[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: Review of draft-ietf-v6ops-nap-02.txt
Tony, do you have any comment on this?
My initial thought is to identify the things that impact scalability,
and make recommendations - perhaps in a separate document or in an
appendix to this one. This comes from Dave's comments regarding "a
gigabyte of more of memory, gigE's for interfaces, and multi-CPU
route processors." While routers most certainly exist that have this,
such routers are not found on the shelf at Fry's...
On Jun 28, 2006, at 1:19 PM, David R Oran wrote:
On Jun 28, 2006, at 4:00 PM, Fred Baker wrote:
Radia and Dave:
There is a dispute going on regarding the scalability of topology
hiding by what amounts to IS-IS level 1 routing (identifying hosts
in a multi-LAN network by their host identifier and using a common
subnet ID for the domain). Do you know of available documentation
of IS-IS level 1 routing and its tested scalability?
Nothing recent. Tony Li did some testing with IOS ISIS and host
routes about 10 years ago which as I recall scaled to about 10**4
as opposed to 10**3. This was on much wimpier boxes than we have
today and the details are lost to me in the mists of time.
My intuition is that 10**3 is very low, even with moderately high
updates rates due to host movement (Tony was looking into host
routes as a host mobility solution). It would be useful to add him
to this thread.
I'm not close to this stuff anymore so appropriately de-rate my
opinion, but 50K host routes in an ISIS area does not sound the
least bit scary to me on today's boxes with a gigabyte of more of
memory, gigE's for interfaces, and multi-CPU route processors.
As Radia knows, this question is also being looked at in IEEE
802.1, which presumes that there is only a single multi-LAN-LAN,
and that MAC addresses are used to route within it.
On Jun 28, 2006, at 12:40 PM, Margaret Wasserman wrote:
50K is an order of magnitude higher than the analysis in Alex
Zinin's presentation would seem to indicate. His presentation
indicates that flat routing will only scale to something on the
order of 1K nodes. Please feel free to check with Alex, though,
as he certainly has more understanding of IGP scaling than I do.
If this document is going to recommend doing flat routing for
topology hiding, I think that the WG needs to do the analysis to
determine that this is a valid and scalable technique.