] They still I think leave an interesting question. when you have IP
] networks as large as sprints and mci's and you are having a lot of
] serious problems, if this frame relay architecture is not necessarily the
] answer, then what is?
I'm not sure there is _the_ answer. Right now many solutions are
'working', some better than others, but w/ tradeoffs both ways.
] IE how are you going to avoid the crash into the brick
] wall at 60 mph? One way is to cidrize like crazy.....ie aggregate
] routes, clean up the swamp, etc. Any other solutions out there? From
] what I can remember everytime someone suggested a solution someone else
] showed that said solution produced an undesirable side effect.
"Everything that can be invented has been invented."
--Charles H. Duell, Commissioner, U.S. Office of Patents, 1899.
The issues here are quite simple:
Large Pipes - used to be T1s, then T3, now we're thinking
about OC3 or 12.
Fast Routers - Speed needed is a function of several variables:
- speed of traffic (see above)
- number of routes (function of Inet routing
table)
I'd be curious to hear the following:
1) Can we 'evolve' to the next level w/out going to a
switched network backbone?
* we = Internet type folx
* backbone = The "Biggies" backbone (ANS, MCI,
Sprint, etc...)
2) What is being done w/in our favorite vendor (Cisco) to
provide products capable of the next order of magnitude
of power required for Backbone routers?
Mention is made of the switched folks, what are they
using? Is it reliable (enough)? Is it cost-effective?
-alan