jach, et al, I just have to say something at this point.
I'd like to elaborate for a moment on some notions that seem to be running wild these days with regard to the alleged death of TDM (as a function within SONET), the impending demise of SONET in general, and the assumed solitary ascendancy of IP over lambda at the expense of all other forms of transmission.
It's just not happening at this time, and it wont happen for a long time to come.
I speak with the largest interstates and locals every day from an engineering perspective, and with the sales and marketing folks from a market sizing perspective.
Their SONET platform acquisitions are running off the map (it's the only thing they're putting on all of that DWDM'ed bandwidth, where the voice carriers are concerned, and almost all but for a few of the ISPs as well !), and where they are not, the installation intervals for new services to the larger end users are beginning to suffer dangerously. And that's no way to keep large customers happy.
Carriers of all types and Tier One/Two ISPs are buying new digital ADMs and cross connects. Soon these will include increased numbers of optical (SONET- and lambda- enabled) ports on DCSs.
End users are calling for IP access to the Internet and into VPNs over larger and greater numbers of legacy fat pipes (T-3s, OC-3s and OC-12s) every day, and they can't get enough of them. There is no other way, in fact, to get most of them from here to there at this time, and one begets another, and another, etc.
Lead times for T-3s and T-1s are increasing again, to precarious levels, to where they were a couple of years ago, and larger institutions and other large users are starting to get nervous about a shortage. In some places, even the largest MSAs, there are port exhausts happening in end offices in SONET and DCS boxes that are putting major end user projects on hold. It's some serious stuff we're talking about here, when even competing carriers themselves have to start calling in markers to make ends meet.
NOW.... IP over lambda will increase dramatically, enormously, beyond normal comprehensible levels, even, but while this happens it will have very little impact in a negative way on the growth of SONET over traditional means, and on digital cross connects... in fact, perhaps just the opposite will unfold.
With more capacity in the core facilitated by improved IP over x technologies, there will be a sucking sound at the user levels of utilization towards the core... users, whose capacities do not now, nor ever will in the foreseeable future, justify DWDM, WDM, or even direct IP mapping onto anything at all.
End users, instead, will increasingly be wanting to consolidate their various and disparate feeds of dissimilar protocols and traffic types on aggregation pipes, and that means more T-3s and OC-xx's which have to be switched and mapped somewhere into the premises and into the cloud.
That means more digital cross connects. Even in cases where everything is IP, users will want to separate their feeds into logical and physical partitions for security and political reasons, hence the furtherance of channelized (sonetized) pipes. Heck, even NAPs and other forms of Internet exchanges are now using DCSs.
Again, nothing that I've said here should be construed to be an indictment against IP over lambda and DWDM. Those techs will very likely and very soon take logarithmic excursions of their own, with even greater acceleration than anything we've contemplated in the past. The point I'm making here is that during the visible horizons that exist, neither SONET nor DWDM will work towards the exclusion of the other, rather, they will serve to mutually provide fuel to one another.
I had to say this, even though I find it difficult to believe myself when I think about it. But my daily experiences and the trends I am witnessing don't lie. Comments and corrections are welcome.
Here's wishing everyone here the best of luck on Monday, and the best regards,
Frank Coluccio |
