The New QLogic (ANCR) Message Board

STOCKTALK

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)

QLGC 16.07

0.0%

Aug 24 5:00 PM EST

Public Reply Prvt Reply Mark as Last Read File

Previous 10 Next 10 Previous Next

To: Greg Hull who wrote (12156)	11/2/1997 5:04:00 PM
From: George Dawson	Read Replies (1) of 29386

Greg, I think I know what you are getting at, but I'm not sure the data is publicly available. If you look at texts on networking that use ATM as the predominant technology (like *Data Network Design* by Darren L. Spohn - you encounter a concept called Utilization. This is the fraction of a link's capacity that is used over a long period of time. You can use queuing theory to calculate crossover bandwidths for various file sizes at various percentages of link utilization. The point where the time to transfer the burst exactly equals the propagation delay is defined as the latency/bandwidth crossover point. (see p. 696 of Spohn's text). The practical significance is that when the link utilization is less than the crossover bandwidth the transmission is bandwidth limited. That is the bandwidth is limiting the transmitter from releasing the entire transmission. When the link utilization is greater than the crossover bandwidth point the transmission is latency limited, where the latency response of the receiver limits further transmission. I do not have the data for FC links, but I expect them to be significantly different snce the ATM data is for long distances and ms ranges of latency. The other thing to consider, that both Craig and I have confirmed is that class 1 service essentially gives you point-to-point latency figures (< 100 ns range)and that when a class 1 connection is made all of the latency is in the electronics, since the routing instructions are read from the first frame only. I will continue to look for similar FC information, but I expect that the server characteristics including the bus, buffer and processor will be the major limiting factors if you want to get internet 3 type speeds to your desktop. It took me a while to get this concept. One of the top experts in FC told me many times that latency has nothing to do with bandwidth. I finally believe him.

Report TOU Violation

Share This Post

Public Reply Prvt Reply Mark as Last Read File

Previous 10 Next 10 Previous Next