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* Frank Coluccio Technology Forum -- Ignore unavailable to you. Want to Upgrade?

To: aladin who wrote (34710)	7/31/2010 10:01:04 AM
From: ftth	Read Replies (1) \| Respond to of 46821

Hi aladin, re: "Among other things I design network systems for brokerages. Their focus is on absolute minimums for latency and jitter. This gets extremely anal - some of the colo's in NY are taken to the level of having all the cabling the same length to equalize inter and intra rack latency." When I read that, it indirectly implies they have squeezed every nanosecond of "protocol latency" (i.e. delays attributable to the implementation of the communications protocols used) and are down to the grand finale of tweaking latency due to physics. So for example, how do they minimize the processing time (for lack of a better catch-all term) of encrypting, digitally signing, authenticating, and non-repudiation of each of these transactions? With so many different implementations of those factors it seems it might be more fruitful (i.e. better latency-reduction payback) to tweak and squeeze all those factors than latency due to physics (e.g. cable lengths). Plus, once you get to physics and propagation velocities, optical fiber is the "slowest" (compared to copper-based cables and to wireless). If memory serves, there's about a 30% difference between those 3 mediums...something like 68% to 98% of c.

To: aladin who wrote (34710)	7/31/2010 10:03:09 AM
From: Frank A. Coluccio	Read Replies (1) \| Respond to of 46821

aladin, I can see where a one-size-fits-all 'best practices' cult might have an influence here, but doesn't equalizing all cable lengths play adversely to the sum effect of all latencies, given the familiar delay skew dynamic of slowing down all links in order to be as slow as the slowest one? Are you suggesting that time-varying delay and jitter are attributable to different cable lengths? I hardly think so, but just wondering. Ordinarily, delay and jitter (where jitter can be discussed in many different contexts) result from time-varying properties of inter-networks caused by unpredictable buffering delays, the use of multiple routes possessing different route costs during packet flight, etc. I could see the cable lengths having more weight if precise synchronization among all elements in the rack assemblies is critical. Is that the case? Or is it, as you noted, merely a manifestation of an anal trait that some circles of IT are known for? TIA. FAC ------