Among its many recommendations, the National Broadband Plan recommended that the FCC “encourage transitions to IP-to-IP interconnection where that is the most efficient approach.” Last week, COMPTEL advised the FCC that the way to achieve this objective is to apply the same rules to IP-to-IP interconnection as apply to interconnection on the public telephone network, aka PSTN. Unless the Commission wants to impose on IP networks the various inefficiencies, disputes and arbitrage schemes that plague the PSTN, it should decline this invitation to import the logic of calling party network pays to IP networks.
Did I really just write “calling party network pays” in a blog? Well, yeah, I suppose I did. So, unless you’re in the mood for some wonkery, now might be a good time to check out what’s going on with the Facebook-Google smackdown or check in on the discussion around whether cable-cutting is mainstream or just for dog food lovers.
But, if you are still with me…and are asking yourself, “Hank, what’s so wrong with imposing PSTN interconnection rules on IP networks? I mean, after all, it’s just a technology change, right? Did Major League Baseball get rid of the infield fly rule when lights, artificial turf, or those goofy relief pitcher cars were brought into the game?”
Well, no, but this isn’t “just a technology change.” It’s about moving to broadband, digital IP networks where voice is a data application – and a relatively low bandwidth one at that.
There are two fundamental distinctions between the PSTN and IP networks that make COMPTEL’s proposal to bring 20th century business practices to 21st networks a very bad idea. The first is the problem I alluded to above. Hardwired into PSTN interconnection is a concept known as “calling party network pays.” Under this doctrine, the network provider of the party who places a phone call is obligated to pay the network provider of the recipient of that phone call.
While one might expect that the ratio of inbound to outbound calling would be generally balanced across networks, making this right to receive revenue for “terminating” a call unimportant, it turns out that entrepreneurs have found it quite easy to offer services in ways that produces significant imbalances in the ratio of inbound-to-outbound calling. Indeed, there are some providers out there with ZERO outbound calling (parents of college students can probably relate to this phenomenon).
In the world of IP interconnection, backbone providers have never cared a whit about which party initiates a flow of packets back and forth. To the extent that they have concerns about “traffic ratios,” those focus on the relative volume of data being exchanged in each direction, and not which network pings the other first. From an IP networking perspective, voice is a symmetric application which is likely to produce generally balanced traffic flows irrespective of who picks up the phone first. It makes no sense to require IP networks to change their practices to accommodate the arbitrary convention of calling party network pays.
The second distinction between the PSTN and IP networks has to do with where networks interconnect. Because of regulation on the PSTN, interconnection occurs way out in the local exchange. IP networks interconnect much closer to the network core. As a consequence, IP interconnection occurs at comparatively fewer points and at much higher capacities.
Needless to say, IP interconnection is far more efficient – and less litigious – than its PSTN step-cousin. Assuming that the FCC wants to keep it that way, it should firmly reject COMPTEL’s invitation to redesign the broadband autobahn in order to accommodate the PSTN Edsel. In an upcoming blog, I will explain why the best way to promote IP-to-IP interconnection is to take the intercarrier compensation meter away.