For starters, I think you were quoting AHhaha's words there, not mine. But that's okay. AH and I have shared [and caused each other] a fair amount of grief together before, elsewhere, on the silk roads of SI.
"Maybe I'm confused here Frank,
Forsooth! (Sorry, but just couldn't help myself.)
"Everyone including you seems to think the big problem starts at the coax leading up to the customer(primarily because of the small upstream path) Milo has stated that their (ATHM's) portion of the network is sychronous (that might not be the right word..."
The word you're searching for is "symmetrical." I am at a loss as to what Milo is celebrating here, if he's referring to the larger backbone cloud portion, which is almost always symmetrical in nature. It's such an obvious and commonplace thing, as networks go. Perhaps they shouldn't always be, due to the way traffic flows take place on intranets like ATHM's, sometimes, but most often they are, nonetheless.
"...it goes both ways equaly.. so it seems to me that ATHM's network is more ready for the next solution more than the MSO's, which everyone would like to see trashed in favor of pure fiber. "
The MSOs don't have to entirely trash their networks, far from it, up front. Rather they need to modify them (even if only for their Internet access and other data services, while leaving the video wiring in place until it dies a natural death), making use of existing fiber backbones that they now have in place, as in the "F" in H "F" C, and replacing the coaxial and RF end sections as they relate to the data portions with fiber.
T is already doing this to within 75% if what I'm suggesting in lightwire, their SLC pilot. A major difference, however, is their continued dependence on Coax to the home. They'll get past this at some point, but in T's situation, they have other concerns. And those are the "good money" that was already spent putting in yesteryear's framework, and the "good money" they'd have to throw at the same problem again. No one wants to admit that the original good money was misspent.
DOCSIS? I'm not prepared to discuss the number of options or what the tradeoffs are there at this time. It's a lot of discussions. Unless they can upgrade those systems quickly enabling multiple partitions, one for each additional ISP which is admitted to the system, which I doubt they can do very quickly at this time.
Speaking of which, the one area that the MSOs can't continue to ignore any longer is that of revenue streams which are now being chalked up as lost opportunities due to the restrictive architectures they now have in place from a plant perspective, and the self-defeating covenants they've agreed to in the charters.
My word, don't the AHTM consortium partners see this yet? REVENUE STREAMS FROM MULTIPLE, THOUSANDS, of ISPs! This is all some very weird kind of math, and a curious set of rules which dictate the nuances and rituals which must be performed for maintaining allegiance and loyalty. Granted, there are problems in getting plant primed for OA, even if they wanted to, as I've listed herein. And the work would be considerable. The odd thing appears to be, though, that they actually don't want to.
They've just lost Big Bird to the competition. Do you think they get it, now? Why do I tend to think not?
Re: Milo's statement on ATHM's backbone cloud symmetry: Even so, if the backbone he's speaking about (which is the dual OC-48 arrangement supplied by T, for the most part) is symmetrical (balanced) in nature, and the last mile section (HFC and Headend) is asymmetrical (unbalanced), then the constraints of the last mile HFC will win out in the end, anyway, and render the user flows lopsided.
But I see your point, and yes, home's backbone is better prepared than today's HFC last mile to advance to a more balanced model. Granted. So, where does Milo take this observation in his next statement? It's nothing more than a non sequitur, the way things stand now.
Re: I "...seem to think that the big problem starts at the coax... "
There are a number of problems with the existing setup, if we are talking about admitting multiple ISPs. Is that the situation we're now discussing, btw? Open Access?
If we're talking OA now, and assuming that everyone in the ATHM camp and their current partners stay put for a while and we ignore pending terminations of contracts, then I see at least three or four problems off the top of my head which exist between the head end location and the subscriber location:
1. The bandwidth bottleneck imposed by coaxial cable plant in the passed home clusters which is shared (and limited to 1 GHz maximum to satisfy all video and enhanced services (voice, interactive, hdtv, internet access, distance learning, etc.), including the analog RF HFC network elements which are costly and prone to ingress (noise). The future overcrowding issue may have been diffused somewhat, by the AOL/TWX thing that took place last week, but capacity constraints will still become a dominant issue of concern as subscriber take up and more demanding applications come onto the scene.
2. DOCSIS was never designed to permit more than one operator at the helm. More on this elsewhere in this message, below (I think).
3. Along these same lines, there is the always present possibility on shared media networks that one network's anomalies (such as a bandwidth hog taking all the bandwidth, or a jabbering modem that wont quit, or more ironically but still prevalent in some cases a network management system that goes into renegade mode and shuts "everyone" down during upgrades, routine maintenance periods, etc.,) could adversely affect another network's users usually at the worst possible times, when they both (not to mention multiple of them) share the same segment.
4. The upstream constraints you spoke of. Not only upstream, but downstream as well. Who arbitrates and ultimately decides who gets what, during disputes which center on "he was using more bandwidth than I was, and I'm paying good money, too!?"
[[5. I will only whisper the word "security." Shhh...]]
Current DOCSIS administration platforms were never designed for multiple providers' simultaneous utilization. Its a cultural thing from a time when cable franchisees never imagined that such a thing would come to pass as OA. If you read my post on this topic over in the FCTF thread yesterday, you might have noticed how it was written from a service provider-centric perspective, with one guy in charge. That's the MSO.. only One MSO, and not multiple MSOs all doting over the same plant.
The assumptions that went into DOCSIS were not dissimilar to those which would be true for designing a car: There's usually only one driver. All of which means that current Head End CMTS systems will not gracefully permit two or more masters on the same platform, much less dozens or hundreds, without some major re-working of the platform geared to permitting partiotioned access to it by each ISP, fist.
An alternative point of interconnection which the Canadians have looked at is deeper into the cloud, towards the core. But if you do it there, can you still hand over control of enhanced voice, security, etc., to those "other" ISPs without physical connections at the head end? Yes, it's possible, but those protocols and practices have not been ironed out. |
