home *** CD-ROM | disk | FTP | other *** search
- From: ajw@otter.hpl.hp.com (Anthony Wiley)
- Date: Fri, 4 Sep 1992 10:17:45 GMT
- Subject: Re: AAL 5 query
- Message-ID: <68780001@otter.hpl.hp.com>
- Organization: Hewlett-Packard Laboratories, Bristol, UK.
- Path: sparky!uunet!spool.mu.edu!sdd.hp.com!scd.hp.com!hplextra!otter.hpl.hp.com!otter!ajw
- Newsgroups: comp.dcom.cell-relay
- References: <1992Sep2.121122.15112@ccsun.strath.ac.uk>
- Lines: 46
-
- In comp.dcom.cell-relay, craa85@ccsun.strath.ac.uk ( D.W.Stevenson) writes:
-
- > Can somebody perhaps clarify my understanding of certain parts of AAL5:
-
- > I'm looking at the situation where a host requires to have more than one
- > VC open at one time. In this case, one AAL entity is required for each
- > connection. ATM-SDUs from each AAL are multiplexed/demux'd by a single
- > ATM layer entity in the interface. This gives the following structure for
- > the interface:
-
- > ATM physical connection
- > |
- > |
- > ATM entity
- > | | |
- > AAL AAL AAL
-
- > CPU
-
-
-
- > In order to transfer information across a virtual connection, the user layer
- > passes a MAAL-CREATE.request primitive (with the far end ATM_CEI and
- > AAL_CEI as parameters) to initialise new AAL state machines for a new AAL
- > entity.
-
- Why are the far end connection identifiers specified here?
- And, is this explicity documented?
-
- > There is no primitive that is passed from the MAAL (or AAL) entity to the ATM
- > layer which says "a new connection is being created, please assign a new
- > VCI/VPI pair for use with this AAL and associate cells to/from this
- > VCI/VPI pair with this newly created AAL entity.".
-
- > As far as I can see, this means that the ATM layer is unable to demultiplex
- > cells to the correct AAL entity, since firstly, the ATM layer doesn't know
- > which VCI/VPI corresponds to which AAL and secondly, there is no parameter
- > in the ATM-DATA.indicate primitive to identify a unique AAL entity to
- > which each ATM-SDU should be passed.
-
- True, however
- I've always assumed (although one should NOT assume anything here) that a
- VPI/VCI was automagically associated with an AAL_CEI in the ATM entity,
- and hence this is how (de)multiplexing was done.
-
- Tony Wiley.
-