SI
SI
discoversearch

We've detected that you're using an ad content blocking browser plug-in or feature. Ads provide a critical source of revenue to the continued operation of Silicon Investor.  We ask that you disable ad blocking while on Silicon Investor in the best interests of our community.  If you are not using an ad blocker but are still receiving this message, make sure your browser's tracking protection is set to the 'standard' level.
Technology Stocks : The New QLogic (ANCR) -- Ignore unavailable to you. Want to Upgrade?


To: George Dawson who wrote (25789)1/25/2000 3:11:00 PM
From: Douglas Nordgren  Respond to of 29386
 
George,

You bring up a good point wrt Latency. Skip Jones, Director of Planning and Technology at QLogic, extolled the benefits of running VI over Fibre Channel. VI over FC provides lowest latency communications, direct access to memory, and bypasses the system stack. I can't seem to locate my latency jottings, but I think they were in the 100s of nanoseconds.

The IB latency was characterized as currently running around ~500ns. They would like to get it down to 100ns or less, but for time to market considerations ~500ns is 'just good enough' for now. For this reason, the memory work group recommends the memory sit close to the CPU, with direct connects if housed separately from the pizza box. With VI by-passing the kernel, whether through FC or IB, clustering and network client latency shouldn't be much of a problem, and should become lower with improvements in protocol and I/O engine chips, imho.

I didn't mean to intimate that Ancor owns a patent to the words 'cross connect.' <G> The fact that they have experience with a cross connect architecture is good enough for me, regardless of whose they end up working with.

I guess one could categorize VI as a 'class 1' class of service in that it too is a point-to-point connection, but 2 tin cans on a string are 'class 1' connections too. I agree, George, not germane except in the very broadest sense.

Douglas