This site is operated by a business or businesses owned by Informa PLC and all copyright resides with them. Informa PLC's registered office is 5 Howick Place, London SW1P 1WG. Registered in England and Wales. Number 8860726.
Scalable Video and Scaling of Deployment
OK I am on a tear about scalable video coding (SVC), given that this is my fourth posting on this topic. You can find the others here (SVC, Packet Loss, and Heterogeneous Endpoints). My initial premise was that this technology will help us scale video conferencing deployments. What I mean by that is SVC helps overcome a number of issues that crop up when we want to deploy video to every desktop, not just to a few conference rooms. Let's look at why I think this is true.First, back to the packet loss issue. Video conferencing, when deployed to every desktop, will easily be the highest bandwidth user on the enterprise network. Today video conferencing requires QoS in the LAN and WAN to ensure low packet loss, low jitter and low delay to maintain video quality. But if video were to grow to be half or more of the bandwidth carried on the enterprise network, can QoS really work? QoS is a priority mechanism. When everyone is standing in the first-class line, no one is getting priority.
So I think there will be a big advantage to running desktop video at a lower class of service or even at best effort. The high priority classes will be saved for voice and for room-based video (telepresence or otherwise), and maybe for the CEO's executive desktop system. But for the rest of us, we will have to get by with a lossy network. In this environment, the resilience of SVC to packet loss will be a big bonus (see post on loss resiliency).
As video conferencing becomes a more pervasive communications style, we will want to connect to customers, partners and vendors using video as well. In fact this might even be more important than using video within the enterprise, as the human connections established with outside partners can really help make the business work well, and video can enhance that connection. So now we want to connect our video systems across enterprises, and the network that does that best is the Internet. But there is no QoS on the Internet, so we are back again to the issue of a lossy network and the need to overcome it.
The next scale issue is bandwidth management. If video is destined to become half of our network traffic, we must pay attention to managing it in some reasonable way. SVC has the potential of allowing each receiver to use the minimum bandwidth necessary for the current video display. If dynamic bandwidth adjustment is part of the video solution, more calls will be able to use the limited network bandwidth available at any given time.
The other half of bandwidth management is that the network will have limits on available bandwidth for the video class, and will start dropping packets (or remarking packets to a lower class) if that limit is exceeded. A desktop video solution will need to continue to operate well, if not perfectly, during congested periods. SVC has this potential, although traditional systems often have dynamic bandwidth adjustment capabilities as well.
In my opinion the biggest scaling issue is the multipoint control unit (MCU). This is the infrastructure component that allows three or more endpoints to communicate simultaneously. Today's MCUs are large, high-powered, dedicated computers capable of transcoding video conferencing calls between different bandwidths and different codec algorithms. Deploying sufficient traditional MCUs to support a full desktop deployment will be expensive, requiring rack space, power and significant capital investment.
SVC coding allows this same function to be done by a video switcher, which merely decides which streams of the SVC to forward and which to discard. his switcher function can be done today with standard computer hardware and these platforms can support many more simultaneous video streams than traditional MCUs. In the future this functionality may migrate into the network routers if SVC becomes prevalent.
Because of the structure of the underlying technology, the cost of multipoint support is much lower for SVC, thus allowing a much more rapid and cost effective scaling of video conferencing to every desktop.
OK, wait, don't run out and buy that SVC solution today. There are many more issues associated with scaling to this level of deployment, such as management, software updating, authentication of users, authentication of appliances on the network and more. New video conferencing vendors may not have all these answers in place yet. In fact they may not even have their business models fully in place so you can be assured of their ongoing existence. Be sure to look at the whole picture and not just the technology before jumping into a video conferencing solution that should keep you communicating for ten years or more.
Oh yeah, and then there is interoperability. I will post about that next week.