Tag Archives: IMS

Tried to replace the PSTN lately?

A great challenge for the current US wireline telephone companies is how to replace the voice telephone network (aka Public Switched Telephone Network or PSTN), build up over 80+ years by the Bell System and its progeny. Same concern applies to the cable companies that built their own version of voice service networks over the past 20 years. Though these have different origins, they share the issue of technology obsolesce (with increasing risk of outages, and higher on-going expenses), all while voice revenues are declining; making the business case for replacement extremely difficult. Over the past 29 years I’ve seen multiple attempts at replacement strategies and many different business justifications.

I had someone argue the first “PSTN replacement” was actually the installation of the SS7 network during the 1980s, which removed inter-office signaling from in-band (where it could be and was hacked!) to an out-of-band separate data network using X.25 technology (actually SS6 did this first, but was not fully deployed). If so, then this was the last “successful” replacement! Every attempt I know of since then has ‘failed’. First there was Integrated Services Digital Network (ISDN), which had a strong start in the late 80s, but has only had long-term success at the interface between public and private switches (PBXs). The residential form of ISDN, known as the Basic Rate Interface (BRI), re-used the existing twisted pair of cooper wires, yet was capable of carrying two digital voice channels and a very low bandwidth 16kbps switched data channel (2B+D). The voice channels could alternatively be used for 56kbps data as well. At the time there were no browsers, no webservers; mostly just FTP, email and private BBSs.  This was long before the World Wide Web became the runaway application of the Internet. In the 90s when the WWW came to be, ISDN BRI was not up that task and was quickly forgotten.

A really massive attempt to replace the PSTN was the Pacific Bell Consumer Broadband Network, which I was involved in from 1993-1996. The plan was to supply 500,000 homes in California (and elsewhere) a combination of voice, data and TV services, based on an exciting (to me, the engineer) mix of technologies from the telco world, but using cable’s hybrid fiber coax architecture. The big issue, though, was that it required running new fiber to each neighborhood and new cable drops to each of those houses. All that digging up of streets, or even stringing aerial, is disruptive and expensive. Being California, there were regulatory issues too (this was pre Telecom Act of 1996). But the final nail was when the Internet started to become a big deal [remember those dialup 56kbps modems?], but the data capabilities were based on ISDN…so this one failed too.

Not long after that the cable industry started deploying broadband cable modems to carry IP-based data, and the telcos followed with Digital Subscriber Line (DSL) services (and later, with IPTV using an upgraded DSL service). The key for both of these was that the implementation was scalable and an incremental overlay on the existing networks (coax or twisted pair), rather than a voice network replacement. No massive digging up of the neighborhoods. Well…, except for Verizon’s FiOS where the fiber is brought to every house directly using PON technology. That’s still a lot of digging, and a lot of expense to put electronics with a backup battery in every home. And so, having already deployed its most economical neighborhoods, Verizon has discontinued any further plans to grow the FiOS coverage areas. Interestingly though, FiOS still carries voice services via the same old Class 4/5 POTS switches in use since the late 1970s and 80s [their plans for IMS still pending].

Most recently? Look at the offer from Verizon Wireless (and AT&T also) to have the wireless network provide voice connectivity for your home phonesets, for only $20/month (on par or cheaper than most basic wireline services), … added to your existing mobile smartphone service, of course.

So what are the takeaways here?

  1. The Bell System (R&D by Bell Labs) did many things well. And that equipment has lasted. Yet, I also have to say, that won’t carry the telcos much longer…
  2. Data has been a growing theme in all these network changes. It’s clear that data (with video via that data) service is now the primary residential service. Not voice, nor linear TV. [It’s also clear that Bell Labs was spectacularly not good a predicting consumer demand for data services!]
  3. Incremental change is a lot easier than wholesale replacement, both in technology and economically. Don’t try to shallow the whole thing in one gulp; cut it up into smaller chucks, especially ones that can be justified in other ways…
  4. Mobile networks are taking over… at least for the 98ish% of the US population area where 4G will be fully built out. [And I didn’t discuss VoIP, except via implication…]
  5. Economically, it’s all about the bundle!, regardless of whether measured in lines, ARPU or RGU!
  6. And lastly, it’s also about moving away from heavily regulated domains [but we’ll save deeper discussion of that topic for a future post].


WiFi RF Engineering? (not an April’s Fools question)

After the tech bubble burst in 2001, I joined the Lucent Wireless Networks organization as the systems architect and engineering director responsible for 3G data applications. I had a really great group of engineers fighting hard to create the wireless world we actually have today – everything mobile data. And yes, I do mean fighting; because at that time the core power base within the organization was made up of RF engineers, those who had invented the first cellular networks during the 1980s (and earlier), plus their apprentices. They were all *real* engineers, with a deep understanding of the theory, math, technologies and challenges in designing and building cellular wireless network equipment. In meetings they were always devolving into discussions of link budgets, dB, hertz, and erlangs. I was mostly a software and systems engineer; yes, with a strong background in outside plant and voice (aka POTS) networks, but I was a foreigner in their domain, for sure. I suspect a few of them would intentionally take conversations into the deep technical details to a) impress their boss, b) let the one VP, who was an RF engineer himself, take the lead in the conversation, and c) avoid any discussion of the economics or business sensibility of an idea. Well, I learned fast that to thrive in that organization I needed to speak, or at least understand, RF engineer-ese.

In this time frame, Wi-Fi was just getting serious standardization and the PCIMCIA card was its most common form factor; the same form factor as the first 3G data devices. Perhaps that was the cause of the anti-Wi-Fi bias that developed in those RF engineers. Or maybe it was due to only being for short range data transport, and *not at all* related to the cellular business of voice service. But more likely, I suspect, it is Wi-Fi’s use of unlicensed spectrum (after all, you can’t calculate performance when there are multiple indeterminate variables in the RF link budget). So what did naïve Jack do? I became an advocate for fixed mobile convergence and “dual-mode voice service” over Wi-Fi and cellular data, using IMS as the enabling technology… And not long thereafter I moved on to Lucent’s professional services organization.

I smile a lot now, ten years later, with all mobile data devices having built-in Wi-Fi, the industry producing integrated Wi-Fi picocells and implementing of mobile data offload onto Wi-Fi. AT&T and Verizon each have many 10,000s of public Wi-Fi access points in operation for their subscribers’ use. The FCC is adding more free unlicensed spectrum for Wi-Fi to use. And I spent last year working with several of the major US cable operators on their massive outdoor Wi-Fi network RF designs and deployments. Of course, IMS technology is still struggling in wireless networks and the few dual mode services that have commercially launched don’t use IMS. But at least three of the best engineers that worked for me went on to found their own, or to work at, small cell and/or Wi-Fi startup companies (Meru, Ubiquisys, Airvana). They and many others have made it happen, not me. No regrets here.

Who Knew?

One of the greatest frustrations in my career has been the struggle for acceptance and commercial deployments of IMS (IP Multi-media Subsystem) in telecom networks. I started work on the ALU version of IMS with our software release 0.1, 0.2 and 0.3 – which were never intended to see the light of day beyond a lab. But we wanted to get the full requirements process engaged and the development team initialized in the right frame of mind. This was late 2002, when I was director of systems architecture and engineering in Lucent’s Mobile Applications Services Delivery. The initial IMS standards (Rel 5) from 3GPP had barely had time to dry, and had already started significant enhancement/revisions (still occurring). Yet many of us shared the strong conviction that moving voice services from the very optimized, but inflexible implementation that was then common in 2G technologies (and beginning initial deployments in 3G) would be the pathway to an explosion of new services and applications offered by mobile network operators, taking advantage of these new high speed mobile data networks that 3G enabled. We all remember what  happened to the wireline ISPs (Internet service providers),.. They had been reduced to only providing a “simple” point to multipoint IP data transport service for a fixed monthly fee, but with constant demand for higher speed (bandwidth) and capacity (data volume). Where was the money for future network investment going to come from, if not from an explosion of new services/applications? How would mobile operators make money once the inevitable influx of these non-operator (aka over-the-top) applications started coming available over mobile networks? Well, at the time, IMS was our answer!

Ten years later, and no 3G network that I know of has implemented IMS as a replacement for its voice core network. Yes, there have been a few supplemental voice services (e.g. push to talk), and several video or other data services implemented using IMS. I worked on a couple of those… None that I know of were commercially successful; though I’m glad the RCS effort continues to show some progress! Using 4G/LTE technology, MetroPCS is still the only US mobile operator with a commercial Voice over LTE implementation using IMS (not ALU’s). And that implementation may not survive the pending acquisition by T-Mobile (sounds vaguely familiar to another wireless IMS implementation that I worked on…). Verizon’s VoLTE implementation has yet to launch, nor any offering from AT&T.

So where is today’s largest IMS commercial implementation? It’s in wireline voice networks, e.g. over 2 million subscribers on just one US implementation. It is most ironic that the first two releases of the IMS standards didn’t include significant support for these wireline networks!! And we’re still waiting for that explosion of mobile operator applications… Oh, wait, I just received notification of some new appls for my iPhone… …uh!