Cell Relay Archive

Cell Relay Retreat>List Archive>month:1996-Dec> msg00051



[Date Prev][Date Next][Thread Prev][Thread Next]  
  [Date Index][Thread Index][Author Index][Subject Index]

Re: assured service with AAL5

  • From: Jim Jackson <jj@scs.leeds.ac.uk>
  • Date: Thu, 5 Dec 1996 09:58:32 +0000 (GMT)


On Wed, 4 Dec 1996, Neil Harris wrote:

> In article <199612040944.KAA19311@enst.enst.fr>, sassos@res.enst.fr (Saso
> STOJANOVSKI) wrote:
> 
> > somebody asked about providing assured service over AAL5 in the user plane. 
> > 
> > you should look for the SSCOP protocol defined in Q.2110.
> > 
> > no need for TCP as somebody else suggested.
> > 
> > 
> >                          Sasso STOJANOVSKI
> 
> SSCOP, TCP, it doesn't really matter what your higher level protocol
> is, it just matters that you have one.  

Not true. Higher level protocols can serious limit your 
throughput. TCP behaviour and tuning is reasonably well understood and 
can deliver line speeds on 155M links (less all the overheads etc) given 
a good implementation, end hosts, and correct tuning.

However I don't know much about SSCOP, so maybe those who do could 
indicate whether it too can deliver throughput and what it's behaviour 
will be like in congestion - does the retry strategy make the congestion 
better or worse etc etc.

Jim Jackson