Re: thoughts on "ideal"

[James Kempf <James.Kempf@Sun.COM>]

Hi Tim,

>there are a few places where the requirements draft points to some "ideal"
>case, it seems prior to some justification.
>
>for example
>"The ideal IP access infrastructure is one where the RAN is an IP network
>where
>base stations, base station controllers, and mobile terminals communicate
>via IP
>protocols. "
>
>I would stop far short of calling this ideal; similar initiatives in the DSL
>environment decided upon an ATM/l2tp access infrastructure; in general the
>access networks for internet services seem to be layer 2.  So I can
>understand that IP in the RAN is an interesting and may yet to be proven
>ideal, but i'd be happier with a requirements draft that doesn't make these
>judgements.  maybe its a candidate?
>

Having been the editor of the MWIF IP RAN documents, I agree that IP RAN is 
really an L2 issue, since application packets are not seen on the RAN. It is 
more like the PWE3 WG in IETF, Pseudo-Wires Emulation Edge to Edge, but they are 
focusing on ATM over IP. IP RAN is wCDMA or cdma2000 over IP. 

Short of changing the radio protocol to one where there is a better
match between the L2 and IP (like 802.11 for example) this is the
best you can do.

I'm not sure whether the requirements draft wants to get into this
particular issue or not. Most of the draft is talking about IP *service*
to the mobile, presumably also for voice and other multimedia data.
IP RAN is more about IP infrastructure, unrelated to IP service.

>outside of the MWIF technical reports has anyone seen a justification for IP
>in radio networks as proposed.
>

3GPP has a study group and technical report in the area, TR 25933-110.
There is also some ongoing work in 3GPP2, on IP as transport, and 
a rearchitecture study completed last month.



		jak