The max is 2.992 Gbps at 1436/1484 bytes. It's never quite 3 Gbps as the
first Cisco cell has an extra 4 bytes of overhead. The average over all the
packet sizes is 2.82 Gbps taking this into account.
Matt.
-----Original Message-----
From: Eric Osborne [mailto:eosborne@cisco.com]
Sent: 19 July 2002 14:08
To: Chris Whyte
Cc: Majdi S. Abbas; Cisco-NSP
Subject: Re: [nsp] Gigabit Ether Channel on Cisco 12000s
On Thu, Jul 18, 2002 at 11:29:45PM -0700, Chris Whyte wrote:
> Just fyi, I worked for GSR for 3 years. I would be happy to send you
> something privately (that will most definitely convince you) if that
> helps.
>
Allow me to add that neither one of you seems to have gotten it
completly right. But feel free to send me whatever cisco internal doc
it is that you have stockpiled, if you think I've got it wrong...:)
5Gbs is indeed a valid bps number, but it's cisco math.
There are
- 5 SLIs from LC to fabric
- 1.25Gb/sec per SLI
- one SLI is an XOR copy of the others, leaving 4 SLIs
- 1.25Gb/sec per SLI doesn't take into account 8B/10B,
leaving a useful bandwidth of 1Gb/sec per SLI
- that's 4Gb/sec full-duplex
- but there's 25% overhead (16 bytes per 64b cell)
- so 3Gb full duplex of IP bandiwdth
If you want to do cisco math *properly*, just claim that 5*1.25 is
6.25Gb/s, and that that's what we can do per slot. And then claim
that since it's full duplex, each slot can do 12.5Gb/sec, which makes
a stodgy old 12008 a 100Gbs/box! (ignoring, of course, the fact that
the RP will keel over long before it can get around to doing that much
switching).
The IP throughput is a max of 3Gb/sec, full-duplex.
The reason the 3xGE cards don't do 3xGe line rate is partly
because E2 can't switch that fast (4Mpps at 3Gbs is about 94-byte
packets at line rate) and also because of cell tax, particularly on IP
packets whose lengths is not a multiple of 48.
eric
> Thanks,
>
> Chris
>
> > -----Original Message-----
> > From: Majdi S. Abbas [mailto:msa@samurai.sfo.dead-dog.com]
> > Sent: Thursday, July 18, 2002 9:38 PM
> > To: Chris Whyte
> > Cc: Cisco-NSP
> > Subject: Re: [nsp] Gigabit Ether Channel on Cisco 12000s
> >
> >
> > On Thu, Jul 18, 2002 at 02:12:05PM -0700, Chris Whyte wrote:
> > > Btw, it's 5Gbps full-duplex, *not* 2.5Gbps full-duplex.
> >
> > No, it's not:
> >
> > "2.5 Gbps switch fabric (80 Gbps switching system
> > bandwidth) - This consists of the GSR16/80-CSC and the
> > GSR16/80-SFC fabric set. Each SFC or CSC card provides a
> > 2.5 Gbps full-duplex connection to each line card in the
> > system. For a Cisco 12016 with 16 line cards, each with
> > 2 x 2.5 Gbps capacity (full duplex), the system switching
> > bandwidth is 16 x 5 Gbps = 80 Gbps."
> >
> > Note also that:
> >
> > "On 120xx models, each 2.5 Gbps chassis slot has up
> > to four 1.25 Gbps serial line connections, one to each of
> > the switch fabric cards to provide a total capacity of 5 Gbps
> > per slot or 2.5 Gbps full duplex."
> >
> > The upgraded fabric in the 124xx series is 10 Gbps
> > full-duplex, so you will not find any 12xxx model with 5
> > Gbps full-duplex to a slot.
> >
> > > The docs don't tell the whole story (as usual). :-)
> >
> > They seem pretty clear to me, Chris.
> >
> > --msa
> >
------------------------------------------------------------------------------
Live Life in Broadband
www.telewest.co.uk
The information transmitted is intended only for the person or entity to which it is addressed and may contain confidential and/or privileged material.
Statements and opinions expressed in this e-mail may not represent those of the company. Any review, retransmission, dissemination or other use of, or taking of any action in reliance upon, this information by persons or entities other than the intended recipient is prohibited. If you received this in error, please contact the sender immediately and delete the material from any computer.
==============================================================================
This archive was generated by hypermail 2b29 : Sun Aug 04 2002 - 04:13:50 EDT