[c-nsp] Something I was thinking about whilst idle the other day.

[Saku Ytti]

On (2008-03-19 14:20 -0400), Drew Weaver wrote:

>                 What are some persistent things about Cisco products that no matter how high into the product line you travel you cannot get away from? I was sitting around the other day thinking about how odd it is that in 2008 the BGP scanner still causes the CPU utilization to jump ridiculously high (on pretty much all routers I've seen..from 7200 to 128xx), and all of the various limitations of the route table sizes in various Cisco products. I realize these issues are either harmless or explainable ("just the way it is") I just think it is a strange/interesting thing to note that years later the same issues are still present in technology no matter how far up you go in the line.

Actually BGP scanner is not needed any more in late 12.2S images. And yes,
I'd consider it broken design that it ever was required. BGP scanner
is still there and does different things and is not CPU intensive.
 One other thing that to me seems odd thing J and C are pushing as
cool new feature, is prefix-independent-convergency, granted I 
have 0 clues on hardware design, but I couldn't imagine I wouldn't
want to have FIB with indrection to next-hop on my first attempt
to design FIB, just seems so obvious.

-- 
  ++ytti