As mentioned in a previous tip, link efficiency is a QoS component used to ensure that voice packets don't have to "wait" as long to gain access to the network. Due to the special network performance requirements that are essential to transmitting voice traffic across a data infrastructure, link efficiency techniques allow voice packets to better gain access to the network by inserting voice packets between fragmented data packets or by compressing the voice packet's large header. Therefore, the significant queuing delays inherent to a network transmitting large data packets can be resolved using link efficiency techniques. This tip provides the details of some of the link efficiency methods available for today's networks.
Link Fragmentation and Interleaving (LFI)
Link Fragmentation and Interleaving is the process of fragmenting large packets to smaller sizes and interleaving small, real-time packets so they can be transmitted as required. Without this capability, large data packets would obstruct the transmission of voice packets. For example, a one way packet delay of 150ms is acceptable to assure high voice quality in a converged network. If a large data maximum transmission unit (MTU) of 2000 bytes were sent over a 56kbps line, it would take 286ms to cross the network. That delay alone would cripple the quality of voice transmission.
Three examples of Link Fragmentation and Interleaving are MLP Interleaving, FRF.12, and FRF.11 Annex C. MLP Interleaving encapsulates the larger traffic into ppp multilink and fragments the data to meet the delay requirements of voice traffic. The real time traffic is not encapsulated when transmitted between the fragmented traffic, but it is placed in a priority transmit queue to be sent before other flows. FRF.12 supports real time traffic over low speed, frame-relay links by fragmenting only the data frames that exceed a defined packet size within a data link connection identifier (DLCI). All smaller frames are then interleaved with the fragmented frames. It should be noted that FRF.12 does not differentiate or prioritize small voice packets over small data packets. FRF.11 Annex C addresses traffic across a DLCI configured for a Voice of Frame-Relay network. This method does recognize voice versus data packets by fragmenting only those frames with a data payload.
Compressed Real-Time Transport Protocol (CRTP)
Compressed Real-Time Transport Protocol is used on a network segment to compress the IP, UDP, and RTP headers from 40 bytes to a range of 2 to 5 bytes. Within a converged environment, voice framing speech occurs every 20 milliseconds. The total packet size is then represented by an IP header of 20 bytes, a UDP header of 8 bytes, a RTP header of 12 bytes, and a payload of 20 bytes. This results in a header size that is double the payload size, and for smaller network segments (below 768kbps) the header uses too much bandwidth. Therefore, CRTP is used to compress the packet headers by 87 to 95 percent.
Richard Parsons (CCIE#5719) is a Manager of Professional Services for Callisma Inc., a wholly owned subsidiary of SBC. He has built a solid foundation in networking concepts, advanced troubleshooting, and monitoring in areas such as optical, ATM, VoIP, routed, routing, and storage infrastructures. Rich resides in Atlanta GA, and is a graduate of Clemson University. His background includes senior and principal consulting positions at International Network Services, Lucent, and Callisma.