Hi, Haim and All:
I am glad that you took the pain to go through AT&T's contribution (APC-1434) submitted a long time ago. You are right. There are some fundamental issues that need to be addressed for scalability especially for the large-scale network(s). Let me summarize my personal view point what progress we have made so far:
* Inter-Domain Communications:
1. We have finalized the H.225.0 Annex G protocol. HopCount and other aspects (e.g., distributed and centralized domain concepts, caching mechanisms [but not how to manage the cache]) of AT&T's contribution were accepted. 2. More work needs to be done. For example, seamless communications (e.g., signaling messages) between the intra-domain and the inter-domain. How do we send signaling messages (RAS/Annex G, Call Signaling) among the BEs when domains are fully distributive (not fully connected)? How do we implement the application layer H.323 QOS on end-to-end basis using the same QOS services format from end users' point of view when pre-call or call signaling messages will go via none (direct), all, or some BEs between the source-destination domains considering the domains are distributive and each domain may implement different network layer QOS services? All these issues are subject to further standardization. 3. More contributions will allow us to proceed in this area in the future.
* Inter-Zone/Intra-Domain Communications:
1. No modification or improvement has been made in H.323. 2. AT&T contributions have been made to allow the more powerful Annex G messages for inter-zone communications as well. I request that all companies examine this proposal. 3. For two-GK model, I see that the existing signaling messages may work (even then we have some problems). However, we have, as you have also pointed out, serious problems in the case of multiple-GK (3, 10, ... ) environment. 4. We also need to provide more contributions solving these problems.
* Going-Forward:
1. We have opened a new work group know as "H.323 Robustness." 2. This group will have to deal with GK reliability as well. In this situation, we have to consider multiple GKs even in a given domain. We definitely need an new inter-GK protocol. 3. My suggestion would be: Let us consider the GK architecture and test the inter-GK protocol. Many of the suggestions made in AT&T contributions may still be applicable to develop the inter-GK protocol. 4. We may have a valuable experience while we develop the new inter-GK protocol to improve the robustness. 5. We may then apply the same experience how we can further improve the existing signaling messages for inter-GK communications as well.
For now, this is what I have to offer to you. Hope that others will also provide their comments.
Best Regards, Radhika R. Roy AT&T + 1 732 420 1580 rrroy@att.com
-----Original Message----- From: Haim Rochberger [SMTP:Haim_Rochberger@EUR.3COM.COM] Sent: Wednesday, November 03, 1999 4:36 AM To: ITU-SG16@MAILBAG.INTEL.COM Subject: Q: Intra Domain/Inter zone communication
I apologize for the wide distribution - I am a "new kid in town". I know that Annex G refers to inter-Domain connectivity... but my current interest is Inter-Zone/Intra-Domain. It is clear that GKs can broadcast LRQ msg to resolve non-local addresses (and cache them) , but this does not sound like a scaleable solution. I found a contribution APC-1434 from AT&T that gives some static and dynamic solution/guidelines. My questions are:
- What is the status of the contribution ? anyone knows if this is
accepted ? 2) Can someone point me to other places in any standard/contribution that address this issue (if such exist). If no special mechanism exist, and yet people agree that this is an issue (scaleability and redundancy) for Inter-Zone then let me know, and I can build a contribution that will offer those requirements.
Yours, Haim. e-mail: Haim_Rochberger@3com.com