, BERs will be high. Even after error correcting channel coding, the remaining BER can be as high as 10-3 (one in 1000) or even 10-2 (one in 100) for some users some of the time. Interleaving is used to improve the performance of the channel encoding, and together with other delays this brings the link roundtrip time to 100-200 milliseconds.
The existing standard for compressing IP/UDP/RTP headers is CRTP, defined in [RFC-2508]. It compresses headers over a single link by maintaining a context, which is essentially the full version of the last header transmitted over the link, at both ends of the link and transmitting only differences between consecutive headers. In this manner, each compressed header updates the context. When a packet is lost between compressor and decompressor, the context of the decompressor is not updated properly, and decompression will fail. To deal with such problems, CRTP has a context repair mechanism which relies on signaling. Essentially, the decompressor signals the compressor that its context is out of sync, and the compressor then sends updating information. Due to the long link roundtrip time, however, several voice packets will be lost before the repair is completed. CRTP thus performs very badly when error rates are high as each lost packet is accompanied by several packets being lost due to context mismatch. The only way to avoid context-related loss appears to be to enable the decompressor to repair its context locally, i.e., without going across the link. ROCCO computes a CRC over the header before compression and transmits the CRC as the compressed header. In this manner a reliable way to detect a successful context repair is introduced. It becomes feasible to have a decompressor attempt several repairs without risking introduction of erroneous headers. As headers of voice conversations are fairly regular, the decompressor will almost always succeed in repairing the context. When there are irregularities, the compressor adds minimal extra information to the compressed header.
ROCCO exhibits excellent performance with near-trivial context related packet loss. It is an enabling technique for cellular IP telephony. The development and standardization of robust header compression schemes are now being continued in the context of the ROBHC working group of the IETF, chaired by Degermark. In addition to IP telephony, the WG will work on robust header compression for low-bandwidth video and for TCP.
ROCCO homepage: http://www.ludd.luth.se/users/larsman/rocco/
Mikael Degermark Systemteknik, Luleå
University of Technology
Tel: +46 70 833 8933
E-mail: email@example.com, firstname.lastname@example.org