When we started Blue Jeans a few years ago we had one simple goal - we wanted to make business video calling as easy and affordable as audio calling. Unlike many startups that start with a piece of technology and try to find applications for it, we started with that very simple use case of business video calling. We then embarked on finding the right technologies that could help us solve it. We wanted to understand why video calling was not pervasive despite many technological advances in the last 5 years. We spoke to many end users, IT personnel, resellers and industry insiders and started piecing together the issues that inhibited widespread use. It boiled down to ease of use, interoperability and cost. And at the center of this old world of business video calling was this beast called the MCU. We had to slay the beast to unlock the next level in the game!
Humor aside ... what is an MCU and why is it so central to business video conferencing? MCU stands for Multipoint Control Unit and the term came from the telephony world. An MCU is needed whenever 3 or more parties want to be in an audio (or video) call. It acts as a mixer and if necessary, acts as a translater. It can be in a central location or co-located with one or more of the parties, but it needs to mix/translate all the audio and video so that everyone is seen and heard. This translation many times involves transcoding, a technique that allows one audio/video codec format to be transformed to another in real-time.
Why is the MCU a beast? Traditional MCUs are heavy metal custom-built hardware devices chockfull of DSPs and FPGAs - specialized chips that require armies of programmers and hardware engineers to build. As we talked to more and more people we heard the same repeated theme. Nobody liked MCUs, except the MCU vendors of course. They were hard to manage and operate. They cost well into the 6 figures $$. We even heard of a project within a large MCU vendor that expected to build an MCU they could sell for a 7 figure $$ amount. On top of all of this, these hardware MCUs were the most inflexible platforms when it came to keeping up with the pace of innovation and growth that we wanted to achieve.
Nevertheless we realized very early in our research that we needed to crack the code on the next generation MCU architecture if we wanted a shot at making widespread business video calling a reality. Why did we come to this realization? It is simple. We wanted to build a service that allowed any video conference room system, software client, video chat client or mobile client to talk to each other. We were not willing to take the "cop out" approach that some recent entrants into the video conferencing world were taking by replacing all your clients with new software and video codecs (like H.264/SVC) that try to make it easier to build a lower cost MCU. This was the easy way out. We were not willing to modify or enhance the end user clients to work with a built-in peer-to-peer MCU either since the approach would not scale beyond simple consumer chat usage scenarios. As I mentioned in my previous blog post - we do not believe the world will converge to one type of video client or codec standard, let alone allow the fork lift upgrade of all existing systems anytime soon. So we decided early on that we had to take the MCU beast head-on.
We had to build an MCU that was scalable, interoperable, flexible, easy-to-use and had a low cost to build and operate. Necessity is the mother of invention - there was no choice but to avoid, like the plague, going down the path of building or buying a traditional MCU architecture with custom hardware, DSPs and FPGAs. Inflexibility of the platform aside, I would have been walked out the door by any investor as soon as I got to the second slide! This is where the survival instinct of startups and their passion to solve an end user's problem kicks in. We started exploring how we could solve this MCU problem by using recent innovations in technology without taking the "cop out" approach. We looked at advances in multi-core server computing, the advent of the GPU (Graphics Processing Unit) to solve more than graphics problems, large scale cloud computing with clustered off-the-shelf hardware, globally distributed architectures like CDN (Content Delivery Networks). We believed that many of these techniques could be applied to build a massively scalable next generation MCU. We essentially took the approach of blowing up the MCU and decomposing it into an architecture that would allow us to apply all of these recent advances in computing to solve the MCU problem. With this our virtualized, cloud-based MCU alternative was born, what I sometimes refer to as the "biggest, baddest MCU on the planet". We could not have done this even 3 years ago had it not been for these recent technological advances.
Some of the areas where we have innovated to bring you this virtualized, cloud-based MCU alternative that will enable the next phase of growth in global video calling include:
So there you have it - the traditional MCU as we know it today is dead. The beast has been slain. We apologize for the inconvenience caused to anyone. I assure you the game is going to be a lot more fun, now that we've unlocked this next level! The scalable, virtualized, cloud-based MCU alternative is here to stay and most users won't even know it exists because - it just works.
Technology is magical when it blends in with your everyday life and you don't even know that it exists. The virtualized, cloud-based MCU alternative from Blue Jeans is one such technology!