It's Time the PSTN Grew Up

As I write this, five days after Hurricane Sandy dealt a devastating blow to the east coast, there are millions that cannot read it because they still have no power. The effects of the storm will be felt for years to come. And in the wake of Sandy there will be many lessons to be learned. One of those lessons is that it is finally time that the Public Switched Telephony Network (PSTN) grew up.

Back in the 1990's when the FCC granted competition in the local telephone market (in exchange for local carriers' ability to get into the long distance game) it became necessary to devise a way to know where to route a "ported" local number. Thus the notion of local number portability (LNP) was born. And to achieve this, a ported number would use a Local Routing Number (LRN) which is used to uniquely identify on which switch a number resided, and therefore where to route a call. Hurricane Sandy magnified the weakness in this approach.

One of the largest examples of such a failure was a Verizon switch in lower Manhattan that lost power due to flooding. With the switch, a Nortel CS2K (which is essentially a DMS with an IP head end), out of commission for several days so were the thousands of numbers that were served by it. The PSTN was designed to allow multiple routes, or trunks, to get to any given switch in order to give redundancy to the weakest link in its chain: the transport. And within the switch itself all hardware is redundant. But it still assumes that switch will always be up. An epic single point of failure.

In an IP world it has become possible for a number to be hosted virtually anywhere; you can literally serve a number across the globe from a server in your basement. Smaller competitive phone providers have been serving distant locations (probably not across the globe, but certainly across state lines) since LNP began. And Internet Telephony Service Providers (ITSP) have extended this reach. In many - perhaps most - cases these ITSP's will still home a number from a single location. Some redundant Voice over IP (VoIP) architectures allow for geographic redundancy; that is, that single point of failure may now be spread over thousands of miles, or truly all over the world. To take it a step further, large scale IP Multimedia Subsystem (IMS) networks allow an even greater level of geographic redundancy and roaming. Though they still rely on a centralized IMS core, that core can be geographically disbursed.

Verizon and the other large carriers have already invested in such VoIP architectures. So, how could such a failure occur? Why are we still being served by switches that were designed in the 80's and refitted for LNP in the 90's? The answer is simple. They want to squeeze as much out of their existing investment as possible. I am not saying it is easy to migrate to the new architecture or that it can happen over night. But it can and should happen.

It's time.