Gartner gives prescription for facing down mobile adversity

ORLANDO, Fla. -- Considering mobile enterprise solutions?
Written by David Berlind, Inactive

ORLANDO, Fla. -- Considering mobile enterprise solutions? You'll need to start thinking "disposable" when you think "mobile." That advice comes from Gartner's expert on mobile solutions Nick Jones, who gave the lead presentation for Symposium's mobile track. "You should be able to build an application, deploy it, get value from it, and throw it away in about 18 months."

Jones noted that there's a significant degree of uncertainty in the future of the various wireless networking technologies--which are the key enablers to mobility. Between the myriad short- and long-range wireless networking technologies (standard and proprietary), the pace at which they are evolving, and their changing applicability to various application types, the volatility in the networking area alone over the next 12 to 48 months is enough to make most recent deployments look prehistoric within a couple of years.

For example, in showing the 2G to 3G roadmap for long range (carrier-based) wireless networks, Jones showed how, starting in 2001, the GSM and TDMA infrastructures in the U.S. and Europe took 18 months to merge into GPRS (general packet radio service) and then, took 12 to 18 months to bifurcate into two paths (GPRS and EDGE), each of which subsequently will take about 24 months to converge on WCDMA. Running parallel to that, Jones' roadmap shows the evolution of CDMA, the predominant alternative to GPRS in the U.S., from its plain vanilla incarnation in 2001, to CDMA2000 1xRTT (what's available from Sprint and Verizon today) to CDMA 1x-EV DO and CDMA 1x-EV DV, the former of which Verizon is testing in San Diego and Washington, D.C. but whose future is cloudy.

The incremental steps on Jones' roadmaps are associated primarily with breakthroughs in bandwidth, which creates a moving target in terms of what mobile solution architects can ask their applications to do.

Further complicating the mobile solution architect's job are the confusing labyrinths on the short-range wireless and device fronts, as well as the changing range in the ongoing costs associated with the provisioning of wireless connectivity. Jones divided the roadmap into adjacent (Active RFID, Passive RFID, and near field), personal area networks (ultrawideband, 802.15.4, Bluetooth, and proprietary PAN), LAN (proprietary mesh, 802.11 a thru f, and 802.15.4), and metropolitan area networks (802.16, proprietary fixed mesh, and 802.11b). As can be seen from the technologies listed in each of the short-range silos, some technologies are strictly vertical applications while others can be applied horizontally. Between that complication and the expected evolution within each silo, those deploying mobile applications have to decide whether they'll try to straddle silos and, if they do, which networking technology is best to use.

Incidentally, in a meeting with Intel chairman and CEO Craig Barrett here in Orlando, Barrett mentioned 802.16 (otherwise known as WiMax) enough times to make clear that it's a strategic technology for Intel. Intel's vice president of sales and marketing has also discussed the company's plans for manufacturing 802.16 silicon.

"The entire value chain is shrouded in a veil of volatility," said Jones. "Mobile success is not about strategy, it's about tactics." Perhaps Jones' 18-month window is a blessing for those seeking job security. If you can convince your boss that your company should be prepared to throw out today's deployment in 18 months, that about coincides with the time you (and your management) might begin regretting a "legacy" technology decision. Once expectations are properly set, you may not be held accountable for more strategic decisions.

According to Jones, every organization must be prepared for multiple device connectivity. Speaking of Symbian, RIM, WinCE (and derivatives like PocketPC), and PalmOS, Jones noted that comparing the mobile platforms to each other is like comparing apples to oranges. Because each platform has its own strengths and because of the range of needs of individual users, IT departments will have to accept the fact that setting device standards, especially against the backdrop of device evolution, could be an impossible task. The resulting conundrum is what Jones calls "managed diversity."

Jones did, however, declare Microsoft's PocketPC the winner in the thick mobile client category for enterprise deployments. He failed to mention, however, that in a few years just about every platform will be pretty thick by today's standards. If Jones is right about PocketPC, it may not win only in the enterprise applications category. It could win across the board--an outlook that doesn't bode well for Symbian, RIM, or PalmSource and perhaps Java. According to the research firm Canalys, shipments of HP's PocketPC-based handhelds recently surged ahead of those from Palm in Europe, which has been a traditional stronghold for both Symbian and Palm.

I asked Jones whether it might make sense to target Java, one of the most common denominators across the various platforms. In response, Jones lamented the fate of Java, declaring the mobile versions crippled, and blaming Java's current custodians (primarily Sun and the network operators) for dropping the ball. Ultimately, Jones believes, Java could help to manage the diversity if only another custodian would step forward. Without naming names, Jones sees a 70 percent probability that "someone will produce a more standardized corporate Java."

That someone is undoubtedly IBM. In what I think is Big Blue's first attempt to spread its Java gospel beyond the domain of its own offerings, IBM recently announced that its version of mobile Java --- known as WebSphere Micro Environment --- would be distributed with Palm Solution Group's Tungsten handhelds.

Managing that diversity, however, has its challenges. To face that challenge, Jones offered another idea: Turn over your mobile projects to an IT service organization. "You have to be the integrator," said Jones. Unfortunately, there isn't even a middle of the road one stop solution. Te safest bet is to go to an IT services provider and, where they don't have the coverage, get them to at the very least manage it. The lowest risk is to go with a large IT services provider."

You can write to me at david.berlind@cnet.com. If you're looking for my commentaries on other IT topics, check the archives.

Editorial standards