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: pjd777 who wrote (3158)	6/30/2001 2:04:33 PM
From: Frank A. Coluccio	Read Replies (1) \| Respond to of 46821

One reason that enterprise network planners decide to limit end users to 10 Mb/s has nothing to do with the ability to feed higher rates locally to the desk. In most cases the ability to support 100 Mb/s is built in, and is only a matter of a configuration change invoked by a keystroke. Desks are instead often throttled down to 10 Mb/s, sometimes to half duplex mode, because of the strain that higher access speeds would cause on in-house lateral- and collapsed- backbones, and the costs associated with increasing the capacity of wide area backbones that would be necessitated if users (especially just a few power users) were given higher speeds. In actuality, configurations are made on the basis of need and justification, and some users do get to go full bore. But the greater number are held to a point that only satisfies normal workplace requirements, and not necessarily suitable for gaming or full motion video at NTSC quality. The gb/s access speeds mentioned in the uplinked release would probably be fine for the walled garden, for accessing servers and colocated cache, and for peer links on the same network were all that were required. Anything looking inward (from a center of the 'Net topology perspective) going to the edge and core network, however, would have to be very amply provisioned, and this is something that service providers are very wary about. And then we approach the other conundrum concerning the ability of the targeted site to deliver speeds that were in kind. Again, this in many ways revolves around the costs associated with the higher-capacity links to the upstream providers that would be needed to support the higher levels of suction that would come from neighborhoods in such a model. To put it in simple terms, it could mean the difference between supporting a multiple T1s or a T3 costing anywhere from 2K to 10K per month within a locale or region, and the higher cost alternative of supporting a fractional- or full-10 Gb/s link or greater to the upstream provider when Gb/s speeds are delivered to end users. The latter could wind up costing many times more than the cost of multiple T1s or a T3 pipe. Of course, implied in this model is the ability to burst to much higher speeds and even experience some sustained near-gb througputs, perhaps approaching 1 Gb/s, but not necessarily the ability to provide sustained throughputs over longer periods of time at that rate, due to the statistical nature of the beast. The latter reminds me of what's takint place now on Frame Relay Networks. It was once feasible to subscribe to a committed information rate (CIR) of say 32 kb/s on a 128 Kb/s pipe, and the user would normally be able to "burst" to the full 128 rate for some interval of time. Now that the backbones on those networks are experiencing more strain, folks can't as easily get that extra umph anymore, and they are hitting a wall at the CIR, so they are increasing their CIRs, and it's costing them more. I wish that I'd saved my posts and those of others from the Frame Relay Forum on Compuserve from 1995-6. They would reveal a direct hit on this point, at a time when the carriers collectively were making claims that they would increase their capacity in the cloud as needed. Yes, but as needed by whom? New age GbE delivery models, such as those that are offered by Yipes, Telseon, Cogent et al might have a softening effect on such financial trauma, if they survive and can bring such services to where they are needed. But let's not forget that we're basically talking about residential networks here, and not inner city commercial ones.