Jay, those are all very good questions. It would be helpful to know more about the specific architecture they plan to use.
This subject calls to mind the value proposition effected by the economies of scale in Cogent's planned offerings, as well. Cogent, you might recall, is offering to deliver 100 Mb/s, unobstructed (read: dedicated), to enterprise end users for $1,000 per month, across the board. This, they say, will match the speeds of their corporate LANs and cause the Internet cloud to seamlessly interface with their LANs, and look like it was merely a server on their network. That remains to be seen. We should speak more about Cogent in another sub-thread in this regard, under its own heading.
Getting back to Blockbuster, if you think about the Enron network in the same context as inter-exchange carriers voice clouds --which, by hook or by crook, wind up having a presence in each of the, or most, ILECs'/CLECs' POPs-- then it's easier to envisage an architecture as one where their traffic management issues are separate from those of the ILEC's or CLEC's "other local traffic" loads, hence, free of congestion from same.
Under the heading of speculation: Here, one could also envisage the ENE cloud mapping streams ATM virtual connections directly to the DSLAMs [Router port to DSLAM, using ATM], as opposed to subjecting those streams to "best-effort" scenarios from the edge to the core amidst other traffic flows headed for the 'Net.
I should clarify that the DSLAMs in most of the ILECs --if not all of them-- that were mentioned in yesterday's release by ENE & Blockbuster (didn't see any mention of Telus in that one) are virtual dedicated links to end users by virtue of their use of ATM PVCs.
This stands in contrast to other forms of DSL aggregation devices (collision domain-oriented such as those which some MDU UTROs are using) and Cable Modem approaches, which are entirely contention based. And where some cable modem approaches are slightly different than pure contention based, they nonetheless vye for the same downstream bandwidth even though there may be arbitration tricks taking place somewhere in the head end. But in the end, where bandwidth sharing is used, the conservation of matter and energy hold true. If it ain't there, you can't have it.
Now, let me balance what I've just stated about Cable Modem by also offering that VoD and certain other forms of switched digital video and streaming content will be available over HFC systems independent of the cable modem architecture which uses CMTS (cable modem termination sytem) elements at the head end. Instead, these services will be available via the set top box under a separate operating system. And of course, these new channels that will be accessible through the set top box will not be sharing bandwidth with cable modem spectrum.
What remains to be seen --and this will probably fall under the heading of content title horsetrading-- is whether cable modem subscribers who have enhanced set top boxes can access the same services as those offered by the ILECs in combination with Blockbuster.
I do believe that this video distribution application, along with gaming and some other multimedia apps yet to be announced, may just be the much-awaited class of "killer apps" that will drive dsl into hyper-acceptance, at first, and then into periods which could become shaky, later on. The latter would hold true if the ILECs don't adequately anticipate demand (something they usually don't do a good job at) and bolster their transport networks. This assumes, of course, that transport from each CO housing DSLAMs to the ENE POPs will be supported by the ILEC. The situation might be, however, that ENE comes directly to the CO with fiber and colos with the ILEC with whom they are partnering, in most cases. But given the number of COs (in the tens of thousands in the ILECs mentioned), they will probably lease ATM pipes from the ILECs, instead.
Since they "are" partnering with the ILECs, in fact, there exist a number of mutual incentives -- at least one would imagine this to be the case -- for them to optimize the transport situation, one way the other, to sufficiently handle the increased traffic loads between the edge and the streaming content servers, upstream.
Another gating factor, of course, would be the aggregate switching speeds on the backplanes of the DSLAMs that are used. Some of them will "block" traffic when every port on the subscriber side of the box is operating at full throttle.
If anyone comes across a detailed depiction of this proposed architecture, please post.
FAC |
