[c-nsp] new 8-port 10 G bade

[Jared Mauch]

On Thu, Sep 28, 2006 at 06:12:27PM +0200, Arnold Nipper wrote:
> On 28.09.2006 16:32 Justin M. Streiner wrote
> 
> > On Thu, 28 Sep 2006, Jared Mauch wrote:
> > 
> >> I think it's critically important for folks to continue to ask 
> >> Cisco (and your other vendors) for a 100GE solution.  IEEE process
> >> is a bit odd, so could possibly stall.  If that's the case, us
> >> "nsp-types" may need to band together and have a SP nonstandard
> >> 100GE bakeoff.
> > 
> > The issue I have with that is that if we get too far ahead of
> > (ourselves, IEEE, $other_standards_body), the risk of having more
> > neat vendor-specific compatibility and interoperability issues down
> > the road grows dramatically.  There's already a pretty well
> > established track record of vendors jumping the gun on standards,
> > with varying levels of headaches. ISL vs. dot1q, IETF frame
> > encapsulation vs. Cisco frame encapsulation, HSRP vs. VRRP, etc...
> > 
> > There will always be interop concerns, especially in new technology,
> > but I'd prefer not to give vendors the opportunity to grow interop
> > issues in a petri dish :)
> > 
> 
> Actually, first was ISL, then came dot1q. First was HSRP, then came
> VRRP. Wasn't it?
> 
> Esp. for 100GE I *hope* that vendors don't wait for the standard being
> finalized.

	Typically folks see "pre-standard" hardware come out prior to
things being actually available.  These usually support all
the right encoding on the "wire" but may have minor differences
between the "final" hardware.  (usually once you finish "baking"
things and find that vendor X and Y don't work perfectly together).

	 I know there's a number of places facing challenges, eg:
AMS-IX, LINX with how to deal with the lack of 100G as bundling and
hashing can only get you so far.

> Currently with Cisco 80Gbps is the maximum you get when switching ...
> that might be too little in two years from now,

	Yeah, this is my concern.  I'd like to see the ability to put
more interfaces in a bundle.  I believe the current limit on the "76k"
platform is 8, with 8 in a backup/standby state.  This of course
may need some hardware redesign work.  I don't want to seem radical or
alarmist on this but this is I think the critical issue facing SPs
now or in the near future.

	Trying how to get 100,000,000,000 distinct samples
per second of a transmitter is likely very hard (in a serialized
fashion).

	- jared

-- 
Jared Mauch  | pgp key available via finger from jared at puck.nether.net
clue++;      | http://puck.nether.net/~jared/  My statements are only mine.