ZapThink has recently noticed an important trend in the service-oriented architecture (SOA) world -- an increased interest in SOA from telecommunications (telco) carriers, as well as the ecosystem of vendors with telco-related offerings. Now, the SOA story in telco has been quite strong for a few years now, as many of the large carriers have targeted SOA for reducing internal redundancy across complex, heterogeneous infrastructures. The trend this year, however, is that telcos are looking to leverage SOA to provide better, more flexible products to their customers, and to roll out those products more quickly. So, while this ZapFlash will be of interest to the telco ecosystem, the real story here is in how service-oriented best practices will impact any consumer of telco services -- which of course means all of us, from the individual up to the largest enterprise.
Convergence and SOA
Ever since the Internet came on the scene in the early 1990s, the overarching business driver in telco has been convergence. Convergence first and foremost between the traditional analog, circuit-switched world of traditional phone lines with the digital, packet-switched world of the Internet, but now, convergence among landline and mobile phone service, wireless and wired Internet service, as well as cable and satellite television service. Needless to say, the telcos have had their hands full, especially considering that they are all immense, extraordinarily complex organizations that have typically grown through mergers and acquisitions.
A simple example of a convergence application would be purchasing ringtones on your mobile phone. You use your phone's interface to select and download the ringtone, and then you can go to your carrier's Web site to view your account, where you can see the purchase. If you call their call center, the rep can manage your ringtone purchase as well. And behind the scenes, a transaction takes place between the carrier and the owner of the rights to the ringtone. Building an application with such converged capabilities would be enormously complex in a traditionally integrated environment, and thus begs for implementation as a Service-Oriented Business Application (SOBA), which is a composition of Services that implements a business process. Leveraging SOA in this way promises more flexibility for the telcos as they develop such convergence products.
This flexibility benefit of SOA is not lost on the carriers, but the time-to-value is even more important to them. It can take almost a year to roll out a new product like ringtone purchasing, local number portability, or one current challenge, which is location-based services that leverage global positioning satellite (GPS) technologies in next generation mobile devices. If they can, say, reduce the time it takes to launch such a new product by a few months, they can save substantial amounts of money and also improve their competitive position against other carriers. In other words, the business agility benefit of SOA goes directly to both the top line and the bottom line for the telco carriers.
Laying the Groundwork for Telco SOBAs
The telcos, in fact, are all building SOBAs, and the first thing you need if you want to build a SOBA are some services to compose. As a result they are rapidly building out a range of services that they can compose into various SOBA-based products. Many such services fall roughly into the following categories:
- Session -- think making calls and maintaining state. The metaphor of the simple phone call still lies at the core of each telco's DNA, only now, we're leveraging the Internet, which pushes intelligence to the endpoints. Innovation around the Session Initiation Protocol (SIP), and SIP-related specs and technologies, drives this set of capabilities.
- Access -- think getting on the Internet and distributed functionality. On the one hand, the carriers are in the connectivity business, whether via dialup, DSL, or wireless. On the other hand, as the market for connectivity approaches saturation, the big win focuses on what they can offer via that connectivity. Here lies the full spectrum of Internet-based services, from Web Services to Software-as-a-Service (SaaS). Getting a lot of attention this year is the IP Multimedia Subsystem (IMS), which is a telco-centric architectural framework that supports access-based Services.
- Location -- think GPS and location-based services. Once everybody's mobile phone supports GPS, then what? If you think the point is to serve ads for stores as people walk by them, you're missing the picture entirely. What the true opportunity for location-based services is, however, is currently in the hands of the entrepreneurs and visionaries.
- Presence -- think buddy list and communities. When you place a normal phone call, you have no idea if the other party is currently able or willing to take the call. If you wish to send that person an instant message (IM), however, you can quickly see from your buddy list whatever availability status that person wishes to share with you. But why stop at IM? How else might we be able to use presence information in a converged world?
In addition to the four service categories above, the telcos must also work out two more infrastructure-centric, yet crucial Service categories as well:
- Identity -- the telco world currently deals with a complex range of different identities for their customers. There's you, your phone number, your phone, your family's phone, and your account with the phone company, to name a few -- and each typically has a separate unique identifier.
- Billing -- you didn't think you'd get anything for free, did you? One way of looking at the telcos is as massive billing engines looking for new products and services to bill for. These are the guys who figured out how to sell their customers individual items that might cost one or two cents at a time, and make money doing it. If they can do it with phone calls, they can do it with just about anything else.
From SOBAs to Mashups
By leveraging the Service abstraction at the heart of SOA, the telcos are abstracting the above capabilities as services, which enables them to assemble products by building SOBAs. Perhaps one of the more successful telco SOBAs to date is Verizon iobi, which combines voicemail, Short Message Service (SMS), the Web, a weather service, a mapping Service, a Windows-based desktop app, and integration with Microsoft Outlook's contacts, along with the required identity and billing infrastructure running behind the scenes. When a call comes in, an alert pops up on the desktop with caller ID information, and an SMS goes to any mobile phone with the same information. Click a link in the popup and see the location of the caller on a map.
Even though iobi rates high on the cool (and useful) meters, it's still essentially a first-generation SOBA, for two main reasons. First, it's essentially proprietary -- it requires Microsoft Internet Explorer, and blatantly states on its login dialog that it's "not compatible with MacIntosh [sic] computers." The other reason that iobi is stuck in the first generation is that it essentially falls short of being a mashup. In other words, iobi users have no control over the combination of services that iobi composes, beyond basic configuration capabilities. We can't switch the mapping capability to Google Maps, for example, or add our instant messaging buddy list to its capabilities, that is, unless Verizon happens to do that for us. And there's the rub: the true opportunity for the telcos, as well as for all telco customers, lies precisely in enabling this mashup capability.
The ZapThink Take
Therein lies the true challenge for the telco industry as a whole: when will they realize that the big win for their customers, and hence for their competitiveness overall, is in enabling their customers to build products themselves? They are rapidly driving toward the service-oriented vision of their world, and yet all indications point to the strong possibility that some telcos may not change their product development culture to allow customers to compose the services they offer as they see fit.
The irony here is that the telcos are devoting enormous resources to building better trees, but some of them may still miss the forest entirely. This pattern, however, is actually a familiar one, and one ZapThink frequently writes about: how technology change is the easy part of implementing SOA, when you compare it to the challenges with organizational change. Rethinking how to build and deliver products to customers is essentially a change in human activity -- more a change in what they do with the services they provide, more so than a change in how they provide those services.
Telco customer service is one case in point. Even those telcos who have service-oriented their customer service-related capabilities still have a miserable reputation as being siloed, impotent, and fundamentally hostile to their customers (speaking from personal experience, by the way). If shared services hasn't changed those telcos' attitude toward customer service, then service-orienting the products they deliver to their customers isn't likely to either.