Silicon Investor (SI) -- The First Internet Community

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) -- Ignore unavailable to you. Want to Upgrade?

To: Craig Stevenson who wrote (18333)	10/1/1998 10:35:00 AM
From: George Dawson	Read Replies (1) \| Respond to of 29386

Craig, Interesting but confusing article. If the PCI-X is a faster I/O bus it would seem like a positive for any connected I/O device. The problem of course is that speed has characteristically increased about 100 fold between the CPU and the I/O bus. This is probably what the Intel engineer is referring to with the "wait faster..." comment. I think we went through that exercise last summer when we looked at how fast disk arrays could read/write data and concluded that they were a lot slower than a Gbps. The comment about switched fabrics in the future is interesting and I wonder what Intel is up to. There are several proprietary clustering schemes that depend on low-latency routers. Some of them get around the I/O bus speed problem by attaching to the main system bus. Makes me wonder if Intel is getting into FC fabrics? George

To: Craig Stevenson who wrote (18333)	10/2/1998 7:20:00 AM
From: KJ. Moy	Read Replies (2) \| Respond to of 29386

Craig, Sorry for the late reply. I've been busy. My read on this PCI-X architecture is that it is trying to eliminate the conventional bus structure. Datas move within chips, minimum or no storing in a 'bus', thus makes it super fast. It wants 'fabric' implementation. I believe they are talking about switch/fabric in a generic term, not necessarily FC. Will it post a threat to other fabric manufacturer such as FC? I don't think so. FC was designed to be a hybrid channel/network. It allows distance, more devices to be connected via channel or a 'FC NET'. What Intel/IBM/HWP are talking about here is an infant product, no definition of how it would interface with other products. In other words, there are no standards. It focus on a design which would accelerate the channel speed and channel speed only. I think I read a comment from the article, 'faster to wait' . Look how long FC has come along. Even with a well defined standard, many twist and turn, existing product supports, private loop, public loop, fabric, pt to pt. Private loop devices were not supposed to participate in 'fabric login', somehow companies manage to force switch vendor to put in supports for them for legitimate reasons. Bottom line, this new design may be for some specific applications yet to be defined. KJ