Open Letter to Wainhouse Research Senior Analyst Andrew W. Davis

Dear Andrew,

I read with great interest your coverage of Cisco’s future multipoint video architecture. I am not sure if it was intentional, but your title selection resonated well with us: we all know what happened to the Empire! We (in the rebellion camp) are happy to see the big players admitting one after the other that all these expensive MCUs they are selling are a really bad investment – thank you for pointing this out.

This confirms what we have been saying all along: it is all about an efficient architecture that can replace the MCU. I did, however want to share my thoughts on some of the technical aspects around the Cisco solution.

The big question is HOW to replace the MCU, without losing functionality? Let me touch on a few aspects of this:

Error Resilience: As you mentioned in your analysis, Cisco provides error resilience using Forward Error Correction (FEC). FEC is a wasteful technique because you spend bits even when you don’t need them. Furthermore, it is particularly bad when error conditions are bursty. You identified this point much earlier in your November 2009 report “H.264 SVC: A Technical Assessment” (P. 25-30, “FEC overhead can range from 20% to 100%”). To make matters worse, the FEC overhead has to be added to the multi-stream overhead. Not to mention the fact that the FEC approach cannot avoid sending I frames, which degrades the quality of the call for all participants, not just the ones experiencing problems. FEC did not work with single-layer AVC and will not work well with multi-stream AVC either.

Personal Layouts and Bitrates: For years the key selling point of expensive MCUs was that you must be able to adapt the bitrate and viewing options to individually accommodate each participant. With the new Cisco solution everyone gets to see “Exactly the Same Thing” (like identical Storm Troopers). What happens with Cisco’s multi-stream solution if you happen to have 3/4 of the bandwidth required for full HD? Do you display the stamp-size 240p? And what if you want to see a 2×2 layout of 4 participants, you show them as 240p images on a 60-inch screen? What happens if some people want to see different people at HD? As you so rightfully point out in your commentary, “the difference is in the significant details.”

Error resilience matters, personal layout matters, and bitrate adaptation matters. And you can get all of them today, not in the future, with Vidyo’s implementation of video switching, using standard H.264 SVC. It’s called Adaptive Video Layering (AVL).

Interoperability: Arguably the most interesting feature of multi-streaming is the promise of better interoperability with legacy systems. From your note, however, it seems that the “native interoperability” claimed for Cisco’s architecture requires a Codian MCU to “act as the gateway between the multi-encoding world and the non-multi-encoding world for multi-point calls.” MOST Legacy devices participating in multipoint calls cannot do their own composition of the multi-stream, so they need someone to do it for them. The advantage, if any, will be limited to a small subset of Cisco’s newer endpoints.

In reality the market has already indicated where it is going, with a roster of Vidyo’s licensees that includes Google, Ricoh, and Hitachi, among many others, and Polycom and Microsoft announcing their plans to support SVC. Wainhouse Research has once again done a nice job in providing its readership with the information necessary to connect the dots with regard to where the video conferencing industry is heading. Everyone now agrees – there’s no future for the MCU.

- Ofer Shapiro.

———————————————————————————————————————

Vidyo Co-Founder and CEO Ofer Shapiro

Ofer co-founded Vidyo in 2005 and pioneered Personal Telepresence, enabling a new generation of software-based natural, multi-point HD video conferences on desktop computers, room systems, and mobile devices. Prior to Vidyo, Ofer spent eight years at RADVISION where he was responsible for the development of the first IP video conferencing bridge and gatekeeper technology and the first commercially successful video conferencing architecture. He also served as senior vice president of business development responsible for strategic sales and relationships. Ofer was a contributor and one of the editors of the H.323 standard. Ofer has over fifteen years of experience in bringing disruptive technologies to market. He holds a Master’s Degree in Applied Physics.

5 thoughts on “Open Letter to Wainhouse Research Senior Analyst Andrew W. Davis

  1. Good technical perspective. Next challenge is that all will be protecting their past investments in MCU’s or MCU development (customers and manufacturers alike). In that space Vidyo is also a very good solution as it allows interconnectivity without the ‘old-school’ vendor lock-in.

  2. Vidyo is perfect in the video world especially in terms of working under adverse network conditions, efficient way of handling multipoint calls and scaling without limitations. However will it be able to have the success that webex had still remains to be seen.

  3. Nice article and I understand Vidyo is really good company to make many people’s dream true.

    But well, I have a few doubts or comments:
    - Error Resilience: Really? Do you believe FEC is only the solution for it? I have developed bitrate adaptation algorithm in 3GPP world for 10 years and unfortunately no vendor used FEC actually with it in our industry. In VC, yes we can use a few vendors’ one but so far no compatibility due to no standard. Now recently IETF released new standard for FEC and many people are researching FEC framework, then more vendors would use this. And deliverying I-frame is traditional way. In H.264 world, thing is little different and possible without I-frame.
    - Personal Layouts and Bitrates: agree. I know SVC is really attractive to MCU makers. But I am still not sure if this is really being the next standard like H.264 AVC. Many people starts considering H.265 having SVC scheme. If we should invest, should it be UMTS or CDMA 1x EVDO from 2G? In Korea, we know that choosing 1x EVDO was mess and eventually we encounter cost matter, investment efficiency, and interoperability(think why IMT2000 idea came out). I mean changing legacy infra is not easy, however, still there is some places how Vidyo works well because nowadays we leave in new era using Tablet and Smart phones.
    - Interoperability: Interesting. No MCU like P2P concept of multi-streaming? Then what will be the result of cloud base VC solution which heavily depends on MCU system? If there is no MCU, still Vidyo’s business model works? I think this is irony. Many telco still believes a MCU features backed by a company like Vidyo to bring up their million seller business of tablet, while 4G is being legay but again not easy to find killer applications for them. And how can you contribute big VC rental project and Telco’s margin? And how you can provide management solution without MCU? Probably possible but limited. So, my conclusion is still MCU is going on in future; at least in Enterprise market.

Leave a Reply

Your email address will not be published. Required fields are marked *

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>

Current day month ye@r *