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 (19660)	12/9/1998 12:28:00 AM
From: trendmastr	Respond to of 29386

from Yahoo <<Zacks Upgrade by: gcfats 1411 of 1412I just checked Zacks the other day, & they have a buy on the stock, with the report dated 11/30 I believe. They previously had a hold. I think we'll see ANCR in the dbl digits again if they can close an OEM soon. Posted: Dec 8 1998 6:47PM EST as a reply to: Msg 1408 by pallaura>> I can't seem to find this....can anyone else? tm

To: Craig Stevenson who wrote (19660)	12/9/1998 1:55:00 AM
From: Roy Sardina	Read Replies (2) \| Respond to of 29386

Ok Guys let's try this Fabric discussion again.... Switch latency is the time it takes from the beginning of the packet entering a switch until the beginning of that packet leaves the proper port (route) of the switch. In a memory based switch this is the time t takes the memory pointer to be directed ot the proper port and that port to begin extracting the packet from the memory (therefore a function of memory speed and switching (routing) speed) Lets not confuse that with switch throughput which is the maximum amount of traffic that a switch can pass through it's ports (this is typically a function of the memory bandwidth in a memory based switch) It can also be dependent on backplane bandwidth in switches with blades and backplanes. In a sixteen port FC switch, 16 Gigabits is the max throughput needed since eight ports would be input and eight would be output. Now in a Fabric you put switches in cascade (E-Port) and you choose the number of hops that will be allowed in the fabric (for example 3 is the number that comes to mind) using 16 port switches that's a couple of hundred usable ports. That means a 2 microsecond latency gives you 8 microseconds of latency in the fabric. (three hops means there are 4 switches end to end). NOw the real bottleneck in these fabrics is the single Gigabit link between the switches because if several ports on a swtich want to send data from that switch to another across the fabric then the link between them because the slowest point. That now becomes a blocking fabric. The way to combat this is to "gang" the number e-ports between switches, but this increase in cost can be high for some fabrics. The longterm solution is either higher port count switches (read high speed backplane and expensive)or e-ports that are 2, 4, or 8 Gigabit to give added bandwidth. hope this clears this up Roy