To Boldly GoTo Boldly Go

If your organization is fairly typical, you may have already gone through more than a few cycles of frustration in attempting to deal with and plan for deployment, maintenance, and support issues around personal digital assistants (PDAs). You probably have weathered the first wave of unsupported, privately owned devices sneaking in through the back door, and you may already be using PDAs in an official capacity for access to enterprise business assets such as email and calendaring. You may have already deployed, or be considering deploying, PDAs in a more strategic sense for ubiquitous access to distributed enterprise data, and are wondering what implications that has for your enterprise. This article should give you a high-level overview of some of the issues involved with enterprise PDA deployment and provide an underseat flotation device in case you find yourself drifting in uncharted parts of the PDA seas.

Over the last decade, PDAs have climbed the peak of overheated hype and unreasonable expectations, fallen through the trough of disappointment, and are now well on their way into the plateau of reduced expectations, technological maturation, and actual functional strategic deployment. The good news is that both the devices themselves, and the infrastructure and tools necessary to support them, have gone through their trial by fire, and are now in most cases genuinely enterprise-ready. Deployment of mobile devices no longer puts you on the bleeding-edge — you are now firmly in the comfort zone of mainstream support.

The bad news is that as with any computing device, you'll find that by far the greater part of the total cost of ownership still comes from the long-term maintenance and support costs, along with application development and deployment. PDAs have gone from being low-power personal information management (PIM) devices to full-fledged computing systems: and with high-end capabilities come high-end support prices and management headaches. Such features as increased memory, third-party device support, and Wi-Fi and Bluetooth capability not only extend device functionality, but also multiply the complexity of support issues. Gartner Group estimates deployment costs of a typical PDA at $3,000 per device, which is about 10 times the price of the average basic PDA. My experience is that this figure may actually be low; it all depends what you do with the devices. Clearly, wireless support and real-time data access will cost more to support than basic PIM functionality. The simpler you can keep your PDA deployment, the less it will cost.

Organizations tend to be at one of three discrete stages in their use of PDAs and mobile computing devices. Stage one is where privately owned devices show up in an ad hoc fashion. Stage two is the use of employer-purchased PDAs used solely for email and PIM functionality. Companies in stage three have business-critical applications or data residing on their mobile devices. It is typical, but not universal, that the three stages are passed through sequentially — the exceptions would be organizations with a historic need to deploy industrial or other vertical mobile applications, such as utility companies, shippers, mobile sales forces, and so on. These kinds of organizations have often been at the forefront of using mobile computing devices in a stage three scenario even before considering their use for PIM or email purposes (stage two). In any case, the complexity and cost of support tends to increase significantly as you go upward through the three stages. So, the obvious first question to ask is, should you deploy PDAs at all?

To Boldly Go

The primary, overriding business need that should drive enterprise PDA adoption is a requirement to push data to the edges of the organization. Do your users need immediate access to enterprise data and resources when they're away from traditional workstations? If not, despite the hype, you probably don't need PDAs. If, however, you have a compelling business need for on-demand access to critical data, PDAs can be a cost-effective and highly portable way to push that data to where it needs to go.

Keep in mind that the PDA form-factor carries significant physical limitations that go along with the benefits of mobility. The first is screen size; despite improved high-resolution color screens, you aren't going to want to do complex data visualization, or even typical Web browsing, on most PDAs. For some business purposes, however, advances in large-screen devices such as the PalmOne Tungsten T3 (3203480 resolution) will give you more to work with.

Ease of data input is another issue. Whether your users have devices with thumb keypads, are typing on portable fold-out keyboards, or are trying to enter data via pens, they won't want to enter large amounts of freeform data into PDAs: and when they do, their error rate will be high. While some talented thumb-typists can pull of virtuoso feats with their mini-keyboards, most of us mere mortals are better off using PDAs largely for read-only data or applications that have limited or well-defined data entry requirements.

There are niche devices, such as Tablet PCs that can bridge the gap. However, despite improvements in processor speed and usability, hybrid devices such as tablets and clamshell form-factor PDAs have so far proven to be too much of a compromise for most mainstream users — too big for true portability, too small to be a desktop substitute.

Stage One: First Encounter

Coming on the IT scene in the late 1990s a little like miniature unwelcome houseguests, privately owned PDA support was for most IT organizations their first encounter with such devices. Early adopters — often executives and IT staff— typically brought in their devices, attached a docking cradle to their PCs, and handled their own support ... until something went wrong.

Help desks received frantic calls regarding performance problems, which often were the first indication that these devices had snuck in under the corporate radar. Having to make reactive policies after the fact isn't the best way to manage an IT shop, but many organizations had no choice but to go through this process. They quickly drafted policies to deal with ad hoc PDA use in the enterprise. Key issues became data security and synchronization of email and calendars — the latter two perhaps officially unsupported but nonetheless requiring infrastructure changes.

On to Stage Two

Despite the fact that the Palm Pilot pretty much invented this PDA space and WinCE/PocketPC devices imitated and then extended what Palm did, the best term for stage two is the "Blackberry Effect," thanks to the explosive popularity a few years ago of RIM's mobile device. Many organizations are currently at stage two, which is dominated by deployment of PDAs for PIM activities — primarily email and calendaring. Organizations at this stage have done some pilot projects with early adopters, who synchronize handheld devices locally with their desktop computers. Next, stage two organizations will have decided to deploy some form of centrally procured PDA as a limited-purpose tool for remote access to email, possibly calendaring, and other PIM applications.

At stage two, some organizations are using PDAs based on the PalmOS and PocketPC operating systems: but perhaps most common are point-solution devices such as the ubiquitous Blackberry. For many companies, always-on email access through a Blackberry or other device has become as essential as a cell phone or pager — and as expected by clients, customers, and internal colleagues. With that kind of widespread usage come expectations for uninterrupted service and the need for enterprise device management policies.

One thing you won't find with the Blackberry type of point device is a plethora of third-party tools and programs, or even a robust set of PIM tools aside from email (they exist, but aren't as capable as those available for more full-featured PDAs). Of course, the small number of choices for third-party programs to support can actually be an asset, if all you're looking to support is remote email access. Blackberries and similar devices are appliances — like cell phones — and that's precisely all that many organizations desire. With dedicated point-solution devices, back-end integration with corporate email systems is generally excellent.

In sum, the major issues at stage two are physical inventory and device management; data security; and integration and synchronization of point-solution devices for PIM activities, primarily email and calendaring. If you anticipate moving to the next stage, your organization may be better off looking at PalmOS or PocketPC devices, which offer a larger set of options for expandability and present more third-party and custom application alternatives.

Stage Three: Data Access

The third stage of PDA deployment includes the implementation of custom or off-the-shelf third-party applications that make business data available to enterprise users. This data can be synchronized inside and outside your physical facilities via both wired and wireless means, including Wi-Fi, Wireless WAN, modems, direct-attach cradles, Ethernet cradles, and even (though less commonly) via sneaker net, with data uploads stored on Compact Flash or Secure Digital memory cards users can insert into most PDAs.

With all three stages, but especially at stage three, data security should be a major concern. With portability come increased odds that a device will be lost or stolen. The mobile device might contain highly sensitive, proprietary corporate data, such as engineering documents, internal applications, client lists, or even just internal phone lists. If you're at stage one, security is hard to enforce; but even here, you should have a policy of requiring on-device encryption and access controls. Devices typically support at least passwords at power-on. Hewlett-Packard, like some other vendors, has begun putting biometric authentication devices on high-end iPaq units. PDAs now on the market include the capability to automatically encrypt data on the handheld, which should help keep your data out of unfriendly hands.

Policies will be difficult to enforce on devices that you don't own or control. However, if you allow employees to put company data on personally owned devices, at minimum IT has a clear mandate to enforce a clearly communicated security policy.

At stages two and three, security is theoretically easier to enforce since your organization has procured and configured the devices. But typically, PDAs are extremely open to reconfiguration by users who like to install a variety of third-party programs. One option is to deploy a full-fledged mobile system management package, such as Mobile Automation's management suite or XcelleNet's Afaria. This is a rather complex, but centralized way to manage which software and configurations are on each device. It will allow you to do remote inventory and configuration management, such as pushing a new version of a database to PDA users whenever they synchronize.

Synchronization

In the past, on-the-road data synchronization was nearly an insurmountable problem. In many cases, the standard method was to use a dedicated phone line and modem on each local PC workstation, and then allow users to dial in remotely and synchronize their PIM data. Obvious security concerns aside, having an always-on, dedicated PC, modem, and phone line on each user's workstation isn't a scalable or supportable solution.

The current generation of PDAs generally come with a much wider range of synchronization options, both wired and wireless than we had in the bad old days. The challenge is to arrange your applications to function with data that you may refresh only occasionally. Circumstances exist where you can depend on ubiquitous network connectivity for PDAs (Wi-Fi-enabled corporate campus, wireless WAN-enabled PDAs, and so on), but most often, users should consider data on a PDA to be less than real time and refreshed only occasionally. Sybase iAnywhere, among other solutions, offers technology to help maintain data consistency with mobile platforms. In general, however, it's best to start your application development with a baseline assumption that PDAs will synchronize only intermittently — and that the data may be hours or days old.

Converging Worlds

Communication and computing worlds are converging fast. The growing ubiquity of wireless networks and mobile devices such as PDAs could soon tip the balance, leaving the majority of organizations in a hurry to get to stage three, where they will be managing data issues more carefully and methodically. Strategic interest in developing real-time data access and business intelligence to support automated business processes will also push the need for more sophisticated architectures that include PDAs and mobility as part of their enterprise view.

Nonetheless, the practical approach is to look at business need — often already pretty clear based on use of personally owned devices — and match technology approaches to fit the need. We live in exciting times — times that demand sober, hard-headed judgment about technology initiatives, rather than a race after the latest silver bullet.

Richard Hoffman is a technology architect for Web/portal-based, mobile, and wireless systems, and a former technology editor for Network Computing Magazine.