Last week Light Reading hosted a seminar on Session Border Controllers (SBCs), sponsored by four vendors of SBC. One of the presenters made a point that SBC could be used to support CALEA. The idea is that when a call is being established that involves a target, the SBC could force the media to flow through it and the SBC could provide a copy of the stream to the LEA. There is a small problem with this approach.
If the media is exchanged directly between the end-points, the headers in the RTP packets will contain the IP addresses of the end-points, but the intercepted calls will contain the IP address of the SBC. So a target can easily detect whether the call is being monitored or not, violating one of the CALEA requirements. It doesn’t matter whether the latency introduced by SBC is not detectable. So if SBC is used for CALEA, then the media for all the calls must be routed through it. That means, SBC and the network connectivity must be sized accordingly.
First news came from FCC that even if a call is carried on an IP network, if it originates and terminates in PSTN, then access charges will apply. Then Qwest took a press release saying that it is setting a precedent by announcing that it will not collect access charges from pure VoIP service providers. I had already anticipated FCC decision. But the action by Qwest is perplexing on two counts. Firstly, it looks like they are easily forgoing a revenue source. The stated justification is that, “it's hoping that cheaper prices will attract more business from commercial Net phone providers.” Then why not eliminate it altogether? Be that as it may, the question I have is how are they going to determine the nature of the origin? For example, AT&T CallVantage will be eligible for access charge relief whereas AT&T LDS will not be. But how will Qwest determine a call that AT&T delivers is from CallVantage and not LDS?
I have written earlier on what aspects of VoIP must be, needs to be and could be regulated. Recently this topic is again in the news. CRTC has issued their preliminary ruling; there are stories about FCC may be getting ready to announce their decision on AT&T petition; bills have been introduced both in the Senate and the Congress regarding VoIP. This entry is an analysis of these developments.
It is interesting to compare and contrast the “preliminary view” from CRTC and the NPRM from FCC. Whereas the NPRM is lengthy and has posed many open questions explicitly soliciting opinions from the public, CRTC is very concise and states pointed conclusions. But there is a remarkable agreement (there are other opinions) in the philosophical thinking and the anticipated regulation. But in some places, I anticipate CRTC will reverse its position (at least they should do it) and FCC and CRTC will be in total agreement. Let me summarize CRTC ruling and indicate where they should reconsider.
There are some regulatory requirements that are worth noting:
Now let me give the reasons why CRTC and FCC are closer than others seem to think. With the Pulver decision, FCC has indicated that end-to-end IP is an information service – no interworking with PSTN, no “voice” regulation. There are strong indications that FCC will reject AT&T petition, which suggested that if even one leg is VoIP, then there shouldn’t be access charges. So FCC probably will indicate regulation applicable at the point of interworking will be in effect. FCC also does not want to lose USF and the idea of contributions based on phone number is getting consideration. Of course Powell and others have openly indicated that 911, relay service and wiretapping must be made available in VoIP as well, though they have left the door open on the timeframe. I do not see how much closer two agencies can be on the first round.
On the other hand, I am not as comfortable with the Sununu bill (I have not located the House version). That bill is also very concise and to the point, except that I have a serious reservation.
On the surface it looks like this bill is along the same lines of CRTC and FCC. But items 2 and 8 when taken together essentially kill the whole voice regulatory regime. I do not think that was the intention of the Senator. Firstly, item 8 is inconsistent. If a service provider packetizes the traffic in Class 5 switch, does the call originate in PSTN? If it does, then are we not discriminating against a particular transmission technology? If on the other hand, we allow this, then all PSTN traffic can be artificially declared to be VoIP and use item 2 to undermine the settlement charge system that is in use. We might as well declare that this bill is applicable to all voice applications. So the glaring failure of this bill is to precisely define VoIP application.
David has summarized a talk given by Henry Sinnreich in the recent VON meeting. Sinnerich said to the effect that “People say they have VoIP, but if you can't give me a SIP URI, then you don't have VoIP. You have a contraption of some kind, but it is not VoIP, as far as I'm concerned." I do not cheer this remark. Let me explain.
As I have said it before, the only real benefit of VoIP is that users can use one or more service providers concurrently; or they may not use any service providers at all. Any implementation that stifles this freedom is not to be encouraged. Use of SIP URI does not guarantee this freedom; use of another URI, like h323 does not curtail it. Windows messenger requires you to register to a proxy, violating the basic expectation. On the other hand, NetMeeting, using H.323 allowed direct communication. Equating SIP to the basic value of VoIP is a bit categorical.
Today Cox announced that it will use Verisign’s NetDiscovery service in its effort to support CALEA. This is a good development, not for the obvious reasons.
NetDiscovery is a back-end service – it interfaces between the service provider and LEA. The question is where will Cox tap the media? It has to do it in one of its network element, like CMTS. This will naturally put pressure on ASP’s like Vonage, who will be at a disadvantage. To offset this disadvantage, these service providers must be afforded equal access to the tapping point and in the same manner. This gives rise to the most natural solution to the CALEA problem: call control information is collected from the application service provider and the call content is collected from the access provider (if the former does not have natural access to it) with the coordination done by the LEA itself.
Copyright © 2003-2009 Moca Educational Products.