Hi, Chris:
I may have misunderstood your questions. I have provided the reply assuming that it is related to the H.323 terminal, user, and service mobility. It is related to H.323 Annex H that deals with mobility in the H.323 layer.
I think that your questions have been related to Annex I where the lower layers (below H.323 layer) are taken care of when mobile users move from place to place. If it is so, my reply does NOT answer your questions. If it is so, I am extremely sorry.
With respect to Annex I, I have to refer to my original clarifications what needs to be done in the context of H.323 Annex I. I mailed my notes to the SG16 reflector sometime ago. Email discussions were held and more actions will follow on this item.
In the light of Annex I, I do not have clear answer for your questions. We have not yet finalized what it will look like. I made some suggestions earlier. We like to get your inputs how we can shape H.323 Annex I. So far, Annex I have mentioned what needs to be done to improve the performance of the radio link layer. We need to go further how the H.323 mobility solution can be implemented over the lower layer including the IP layer. The IP layer may contain mobile IP, cellular IP, and/or any other services. My personal wish is that we would move to this area when finish the extension of H.323 for mobility. For example, I'd like to addess how H.323 mobile protocol can be used to satisfy the needs of the INM-2000/3GPP's wireless architecture in the context of the mobile H.323 terminal (MT), base station (BS), base station switching center (BSC), mobile switching center (MSC), and IP network. I believe that BSC can dela with link/MAC multiplexing layer while MSC will be an IP router with the knowledge of the IP layer. Wireless cells will be between the MT and BS. Now it depends where one puts the GK (and defines zones/domains) and how it communicates with the H.323 MT. It will then automatically bring all related issues: IP addresses, domains, and others. I am also thinking whether we will make this as a part of H.323 Annex I or not.
I am fully confident that our H.323 mobility architecture (please also see AT&T's and Motorola's contributions) will be cosistent to complement the architecture defined by IMT-2000/3GPP.
Now I may come to your questions: How do we related the zone (or GK) in the context of the lower layer? Is that relate to the IP addresses or IP domains only as you have mentioned? I will assume that the lower layer may contain the cell (pico, micro, macro) and/or wireless LANs (with or without cell concept) in wireless environments. In wire-line networks, the zone may consist of the usual network layer as H.323 system envisions. So, a zone may contain both wire-line and wireless networks as well.
I guess that we are not in a position to answer your questions yet. Is it possible to suggest some framework how we can drive the work in H.323 Annex I if you do not find answers of your questions provided above?
Best regards, Radhika
-----Original Message----- From: Roy, Radhika R, ALARC Sent: Friday, December 24, 1999 10:06 AM To: ITU-SG16@MAILBAG.INTEL.COM Subject: Re: H.323 Annex I
Hi, Chris:
We have an H.323 Ad Hoc Mobility Group to perform this work. We have been discussing this for almost a year why we need H.323 mobility. All wire-line and wireless vendors and many service providers are also a part of it. All of them brought contributions to extend the H.323 protocol for mobility.
I am very much surprised to see your comments. I guess that all people who are working to address this issue will also feel the same.
I do not want to confuse anyone by making a simple statement that will make you feel why we need extensions in H.323 for mobility. So many papers have been published by many researchers across the world for extension of H.323 for mobility. One of the excellent papers had been published in INFOCOM'99 referred in AT&T's contribution as well. I would suggest to kindly read this paper first before making any comments like "misconception" or others, etc. May be I have not been able to understand your points. I'd would also suggest to attend the discussion of the Ad Hoc mobility conf. calls.
Finally, I'd only suggest to examine how a GK will account the resources when an H.323 mobile moves from one place to another. The resources of one place have to be de-allocated in one place, while the resources in another place are to be allocated. When people think about resources, these resources are in terms of the H.323 layer. It does not matter how you think about the zone whether it is linked to physical locations, IP domain, or something else.
May be you might have a better solution to deal the mobility. May we suggest to bring contributions to solve the H.323 mobility problems that you might think to be appropriate.
Best regards, Radhika
-----Original Message----- From: Chris Wayman Purvis [SMTP:cwp@isdn-comms.co.uk] Sent: Friday, December 24, 1999 8:27 AM To: Roy, Radhika R, ALARC; Mailing list for parties associated with ITU-T Study Group 16 Subject: Re: H.323 Annex I
Radhika, All,
I've finally managed to read that contribution, and it seems to me that the requirement for a mobile H.323 protocol arises entirely out of the misconception that gatekeeper zones are necessarily linked to physical locations or IP domains. Remove that assumption and I am still
completely
baffled as to why any protocol above the network layer is required.
I will look at the MTD contributions when I find time (which is depressingly short for this sort of activity at the moment), but would be glad of anybody's simple explanation of why H.323 mobility is required without the above assumptions on gatekeeper zones.
Regards, Chris
"Roy, Radhika R, ALARC" wrote:
Hi, Chris:
I guess that you have missed all discussions and contributions related
H.323
mobility (SG16 meeting in May'99, Berlin Aug'99, and Red Bank NJ
Oct'99).
H.323 is in application layer.
IP/IPX is in the network layer.
Radio/ATM is in the link layer.
Mobility may have an impact in all layers. If the link layer mobility
is
transparent to the network layer, nothing should be done in the
network
layer. Similar is the case for others.
When the mobility has an impact in the H.323 layer resources, we need
to
take into account in the H.323 layer.
How does the H.323 mobility work?
Please see AT&T contributions - APC-1651 provided in the Red Bank
meeting.
The 70-page contribution has proposed a complete solution for H.323 mobility. There are contributions as well.
Hope this will clarify your questions.
Best regards, Radhika
-----Original Message----- From: Chris Wayman Purvis [SMTP:cwp@ISDN-COMMS.CO.UK] Sent: Wednesday, November 24, 1999 4:31 AM To: ITU-SG16@MAILBAG.INTEL.COM Subject: Re: H.323 Annex I
All,
Might I be permitted to attempt to summarise Mr Roy's mail while
asking my
own question (my apologies to Mr Roy if I've misunderstood his point!)? I'm not an expert on Mobile IP. However, I don't understand why
there
is
anything at all involved in H.323 mobility beyond making the
statement
"In
IP networks, mobility issues are handled by using Mobile IP", and
expecting
users of other transports (IPX, native ATM etc) to make their own
arrangements.
So, the question: Why is the effort on mobility required? Obviously if this question is answered in contributions that I've
missed,
I'll be happy with a reference rather than a full explanation on the
list!
Regards, Chris -- Dr Chris Purvis -- Development Manager ISDN Communications Ltd, The Stable Block, Ronans, Chavey Down Road Winkfield Row, Berkshire. RG42 6LY ENGLAND Phone: +44 1344 899 007 Fax: +44 1344 899 001
-- Dr Chris Purvis -- Development Manager ISDN Communications Ltd, The Stable Block, Ronans, Chavey Down Road Winkfield Row, Berkshire. RG42 6LY ENGLAND Phone: +44 1344 899 007 Fax: +44 1344 899 001