Browse Prior Art Database

RTP Header pre-compression with Mobile IP

IP.com Disclosure Number: IPCOM000022078D
Original Publication Date: 2004-Feb-23
Included in the Prior Art Database: 2004-Feb-23
Document File: 10 page(s) / 2M

Publishing Venue

Motorola

Related People

Jheroen Dorenbosch: AUTHOR [+2]

Abstract

Standard methods exist to compress the headers of IPv4 IP/UDP/RTP packets. These methods do not work when these packets are sent using Mobile IP. This article proposes compression and decompression methods for RTP packets that can be used at a Mobile IP Home Agent (HA) and Foreign Agent (FA) that allow for standard compression of the resulting packets on the links between the HA and the FA.

This text was extracted from a Microsoft Word document.
At least one non-text object (such as an image or picture) has been suppressed.
This is the abbreviated version, containing approximately 15% of the total text.

RTP Header pre-compression with Mobile IP

Jheroen Dorenbosch, Cyndi Jung

Abstract

Standard methods exist to compress the headers of IPv4 IP/UDP/RTP packets. These methods do not work when these packets are sent using Mobile IP. This article proposes compression and decompression methods for RTP packets that can be used at a Mobile IP Home Agent (HA) and Foreign Agent (FA) that allow for standard compression of the resulting packets on the links between the HA and the FA.

Figure 1. RTP traffic between a 4G IP core and a mobile station. Mobile IP is used for micro-mobility. Downlink RTP traffic is tunneled via a Home Agent in the core and a Foreign Agent at the base station site. On the backhaul downlink packets are wrapped in a Mobile IP header. This makes it hard to use commercially available compression methods.

Context

Many 4G systems use Mobile IP[1] (MIP) for mobility between sites, or even for mobility between sectors of the same site. An example 4G architecture is shown in Figure 1. Traffic (signaling and control) for a Mobile Station (MS) is sent to the MS’s home address. The MS’s home address terminates on a LAN in the 4G core. The MS’s Home Agent (HA) is attached to that LAN and picks up each packet sent to the MS’s home address, wraps it in a MIP IP header; and tunnels it to the MS’s Foreign Agent (FA). In the example architecture, the FA is on the base station and the base station is a layer 3 device. The FA detunnels the packet by removing the MIP IP header and forwards it on the local LAN to the MS on the MS’s home address. (Blue dotted line in Figure 1. The thicker line between the HA and the FA represents the extra MIP header.) In Figure 1 the HAs are implemented on HA servers. The HAs can also be on routers. According to the IPv4 MIP rules, the MS sends its uplink traffic MS directly to the destination IP address. Uplink packets are normally not tunneled (dot-dashed line in Figure 1).

Figure 2. Representation of the uplink and downlink packets on the backhaul.

Figure 2 shows example uplink and downlink packets on the backhaul. The figure is more or less to scale, assuming that an 8000 bps, 50 Hz vocoder is being used. The figure illustrates the large overhead from the headers. The 20-byte vocoder payload is preceded by the RTP[2], UDP and IP headers that are added by the originator. On the downlink the HA further adds the MIP header for IP-in-IP encapsulation[3]. This leads to a somewhat large link layer (L2) header. We assume that the backhaul uses the Multi-Link (ML) Point-to-Point Protocol (PPP)[4] to reduce jitter and to enhance availability, but other layer 2 protocols such as Frame Relay and ATM may be used as well and will have the same type of header overhead problem.

In a 4G system the majority of the packets on a backhaul may well be RTP packets related to telephony, dispatch and multi-media. Hence header compression will be very important. It directly impacts the number of T1 that will be needed per site...