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

Richard A Steenbergen ras at e-gerbil.net
Thu Sep 28 20:31:20 EDT 2006


On Thu, Sep 28, 2006 at 09:28:29AM -0400, 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.

At the risk of not being on the "we need more bandwidth now!!!" bandwagon, 
I don't understand why people think 100GE is so critically important, and 
why 40GE "just won't do at all". Parallel Nx10GE paths is a perfectly 
viable way to scale a network given the commodity technology currently 
available. Even given the 8 member limit that most vendors stop at 
currently, can you honestly tell me you have links where you need more 
than 80Gbps of capacity, and where a sensible architectre wouldn't call 
for adding a diverse path, another trunk, or another router anyways?

This isn't so much directed to the original poster as some of the people I 
see running around screaming the sky is falling because we don't have 
100GE. 100G technology doesn't exist in a commercially viable way right 
now, nor do we really need it to. This is especially true in an IP network 
where you can combine load balancing mechanisms to scale even further (for 
example, ECMP load balancing between multiple 802.3ad trunks), we're in no 
near term danger of not being able to handle reasonable traffic loads with 
a reasonable number of devices. The people who really have a hard time of 
it are te ones who are fundamentally screwed by their choice of 
architecture, the flat L2 networks (like say AMS-IX), where you're pretty 
much forced into the worst of all worlds with trunk size requirements and 
trunking overhead. But even here, when they say "we can't scale because 
we're topping out at 8x10GE trunks", this is a vendor implementation issue 
which is FAR easier to solve by extending the number of trunks than 
waiving the magic "we'd like non-existant 100g technology" wand.

Plus, the near convergence of SONET's traditional 4x upgrades and 
Ethernet's traditional 10x upgrades at OC192/10GE should have been a great 
wakeup call to everyone involved about the benefits of standardization. 
Standardized optics (XFPs) are a great win for networks, as is 10GE WAN 
PHY compatibility for transport. It seems utterly silly to me that anyone 
would want to throw this away on the basis of "but we've always done 10x 
jumps before, next comes 1TbE yes?". 40G technology is real and here 
today, if you really need to handle more bandwidth on a larger pipe this 
is the technology you should be pushing for, since it is quickly and 
easily achievable. And don't tell me "40GE is not enough", because an 8x 
trunk with 320Gbps is a far cry from 80Gbps. 

Bandwidth hyperbole is great fun and all, but it leads to poor decision 
making, inefficient spending, and wasted effort on a global scale. 

-- 
Richard A Steenbergen <ras at e-gerbil.net>       http://www.e-gerbil.net/ras
GPG Key ID: 0xF8B12CBC (7535 7F59 8204 ED1F CC1C 53AF 4C41 5ECA F8B1 2CBC)


More information about the cisco-nsp mailing list