![]()
King of The Road |
|
By Joel Conover
Flying over the city of Ma
dison, Wis., watching hordes of vehicles trying to cram their way through the small isthmus that makes up the downtown area, I was reminded of my backbone network. Packets get bottlenecked at the ingress points, struggling for bandwidth through congested routers that lead to my back-end file servers. Adding bigger pipes would help move the data faster, but putting a freeway through downtown won't defuse congestion problems unless the traffic can get on at a much faster rate. The rampant growth of the switching market has convinced many of us to flatten out our carefully sculpted tree of routed networks. This, in turn, has put enormous performance demands on the few router ports that actually serve the entire network. In some cases, there just isn't enough "oomph" to go around. The situation will only get worse.
Network Computing recently asked distributed routing vendors to submit their Big Daddy devices for competitive review in our University of Wisconsin labs. The goal of our testing was not only to stress packet-by-packet routing performance, but also to see how well these systems responded to the demands of real-world routing situations, such as those found on the University of Wisconsin backbone. To qualify for our tests, vendors had to support either OSPF or RIP v ersion 2 routing protocols. We asked each vendor to submit a multiprotocol router as well as a hardware-based IP or IPX router. In the end, we received products from Bay Networks, Extreme Networks and Foundry Networks. Bay submitted the BCN (Backbone Concentrator Node), a full-fledged multiprotocol, modular router, while Extreme and Foundry submitted hardware-based Layer 3 switching routers.
|
|
|
|
ATM Backbone Switches Hardcore ATM Switches for the WAN The Beginning Of A Brave New (Router) World IP Switching: Battle For The Network High Ground Sync or Swim? Will Your Merged Systems Float Together or Drift Into Chaos? |
|
|
![]() |
|
|


To view the Report card.











