Hey, You Kids! Get Off My Lawn!
At January’s Cloud Computing Summit in Florida, we landed on the topic of new interfaces for speech technology based on cloud technology. These Internet-accessible, remotely hosted services, such as Ifbyphone’s Cloudvox, Voxeo’s Tropo, and Twilio, allow you to add telephony, speech, and related services to applications in multiple programming languages.
At one point, a participant commented on “classic” call control XML-based applications. CCXML uses XML to manipulate the telephone network and can also run on a hosted server, but it requires specialized expertise and complex integration with more conventional programming languages. CCXML as a whole, the participant seemed to imply, was practically ancient technology.
As probably the most senior CCXML expert present, I became disconcerted. My CCXML applications run quite nicely, make my clients lots of money, and are not something I think of as outdated.
I spent the rest of the conference entertaining select colleagues by pretending to be a cranky old man shaking his cane at a group of noisy interlopers: “Hey, you kids! Get off my lawn!” But I admit this incident made me think carefully about these new cloud-based services. CCXML is more than a few years old, so is it ready for the dustbin? Should I advise clients to move on?
The first question that comes to mind, of course, is standardization: The new cloud services do not follow a standard. Will this be a problem?
I can move CCXML applications between vendors, at least in theory, because they all follow the same standard. (Stop laughing. I said “in theory.”) But cloud services are completely different, and applications will not be portable. However, many extremely smart developers simply don’t care about portability. They willingly create businesses on top of Facebook, iPhone, and Twitter, tying their entire business models to the success or failure of another company’s business and making good money doing so.
Fomenting a Revolution?
Next, let’s think about this particular innovation. I’ve lived through several technology revolutions—heck, I’ve written a book on the topic—so I have a few things to say about this. Revolutions come in different types, and the cloud-computing one could fit the same pattern as the computer telephony revolution.
The idea behind computer telephony was simple: abandon complex, purpose-built, and expensive specialty equipment to connect computers to the telephone network. Instead, put telephone-network interface cards into ordinary PCs. These far less expensive PCs might fail from time to time, but we can just put spare ones in place. Good idea. But if you walk through the facilities of a telephony provider that handles a serious number of calls, you won’t find off-the-shelf, cheap PCs.
The migration from cheap to expensive makes perfect sense, if only in retrospect. Pioneering developers who used the technology had applications that could withstand a lower level of reliability. If a computer telephony server failed, then calls on that server would drop, but maybe that didn’t make a difference. As time went on and the technology became more widespread, people decided to use computer telephony servers for applications with more traditional requirements, such as not dropping calls midsentence. Classical telecommunications companies purchased servers and demanded Network Equipment Building Systems (NEBS)-compliant PCs—after all, when you maintain hundreds of them, you want to know they won’t catch fire when a cooling fan fails.
Natural evolutionary pressure drove the transformation from inexpensive PCs to more expensive PCs. At the same time, the machines remained popular. Sure, reliable PCs cost more than standard ones, but drastic reductions in price meant PCs cost less than the “gold-plated” boxes they replaced.
What does that tell us about whether cloud services will replace CCXML? Just like the computer telephony revolution, early adopters will cheerfully trade advanced features for inexpensive simplicity; some might not even realize what’s not included. After all, developers who work on the cutting edge of nontelephony cloud-based applications always look past limitations and work with what’s provided to get their ideas out the door. Despite the rough edges, I expect a steady stream of developers to adopt cloud-based telephony.
Then we’ll see an evolution. Developers will demand more complexity, and cloud service providers will respond. Eventually applications will become just as complex as today’s CCXML applications. So keep your eye on cloud-based telephony. It might very well take over the world.
Moshe Yudkowsky, Ph.D., is president of Disaggregate Consulting and author of The Pebble and the Avalanche: How Taking Things Apart Creates Revolutions. He can be reached at speech@pobox.com.