- TechNote on Infrastructure and Performance & Unified Communications
- Larry Hettick, Editorial Director & Senior Research Fellow, Webtorials
The path to replace the PSTN with a fully VoIP-based network won't be easy, with roadblocks and detours that will need to be solved before any carrier can retire their PSTN. The issues are not insurmountable; however, they are substantial. This TechNote looks at the potential pitfalls, also highlighting possible solutions for each.
Network-to-Network Interoperability
Not every carrier will offer VoIP, and legacy operators will make their transitions at different times. This presents two issues. First, unless two carriers both support SIP signaling, a gateway will be needed to connect calls between the VoIP-based carrier and the TDM-based carrier networks. The good news is that carriers' infrastructure providers have been building and carriers have been operating these gateways for over a decade, and scale is not an issue, even for even the largest networks.
The second network interoperability issue is a little more difficult: how to offer a fully compatible network signaling between two VoIP networks. As we have discovered with enterprise networks, one vendor's IP-PBX may have difficulty establishing a session with a different vendor, and each supports different features.
Today, the PSTN connects different vendors' VoIP platforms, and without a PSTN to serve as a signaling gateway, carriers will need to comply with a standards-based network to network interface specific to VoIP. Fortunately, the Alliance for Telecommunications Industry Solutions (ATIS) and the SIP Forum are working jointly on a series of call routing standards for a network to network (NNI) interface and the associated database operational components carriers need.
Codec Interoperability
Eternal optimists would like to believe that carriers and their suppliers will settle on a single codec standard that delivers high definition (or wideband) VoIP. But the more likely scenario is that multiple codec standards will be used even in the same carrier's network, so transcoding will likely be required in perpetuity. A variety of wideband standards and codecs is already established, using standards such as ITU-T G.722, G.729.1, and Microsoft Real-time audio (RTA), and more standards are expected as VoLTE and audio codecs are factored into the equation.
In some cases, a session border controller or a gateway offers this codec interoperability today--most frequently in an enterprise setting. Going forward, interoperability will be needed to connect residential and business VoIP codecs, cable and telco VoIP codecs, narrowband and wideband codecs, wireless and wireline codecs, video-call audio and voice call audio, and potentially even WebRTC voice and video codecs. Transcoding is certainly feasible, but it will need to scale massively as more carriers adopt VoIP networks and as more codec standards are created.
Managing Security
Today, SS7 is used to set up and tear down calls on the legacy PSTN, and it has inherent security advantages. First, potential hackers are far less likely to be skilled in SS7 software than they are in IP protocols. Second, SS7 network nodes are located in tightly controlled physical facilities. By contrast, the signaling infrastructure and protocols that participate in call set up and tear down for IP calls are highly distributed. SIP sessions can originate or terminate in everything from an IP PBX to a Microsoft X-Box 360, even to a mobile device that supports a VoIP app. SIP's distributed nature and the openness of IP protocols require attention to security that is above and beyond what we need today for SS7.
Fortunately, IP security in enforced today with tools such as session border controllers (SBCs), advanced intrusion detection systems, and other security measures. Still, carriers and their suppliers must continue to develop the tools needed to thwart the hackers who will doubtless also make the PSTN-to-VoIP transition.
The Need for New Regulatory Frameworks: Lots of Questions
As a heavily regulated public utility, the PSTN has regulatory metrics covering network reliability, voice quality, price schedules, and more. However, the regulations surrounding VoIP networks are relatively undefined. For example:
- With the U. S. Federal Communications Commission (FCC) decision to reclassify the Internet as a regulated utility, should the same PSTN regulatory framework apply? Should the FCC distinguish between VoIP and Voice over the Internet?
- Should regulators have jurisdiction to classify a VoLTE or a peer-to-peer WebRTC session just as they would a PSTN call?
- Who should guarantee voice quality and reliability if the call starts as a WebRTC session, and then terminates on the carrier's VoIP endpoint? (Voice quality management was resolved from a technology perspective over a decade ago.)
- SIP session protocols (and IMS) permit user information such as location, endpoint / device type, and select profile data to be included with the call set-up message. Is this information protected under privacy regulations?
The only thing that is clear about these regulatory questions is that they must be answered as carrier shift to VoIP networks.
Conclusions
Carriers have significant experience with IP networks in general because for about decade, more data has been crossing their networks than voice; consequently, carriers have learned to engineer IP very well and they are already addressing issues such as quality of service and security. They are also experienced with tools including session border controllers, gateways, and transcoding needed to transition their own networks to VoIP and to connect to other carrier's VoIP and legacy PSTN infrastructure. It won't be easy, but as was explained in my earlier tech note "Replacing PSTN with VoIP: Not If, But When" the results will be well worth the effort.
This TechNote is brought to you in part due to the generous support of:
This piece begins to address some - but not all of the bigger issues surrounding PTSN retirement. For example, either all non-VoIP phones must be replaced, or somebody will be selling a lot of gateways so Grandma's rotary dial phone in the kitchen will still work.
Another big issue: billing and back office support. Todays back office systems were built to work with legacy networks on quaint standards like EBAF outputs to a 50-year old billing system. Carriers may be willing to offer unlimited calling within their own network, but inter-carrier settlements between legacy and VoIP network providers will get to be very interesting.