High Time for a Change

What’s wrong with traditional video conferencing? Let’s go straight to the heart of the matter: the Multipoint Control Unit (MCU). From a computational standpoint, this is where all the heavy lifting in a traditional video conferencing network is done. And that’s just the problem. Web-savvy applications handle much of the compute load out at the endpoints. MCUs, on the other hand, are like a lingering vestige from a bygone era when mainframes roamed the earth and endpoints were regularly treated as nothing more than dumb terminals. It’s high time we woke up and smelled the coffee.

Look at it this way: MCUs are expensive. From a configuration perspective, they’re neither flexible nor fungible. They don’t scale particularly well, and whenever resolution technology improves, it takes a forklift replacement for MCUs to keep pace. Also, because MCUs are only nominally error-resilient (unable to sustain anything more than 5% packet loss), they all but require the use of expensive dedicated lines. But now, let’s get down to the real problem with MCUs …

For all intents and purposes, it turns out that MCUs are nothing more than quality-degradation machines. And that’s because the transcoding they perform invariably degrades video quality. Also, transcoding is extremely time-consuming (adding up to 200 ms to the time it takes for each frame to leave the MCU). And all that takes place right in the middle of highly time-sensitive video transmissions, which ends up taking a huge toll on the user experience.

So, let’s briefly recap here: greater expense, less flexibility, limited scalability, increased latency, diminished quality. Not terribly inspiring, is it? This isn’t to minimize the important contributions made by the MCU over time. But it is to say that the MCU’s day has most certainly passed.