Cell Relay Archive

[Jim Jackson <jj@scs.leeds.ac.uk>]

UKERNA have done an interim ATM addressing scheme see

          URL: http://www.ed.ac.uk/~george/ukac-index.html


cheers
Jim Jackson

On Wed, 18 Dec 1996, Andy Leigh wrote:

> Dear all,
> 
> the time has come...
> 
> We really need to start tying down an ATM Address structure for our
> organisation (c 25k employees, UK-based, no private links leaving the
> country) and I'm looking into what's needed. I've searched the web and
> found a few references, and it seems that everyone feels as I do,
> namely that something has to be done, but breaking too soon could
> cause problems later on. However, we have to do something fast,
> considering the amount of ATM popping up all over.
> 
> My thinking is this: It's a private network, so don't use E.164. We
> might need multiple Telcos to supply our WAN needs, so don't use NSAP
> E.164. This just leaves DCC and ICD. Given that we're very much
> UK-based, we don't really need an International identifier (although,
> we _are_ well known internationally). Which leaves me with the DCC
> NSAP-type address. 
> 
> My question is this: What is everyone else doing? And, what is the
> most popular system today? (I think I was told that the ICD was the
> most popular)
> 
> Also, the High Order DSP can be broken up into subnets, to allow
> hierarchical information hiding and aggregation (just like IP). I've
> not seen any references on how to go about this. Are there rules, like
> IP (RFC1121). Specifically, how is the mask confifigured? Do the bits
> need to be contiguous? Are any numbers illegal (like IP, where
> network, subnet and host addresses must not be all 1s or all 0s)?
> 
> Any advice or experience you have would be gratefully received! If you
> haven't thought about this before, now's youe time to dig into the
> debate.
> 
> 
> Andy
> 
> 
>