In this article on the Sametime Blog, the new Sametime 9 feature of the VMCU ( Video Multipoint Control Unit ” is well described:
So, now that I’ve admitted to becoming a video convert – my team seems to think I’m ready to become a regular blogger…… Next thing you know, they’ll be expecting me to be tweeting hourly! We’ll see – maybe a video blog entry is the next logical step. No time like the present to jump on the social bandwagon!
Last week, I shared some of my personal experiences and the value I realized from using Sametime 9 video. I thought this week, I would take a few minutes to pop the hood and give you a look at the engine.
The awesome video experience you’ve heard so much about is due to the addition of two new server components and the introduction of a new media framework in our clients. Sametime 9 AV is based on the H.264 SVC technology.
SVC stands for Scalable Video Coding, which enables efficient encoding of video that can be realized at different resolutions, frame rates and quality. Each video image is organized into layers – and a client can request one or more. When bandwidth is scarce, the client can request fewer layers. The SVC client also has the ability to request video from multiple participants and then build the continuous presence view locally. Within Sametime Conferencing and Sametime Complete, individual Sametime users
have the ability to choose the layout they would like to see.
have the ability to choose the layout they would like to see.
The Sametime Video MCU (VMCU) is a software based MCU (Multipoint Control Unit) – commonly referred to as a “video bridge”. The VMCU does not do any transcoding of video which makes it highly scalable. With Sametime’s authorized user licensing model – and Sametime 9’s inclusion of external participants – Sametime customer are also able to deploy as many VMCU’s in as many locations as they choose to best suit their needs.
Sametime’s implementation of SVC supports multiple resolutions and temporal layers. Sametime clients choose the resolutions and bit rates they want. The base layer is H.264 AVC.
In a multipoint Sametime 9 video call, each client transmits a video stream consisting of multiple resolutions and temporal layers. Each client then decides which streams they want, and how to display them. They may request a number of media streams based on the client’s capabilities.
In the example to the right, user 1 has chosen to view user 3 in an Active Speaker Wide format – while user 2 & 3 have chosen grid format seeing both other participants uniformly. In this example user 3 may be bandwidth constrained and has only requested the base layer.
Because the base layer is H.264 AVC, Sametime 9 is also compatible with AVC endpoints – which means it’s backward compatible with Sametime 8.5.2 clients and interoperable with H.264 AVC video endpoints.
The example to the left, shows an AVC endpoint participating in a Sametime 9 video call. The video transmitted by the AVC client will be sent to any SVC client that requests it and those users will be able to see user 2 in their continuous presence view. Since the Sametime MCU does not do any transcoding of the video, it will send only a single “base layer to the AVC client. As a result, user 2 will always see the active speaker.
This would allow you to deploy your Sametime 9 video – and still support those users who may not yet have the new client. However, once they find out what they are missing – believe me, they will want the Sametime 9 client!
Now that you’ve got an idea of how the new Sametime SVC based clients work with the VMCU – let’s me tell you a bit about the two server components and how they can be deployed.
In addition to the Sametime VMCU, we also introduced the Sametime Video Manager (VMGR) in Sametime 9. The VMGR is part of the Sametime Media Manager install package.
Unlike Sametime 8.5.2, the video infrastructure in Sametime 9 can be globally distributed. The VMGR can be deployed in a cluster for scalability and reliability. VMCUs can be deployed in pools. The VMU pools can span geographies. A pool of VMCUs is managed by a single VMGR or a cluster. The VMGR will assign conferences to a particular VMCU based on load and policy.
As I noted last week, Sametime 9 video on the surface is all about rich engagement via a high quality continuous presence video experience which can boost the effectiveness of your workforce. Today, I wanted to give you a glimpse that behind that pretty face – Sametime 9 video also brings critical new features behind the scenes such as efficient use of bandwidth, compatibility with AVC clients, global distribution, and scalability and reliability.
Stay tuned… more to come!
Tips and News for Collaboration solutions 