Clearing the Voice-to-Video Chasm

We live in a world with very few absolutes (beyond, of course, the traditional death and taxes). Even ancient pillars of absolutism such as morality and truth now fall under the crushing wheels of relativism, thanks largely to decades of bashing and bludgeoning by politicians.

This atmosphere, in which facts and figures have a Gumby-like elasticity, places all claims and statistics that are not enveloped in a clearly defined context in under a shadow of doubt and suspicion. Unfortunately, the telecommunications sector is no exception. As the wild, wild world of Web 2.0 seems destined to wreak havoc on nearly every aspect of the telecom industry, one must beware of Geeks bearing benchmarks.

Now more than ever, performance claims by telecom equipment vendors are to be weighed with a healthy degree of skepticism. As voice gives way to other forms of multimedia, mainly video, familiar performance metrics, such as concurrent sessions, can no longer be trusted. Such is the situation with products in the session management space, which is roughly made up of equipment such as routers, session border controllers (SBCs), security gateways, and an emerging class of products categorized as "convergence gateways."

All of these devices carry some responsibility for managing the flow of communications traffic – both signaling and media – across the access portion of service provider networks. Until recently, real-time communication traffic was synonymous with voice, which in most cases requires a 64-kbit/s (or smaller) chunk of bandwidth. Unless you've been under a rock the size of Pittsburgh for the past year, however, you know that media streams connected to real-time communications sessions are migrating from voice to video, a medium that requires several times more bandwidth.

As a result of the expansion of the pipe needed to carry real-time and streaming traffic from one subscriber to another – or from server to subscriber – the performance numbers associated with the equipment that oversees the transport of this traffic must be recalibrated. The benchmark most in question is the equipment's ability to support concurrent sessions, in other words the product's ability to scale and to accommodate simultaneous users. While setting up sessions is largely a function of the control plane of these devices, that number is meaningless if the platform lacks the muscle to move the media associated with those sessions. With the transition from voice to video, the required bandwidth per session could jump from 64 kbit/s to 300 or 400 kbit/s faster than you can say YouTube. Products originally built for voice could end up suffering a severe decline in performance in the Web 2.0 world.

What this means to telcom operators is that performance numbers of SBCs, routers, convergence gateways and other session management devices are meaningless if not put in the proper context. The only way to accurately compare and contrast these platforms is by using the same yardstick, which means using identical bandwidth-per-session metrics when calculating the equipment's ability to support concurrent sessions.

The social networking revolution is turning everybody's world upside down. Just as parents must now troll their teenager's online circle of 2 million acquaintances for would-be predators, service providers have to come up with a new formula for comparing potential products that control the various IP sessions flowing across their networks.

– Joe McGarvey, Senior Analyst, Heavy Reading

Be the first to post a comment regarding this story.
Sign In