Today & tomorrow I'm attending a workshop on "Understanding and Troubleshooting Videoconferencing Networks". I was worried it would be over my head because it's really geared to networking people, but I am learning! The workshop is coordinated by ETOM and attendees are from K12, community colleges, and colleges around Michigan. The workshop is from the University of Wisconsin, and they could come to your state too! Tomorrow I'll post more details on how you could arrange for this in your area.
Here are some of my "take aways" from today's session.
H.323 was designed for corporate LANs
Now I know why sometimes I feel like I'm pushing the edges of this technology – with events like MysteryQuest and Read Across America and really any classroom-to-classroom projects. Because all use of IP videoconference outside of your own network is really beyond what the original designers of the H.323 standard envisioned. Among other things, they did not think of firewalls at all when designing the protocol! Which explains why we're having such troubles connecting to each other outside our networks! While this knowledge doesn't solve anything, it explains the root of the problems that annoy me daily. It also explains some of my frustration with vendors. The vendors are working within the H.323 standard and also mainly considering the corporate world. The design issues that affect the projects I want to do are there because the vendors never thought of using it quite the way we do with K12 projects. Having this perspective tempers my annoyance.
H.323 is for voice too
You probably knew this already, but I just learned it today. The H.323 protocol is used in Voice Over IP. SIP is the competing standard. Interesting to know.
Different versions of H.323
The H.323 standard is evolving and growing. So there are different versions of H.323. This explains several things. Why you can have two H.323 units that have issues connecting. Why you can have an "H.323 aware firewall" and it doesn't work with the endpoint. It also reminds me of the importance of paying for maintenance on endpoints. That way you get the upgrades as they fix issues and implement the new version of H.323. In theory things work better with the latest version.
H.325 is coming down the pike
Another standard is in discussion that addresses several issues with H.323. It is probably a few years in the future at least. Hopefully it will be backwards compatible. Either way, don't get too settled in the status quo! Change is always coming… especially with technology!
The IMTC runs interoperability events where vendors get together and test their equipment to see how they connect to each other. I'm glad to know that they do this considering how much we connect to "unknown" units around the world!
In my first two years of dabbling with IP, we had a Radvision Gateway/Gatekeeper. I'm not even sure now exactly what model it was. It did the functions of both the gatekeeper and gateway. I didn't understand the difference and used the terms interchangably. Well, now I've learned that the gateway converts between H.323 and H.320, and the gatekeeper controls and directs calls. Sometimes they are in the same box, some times the gatekeeper is packaged with a router, sometimes the gatekeeper and/or gateway are incorporated into the MCU. It helps to understand the functionality so that I can use the correct terms!
Endpoints Signaling Firewalls
The industry is very aware of the firewall issues. Much of the current work is now being done on the endpoint signalling the firewall. This sounds very promising and is encouraging to know! Another reason to pay maintenance contracts to get the new software upgrades as they address these issues.
Gatekeepers and Endpoints
The basic H323 call model is from network (IP) address to network address. However all other higher level call models use a gatekeeper (dialing plans, etc.). These call models "increase functionality." I put that in quotes because these gatekeeper call models assume a corporate network world. They don't address the unique issues that seem to arise mainly in the K12 world where we just want an IP address to call the other endpoint so our classes can do projects together. Calls that require gatekeepers, gatekeeper registration, and neighbor gatekeeper registration all increase the complexity of the calling. Although in theory they are intended to make it easier to call other people.
Right now my videoconferencing is primarily all in the direct IP address to IP address dialing world. But from what I learned in the workshop, vendors, standards, etc. are all moving in the direction of gatekeeper controlled dialing. So what does that mean for K12 videoconferencing? I'm still mulling it over.
There is a project called Videnet mainly in higher ed communities to use the Global Dialing Scheme. Some schools have experienced this dialing scheme when participating in Megaconference or Megaconference Jr. Many K12 schools in the United Kindgom are just now switching to IP and they are incorporating GDS. You can review this clear explanation of this scheme and how it is being implemented in the UK. These cooperative dialing schemes require higher level coordination and obviously the purchase of gatekeepers to make it all work. Is this where K12 VC in the U.S. should be heading? If so, how do we get there? Some states have statewide networks that would make this kind of dialing scheme easier to implement. In other places, a district or a building can buy an endpoint, hook it up, and be on their way doing videoconferencing, without a thought of the big picture or even without any knowledge or awareness of a wider dialing scheme. How would we address that? I'm thinking out loud here. What do YOU think?
I'm sure I'll learn more tomorrow and will write again….
Thanks for the more detailed account of the class. I was interested in going, but could not swing the time and budget. Maybe next year.