[ppml] too many variables
John Paul Morrison
jmorrison at bogomips.com
Mon Aug 13 15:45:57 EDT 2007
No, the real point is that there is a big gap between the best current
CPU technology, and what is available for sale today in high end routers
marketed at service providers.
I mean a single core, 1.2 GHz processor is quite a bit behind a 3 GHz,
multi-core processor which is available today on high end, qualified and
tested servers. And probably behind by more than 18 months too.
I think this is proof that there are a lot of other factors driving
network technology, because if there were really enough demand or market
for it, we'd see 3GHz multicore procs for sale in routers NOW, not in 18
- 24 months. Since the control plane side of the equation is at least
18-24 months behind the curve, I think this makes it very hard to argue
that BGP and Internet routing are in dire shape and are about to collapse.
Another point: how many people run their boxes at 80 or 90% CPU in well
managed production networks? Most people I know would like to see them
practically idle, with the line cards or silicon doing the work, or at
least under 50%. So if most well run networks are running boxes with
CPU's 18 months behind the Moore's law curve, and running them at 50%,
wouldn't this mean that they are running about 36 months behind?
I'd say we're not really looking at CPU bound problems.
vijay gill wrote:
> On 8/10/07, *John Paul Morrison* <jmorrison at bogomips.com
> <mailto:jmorrison at bogomips.com>> wrote:
> And yet people still say the sky is falling with respect to
> routing convergence and FIB size. Probably a better comparison
> BTW, would be with a Nintendo or Playstation, as they are MIPS and
> PowerPC based. Even the latest route processor for a decent
> peering box is only a 1.2 GHz PowerPC with 2 GB RAM (RSP720) - so
> basically an old iBook is enough for the BGP control plane load
> these days? I think this has something to do with the vendors
> giving you just enough to keep you going, but not so much that you
> delay hardware upgrades :-)
> There have been big gains in silicon for the fast switched path,
> but the route processors even on high end routers are still pretty
> low end in comparison to what's common on the average desktop.
> I would say that when control plane/processor power becomes
> critical, I would hope to see better processors inside.
> With the IETF saying that speed and forwarding path are the
> bottlenecks now, not FIB size, perhaps there just isn't enough
> load to push Core Duo processors in your routers. (If Apple can
> switch, why not Cisco?)
> I guess people are still spectacularly missing the real point. The
> point isn't that the latest generation hardware cpu du jour you can
> pick up from the local hardware store is doubling processing power
> every n months. The point is that getting them qualified, tested,
> verified, and then deployed is a non trivial task. We need to be
> substantially behind moores observation to be economically viable. I
> have some small number of route processors in my network and it is a
> major hassle to get even those few upgraded. In other words, if you
> have a network that you can upgrade the RPs on every 18 months, let me
> John Paul Morrison, CCIE 8191
-------------- next part --------------
An HTML attachment was scrubbed...
More information about the ARIN-PPML