The MPLS WG Archive

Cell Relay Retreat>MPLS WG Archive>month:2000-Oct> msg00298



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

[IP-Optical] RE: Optical link bundling. Was Re: Draft Minutes FromPittsburgh

  • From: Sanjeev Rampal <srampal@cisco.com>
  • Date: Fri, 20 Oct 2000 15:34:00 -0400
  • CC: ip-optical@lists.bell-labs.com, mpls@UU.NET

Grenville

Of course ATM switches do not look at *IP* packet
headers from the payload and my note did not say that
either ...

This was just a way of saying that cells are in fact special
"packets" and the VPI/VCI label allows the switch to
have some representation of the headers of the packets
(IP/ ethernet/ whatever) being transported in that
ATM-LSP/ VCC. This enables merging operations
etc.

In contrast optical switches work at the granularity of
optical channels/ lambdas/ STSs which are not
"mergeable"

Grenville Armitage wrote:

> aside from the usual caveats of taking analogies too far...
>
> Sanjeev Rampal wrote:
> >
> > Even if we only care about IP clients,  other issues (stemming from
> > the fact that ATM switches can peer into per-packet (cell)
> > headers,
>
> _ATM_ switches do not look inside packet headers (regardless
> of what ATM-packet hybrids you might have seen), and packets are
> not cells. Looking inside cell headers? Well, that's the point
> isn't it...
>
> > do merge operations etc)
>
> In the early days of MPLS, ATM switches couldn't all be relied
> upon to support VC merge either. The analogy to optical switches
> (despite the loud stretching sounds) is somewhat germane.
>
> cheers,
> gja