Cell Relay Archive

[Kari Seppanen <Kari.Seppanen@vtt.fi>]

charlesn@hn-networks.co.uk writes:

> Can anyone clear up my uncertainty as to the role of AAL2 signalling
> please..
> 
> These are my current beliefs that I want to check out:

I think that AAL2 was not developed to achieve faster and lighter
signaling procedures but to allow for efficient bandwidth usage with
low bit-rate voice codecs. Just consider, e.g. 8 kbit/s coder: it
would take 48 ms to fill up one ATM cell payload --- such delay is
intolerable. On the other hand you do not want to waste your bandwidth
using only partly filled cells.

AAL2 (or AAL composite user) solves these problems by offering
multiplexing of multiple variable sized (1 - 45/64 octets) SDUs in one
VCC. With AAL2 you can choose optimal SDU size (depending, e.g. on the
bit-rate of the codec) that provides acceptable packetization delay
and fair bandwidth efficiency.

As a side effect AAL2 happen to lessen the signaling load because
allocation and removal of AAL2 channels involve only the AAL2
connection end-points (e.g. a base station controller and a mobile
exchange). However, same kind of result can be achived with (P)VPCs
(~AAL2 connection) and VCCs (~AAL2 channel). 


-- 
kse