I have an odd question that I'm faced with regarding routing to external
address spaces.
The senario is a large network (global) with multiple points, each POP has
tail sites that
home and default to that Ipop as a gateway of last resort. The question I'm
wrestling with
is at each pop I have a /24 range, would it be best practice to route
internally to those address
spaces versus routing through my pop to each? In otherwords, if say I have a
network thats
located in Russia and I'm attempting to get to it from North America,
wouldn't it make
more sense to route through my own WAN versus the internet? My thought is
that since
I have more control over my WAN than the internet, I can better anticipate
the results, but
one could argue that it may make more sense to use the internet from each
pop.
My thought is that since I have more control over my WAN than the
internet, I can better anticipate the results, but one could argue
that it may make more sense to use the internet from each pop.
Any comments or thoughts are appreciated,
... you forgot to say why it could make moe sense to use the internet from each pop.
Basically it much depends on your needs, your policy and your wallet.
One could argue that it might make "moe" sense as a basis of
redundancy, however my situation is more one of why or why not to
do so. Security, perhaps but since its internet facing the trust from within
is better then the trust outside, and since theres outside facing nets....
The problem is that I see both sides of the fence so to speak reagarding
this issue. The item of cost for the packets flowing on my WAN is
quite little, but more so wondering about best practices.