Hello Dan,
Agreed, most of the discrete applications covered by the article are not earth-shatteringly new. The virtual attachment to the corporate PBX, as you mention, needs to be worked out by next gen CLECs, but even those rudiments are already in place with existing gateways that are on the market.
The mere fact that you are or are not able to attach to the enterprise LAN at this time wasn't even my point, as much as whether or not the CableCo/ISP (in your case, ATHM) would even allow the sort of traffic that was discussed. Granted, you may be able to do data base lookups, exchange files and so on, but are these types of transactions sanctioned by ATHM as a permissible utilization of their resources, in accordance with their present and future use policies?
Given even a small segment size, say, of 100 users, if a nominal percentage of users were pulling down, or worse, sending up, large files, they could (and have actually been found to actually) jam up the works, as things stand today. I think that this issue is more salient than whether or not it works for individual users.
What you say about many of today's users living close to their places of work is not necessarily true, nor is it what I have found to be the case, especially in densely populated areas, where urban and suburban sprawls are taking place.
Increasingly, new housing accommodations are making commuters second job that of traveling to and from, each day. This usually equates to traveling outside one's own head end serving area, compounding the hand off situation to include not only ATHM's intranet, but some other ISP's services, as well. More importantly, however, is the contamination that this introduces to a QoS scheme once disparate service providers need to work with one another in tandem.
It's interesting to note that the trial in question employs Com21 (CMTO) cable modem gear, which I surmise is due to their QoS attributes which are dependent on ATM. Of course, this begins and ends with the cable modem link itself, and doesn't get carried over to the secondary routes upstream.
Instead, from the head end to the ultimate end point, it takes a best effort approach once it leaves the head end, unless some form of Layer 2 tunneling is used, a la, VPN architecture. And while some means of L2TP is doable, it transcends what the trials are all about, using IIXC's local employees in Austin as the control group.
Netting it out, then:
(1) yes these are not groundbreaking ideas. They can for the most part be done today with some minor adaptations of existing wares. In the case of extended reach to remote corporates, additional measures will need to be taken into account;
(2) Cable Cos and their associated ISPs, however, may not permit these forms of SOHO and other telecommuting applications, for reasons having to do with fair allocations of available bandwidth, and
(3) the trials are not representative of most topologies that would be needed by the greater population represented by the potential telecommuting user community. Instead, these applications stand their best chance of succeeding for those (like you stated) who are residing in the same general locale as the enterprises where they are employed.
There are established networking methods that could be employed to get around all of these constraints, even going as far as having the enterprise establish a separate virtual link overlays from their intranets to the head ends. But at what price, both financially and administratively? And are these likely to occur?
Thanks for your reply, and further comments are welcome.
Regards, Frank Coluccio |
