Roy,
I see what you mean. In essence, an 8-port switch lowers the cost of entry into a switched environment over a 16-port switch, just because there are less ports at x dollars per port. Reducing the feature count would also help to reduce the overall cost of the unit, even factoring in overhead costs from the case, power supply, etc.
I've always wondered why nobody makes some sort of Fibre Channel aggregating product, similar to a Fast Ethernet switch with a Gigabit Ethernet uplink. It would seem to me that a lot of companies could go with dedicated 1/4 gig to the workstations and/or server/storage units, as long as the pipe to the other side of the network was big enough. As companies migrate to native IP networks, Fibre Channel actually makes more sense than some of the other technologies, and it certainly is faster than Fast Ethernet, even at 1/4 gig speeds.
Is switch saturation not a problem even with Sequent type applications? Perhaps the lure of a larger scale switch is simply because of the reduced wiring complexity, or perhaps a cost advantage over multiple switches.
At what point do you guys start to see saturation on an Arbitrated Loop? I realize that this will vary widely, based on the traffic patterns, but I'm trying to get a handle on how many devices can realistically be attached to a single loop before it bottlenecks.
Craig
P.S. I do appreciate you taking the time to answer our questions. Having worked for Brocade, and now for G2, you are as close to an industry insider as we can find. |
