Breaking Down the Language Barrier

Proprietary manufacturer standards have always come in the way of synchronization between devices. SyncML has set out to change all that--despite the aloofness of Microsoft.

Over lunch in Paris two years ago, Douglas Heintzman and some friends were grumbling about the number of computing devices that didn't automatically sync with each other. The lunch bunch decided that people needed a standard way to update e-mail, contact lists and other information no matter what device was in use. As the dishes were cleared away, Heintzman was left holding a stack of café napkins with ideas, to-do's and likely players that could create a common language. And so SyncML was born.

During the next 11 months, wireless data companies joined Heintzman's campaign and created what was to become the first flight of the SyncML specification version 1.0.

“We wanted to deliver a core set of functionality and not try to boil the ocean… do it very quickly and maintain momentum and energy,” Heintzman said.

The SyncML initiative parallels what Heintzman does every day at IBM: As manager of business strategies and standards in IBM's pervasive computing division, Heintzman does “landscape strategy.” He tries to figure out what levers to pull to enable markets to develop in healthy ways. The upside to that is IBM can pursue those opportunities and sell products to customers.

The SyncML specification is one lever that could stoke a firestorm of ideas and applications that could have been too expensive to create in multiple device languages. While IBM could have undertaken the SyncML initiative itself, it would have been too costly, and the company would have suffered from having too narrow a perspective, Heintzman said. Involving a number of companies with different perspectives from different computing geographies “enabled us to make the right design decisions as well as amortize the costs across participants.”

In addition to IBM, Nokia, Ericsson, Motorola, Openwave Systems, Symbian and Starfish currently preside over the initiative's development.

At its core, SyncML enables an “intermittently connected” computing model. In other words, it addresses a market in which devices move in and out of coverage areas, are attached and unattached to wires and get turned on and off. Users want to do things whether or not they're connected to the back-end application.

SyncML defines a standard protocol and object representation so that any number of back-end servers and applications can handle updates and changes with any number of front-end devices in operation. Each side of the equation doesn't have to come from the same manufacturer.

“SyncML is like what HTTP and IP are to the wired world,” Heintzman said. “It's the fundamental transport mechanism that has built-in awareness that understands that it's going across potentially uncertain network connectivity. It checks on the message status and checks that the records have arrived intact the way they were intended to.”

Since delivery of the first version, SyncML has evolved from the fundamental concept of enabling end-to-end connectivity between devices and back-end servers. The latest version provides mobile device management and software downloads. Heintzman predicted the next step will incorporate digital rights management.

“It occurred to us that carriers might want to update the configurations of many subscribers, many of whom aren't online all the time,” Heintzman said. “Effectively, subscribers have a local store of the data, and the carrier has a server store of the data,” he said. “Ultimately, in this environment, it's a synchronization application.” Because of that, it only required minor modifications to layer device management on top of the original specification.

In this world, updates to the cell phone, new versions of code, bug fixes, new features and functions can all be delivered remotely over the air — and transparently — to the end user.

Like any protocol candidate, SyncML's success rides on its acceptance by a majority of vendors. While its current contingent of handset, server and application companies is strong, notably absent from membership is Microsoft.

Because the SyncML initiative crosses over into its enterprise sandbox, Microsoft's participation and support could be considered critical. But while the invitation was extended, the Redmond, Wash., company remains on the sidelines.

“Industry specifications aren't necessarily in their best interests. They generally adopt them when they have to,” Heintzman said.

Initially, Heintzman thought it was important to bring Microsoft on board because of its engineering perspective, prestige and critical mass. But as the SyncML initiative gathered steam, Heintzman said Microsoft might need SyncML more than vice versa.

“As SyncML gets implemented into a high volume of handsets, all of Microsoft's competitors will have a lot of devices that can talk with each other, and they can't.”

While Microsoft rules the wired enterprise market, it doesn't exercise the same domination in wireless. SyncML is gaining momentum with companies like Nokia and Motorola that have dominant wireless market share and a longer-term relationship with carriers.

Microsoft is accustomed to driving the market itself and buying its way into other sectors, which could explain its absence, said Andrew Cole, global wireless practice leader for Adventis. “The secret to success in wireless is the relationship with carriers. Microsoft is in significant danger of missing that point.”

Yet Microsoft plans to continue its own efforts with its ActiveSync, which allows for synchronization between the PocketPC and the Internet and intranet. The “value-add of open interfaces is well documented,” but our position ultimately asks if this meets the user expectation, said Michael Wehrs, director of technical standards for Microsoft, who thinks the functionality of the ActiveSync capability fulfills what users want in a mobile environment.

“This is a marketplace that Microsoft doesn't have a footprint in or control over. They don't own the handsets, and they don't own the server infrastructures of the carriers, so we are not so worried right now,” Heintzman said.

Heintzman's movement continues to amass accomplishments. In fact, today there are SyncML clients built for the Pocket PC platform. The initiative also has protocol stacks that run on top of Bluetooth as well as the wide area and local direct connect and infrared networks.

So while customers likely won't see a “SyncML Inside” ad campaign, the initiative is responsible for linking a disparate colony of devices.

“We are comfortable being fairly invisible, of doing the yeoman work of making the plumbing work that allows the houses to be built,” Heintzman said. “When we look back 12 months from now, we will be happy to see that developers can build applications that will work across an intermittently connected computing model without worrying about the vagaries of what that means.”