The IP Transition Part 1: Some ThoughtsThe IP Transition Part 1: Some Thoughts
The PSTN’s transition from legacy TDM technology to IP is coming. It is inevitable. But there are many things that operate over the PSTN that have to be supported.
May 16, 2014
The PSTN’s transition from legacy TDM technology to IP is coming. It is inevitable. But there are many things that operate over the PSTN that have to be supported.
None of these are a deal breaker, but some could create considerable conflict with those using or connecting to devices and services that could be hampered or prevented from using the IP network. There is also the issue of what the transition will cost enterprises and consumers, a bill that could cumulatively be very large.
SIP Trunks Offer a Look at the IP Transition
As I look at SIP trunking and what it does not support, I have less confidence that the IP network replacing the PSTN will be entirely successful. The SIP trunk is an IP network connection. It may be fine for voice communications, but there are a number of devices and technologies that work on the PSTN but are not supported by SIP trunk providers. And it appears they may never be supported. The challenges for SIP providers will be the same challenges facing the PSTN replacement. Take, for example, the poor support for FAX on SIP trunks.
Consider Fax
The T.30 fax standard was designed for the PSTN, not IP networks. It can work if the G.711 codec is used. Fax transmission, however, has trouble with IP network latency, jitter, and packet loss, which limit its success. The T.38 Fax over IP standard does work, but some SIP providers don't support it. Will this be the same lack of support on the IP PSTN replacement network?
Some pundits say fax is dead. Not true. The financial and real estate industries still use it heavily, so do countries like China and Japan. Fax may be dying, but it will continue to exist for a long time, so the IP version of the PSTN will need to support it.
Will Everyone Need a Session Border Controller?
The Session Border Controller (SBC) is a very common device protecting the provider from the enterprise and the enterprise from the provider. I have a network at home that is just as important to me as an enterprise's network. Will I eventually need an SBC or an SBC function in my router to protect my home network? If I need an SBC, will the cost be included in and possibly increase my broadband service charge? This begs the question, "Will the IP PSTN version cost me more than what I presently pay for my telephone service?" It does not seem fair for me to pay extra so the PSTN providers can force me to migrate to an all-IP connection.
SIP Trunks Do Not...
When capacity planning for SIP trunks, most technologists think of the number of simultaneous sessions required (see "SIP Trunk Bandwidth: How Much?").
SIP trunking and the associated Session Border Controllers (SBCs) can scale up to thousands of sessions. What is not apparent is that both SBCs and SIP trunk providers have trouble handling high call establishment rates. I know of one case where the SBC incumbent vendor could not support an enterprise's call rate on any of its models. Another vendor was eventually successful. This condition also means that the SIP trunk provider has limitations. I have also learned that some, if not all, SIP trunk providers will not support auto dialers.
What if the enterprise wants to deliver on-hold music? This will work only if the SIP trunk provider employs G.711 codecs over the entire connection. If voice compression is used, like a G.729 codec, then on-hold music does not work.
SIP trunks are not used for dial-up modems. Most of us do not use dial-up Internet access. However there are many devices such credit card readers, alarm, and monitoring systems that use modems not supported by SIP trunks. The list of devices that the IP version of the PSTN will be discussed in the next blog, "The IP Transition: Enterprise/consumer Impact."
Will I Need to Buy SIP Phones?
If I can keep my legacy phones for a while, will I be able to connect them to a gateway at home? If so, who provides the interface and at what cost to me? Do I need to have an Analog Terminal Adapter (ATA) like Vonage and Ooma use, or will the IP provider offer the ATA? I am not against the IP transition, but why should I have to pay extra for it? This opens another revenue stream for the providers that want to avoid the regulations that govern the PSTN.
Another reason I may have to buy a SIP phone is the limitations of an ATA. Looking forward, I expect that IP providers will want to sell me some Unified Communications (UC) features. Presence comes to mind. The ATA will not likely support this feature, so the SIP phone will be the solution. While I gain access to presence, the IP provider will charge me more for this feature.
Connecting to the Rest of the World
There are absolutely no assurances that other country's PSTNs will be implementing the IP transition simultaneously with the U.S. Therefore, the IP PSTN providers will have to implement technologies for off-IP network calls. This is not new and could be an ongoing IP provider cost that will be factored into enterprise charges. This presents a range of problems for the international enterprises as well as government agencies.
There will be two sets of technology to manage, IP and legacy. The IP transition adoption will be on a country by country basis - except it may be uniform in the European Union. Will poorer countries be able to afford the IP transition? Will the poorer countries buy our old TDM technologies to increase their own network capacity, thereby extending the life the PSTN?
The international IT staff will have to be competent in legacy technologies--technologies they know will soon be obsolete - not a good career move. The international communications regulations will be different. The connection requirements may be different. SBCs may become even more important in those countries that are monitoring enterprises and governments within their jurisdictions.
Will SIP Trunk Problems Be IP Transition Problems?
The implementation of SIP trunks still is encountering problems. Every year, The SIP School surveys over 800 organizations to determine the extent of SIP trunk provider problems (see "SIP Trunk Provider Problems Persist"). A major surprise is that the problems persist year after year and do not seem to be decreasing in volume. This indicates that either the SIP trunk providers are not interested in reducing the problems, they have poorly trained staff, or configuration errors are rampant. If this happens with SIP trunking, which is less ambitious than the IP transition, then we are in for a number of problems.
Back in 2010 I looked at the retirement of the PSTN from a different perspective, "PSTN Closure, the End of POTS, the Challenges." Another recent blog, "A Framework for the PSTN Transition" covered concerns about what the IP version of the PSTN may not support.
In blog #2, "The IP Transition Part 2: Enterprise/Consumer Impact," I will look at the many devices and services that are presently supported by the PSTN and the impact the transition will have on enterprises and consumers. The next blog looks at some of the costs and disruptions that will be incurred by enterprises and consumers once the IP transition begins.