To: Reginald Middleton who wrote (24 ) 12/2/1999 6:14:00 PM From: Frank A. Coluccio Read Replies (2) | Respond to of 68
I'm not sure I'm the one who suggested this, but let's run with it:"..but to locate them on the customer's premises sort of defeats the whole purpose, doesn't it? " I'll assume that by 'customer' you mean enterprise customers. Then it would depend on the size of the enterprise and the reach and makeup of its network. Some of my clients' intranets rival those of some large international ISPs. So, in these cases, or in those instances where the enterprise has distributed data centers, the answer would be 'not always.' But for smaller organizations, or for those whose operations are highly centralized, you make a very good point. However, even yielding on this point runs counter to some assumptions which are based on beliefs that vast supplies of cheap, or near free, bandwidth are soon to be upon us, and that the prior limitations which were imposed by expensive bandwidth are soon to be things of the past. This is germane to the topic at hand because bandwidth can be seen as an enabler of distributed computing, but not necessarily for optimizing the siting of host computing (server) locations, as would be the case with ISP pops or distributed colos. And we should acknowledge that, in the past, the "distance surcharges" that carriers exacted from long haul services was the primary motivation behind doing optimal siting of data centers. To cover the nation, for example, might have required placing host sites in ten or twelve regional office locations. In contrast, if bandwidth is relatively cheap, then it makes more sense to backhaul everything to two or three sites, instead... or back to the enterprise with a fallback in a disaster recovery site. Some good food for thought.