In the old days, you could use just about any embedded operating system for your point-of-service (POS) installation. If you could load your application onto something akin to a cash-register and provide any level of functionality better than the old mechanical beasts, you could make a good market for your embedded system during the register replacement revolution.

As with the horses, however, the POS OS race is now the purview of specialists.

Microsoft is uniquely positioned to play the specialist role in this market. You see, point of service applications invariably have to talk fluently with desktop- and enterprise-level systems in order to deliver the kind of integration that customers crave. That means that it pays for your POS embedded operating system to have connections on the other end of the connection. With more than a bit of ubiquity in those markets, the big MS has a bit of a leg up. Ever hear of Windows? We thought so.

In many embedded markets, it’s difficult for a company with Microsoft’s commanding advantage on the desktop to bleed over into the embedded environment. The factors that drive designers to choose one embedded OS over another don’t weigh desktop compatibility that heavily – if at all. There’s also the tacit assumption that Microsoft’s vast experience with very-large-footprint operating systems will bleed over into bloat in the embedded environment. That’s not a good thing in a market that still carefully counts the cost of each KB of memory required to run their OS.

In the POS market, however, the stars are all in alignment with Microsoft’s strengths. For starters, POS does not have a penny-pinching problem in hardware costs like, for example, cell phones, PDAs, and MP3 players. Nobody expects to buy their next-generation cash register for $50 or get it for free if they sign up for one year of service. Because of that, system designers aren’t as concerned about memory footprint in their POS systems as in other embedded environments.

Second, a host of applications already developed for the desktop-based POS market (many businesses even use regular desktop computers as their POS systems) can move almost seamlessly over to Microsoft’s WEPOS (Windows Embedded for Point of Service). WEPOS is a domain-specific version of Windows, developed for the point-of-service market. It has a smaller footprint requirement (about half) compared to regular desktop Windows XP and is set up to run only POS applications, not general-purpose applications like, for example, Microsoft Office.

From a software development perspective, WEPOS uses the same development environment as other Windows-based platforms. That means that just about every software developer in the world has at least a minimum head start in writing applications for your WEPOS-based system. It also means that there is a wealth of software-IP already pre-brewed that you can use to further accelerate your time to market.

Of course, any time your software is passing around people’s money, security becomes almost as important (but not quite) as applications like avionics where you’re moving around people’s lives. WEPOS comes with robust security features that help protect devices against unsavory access and malicious use. With the preponderance of the world’s security hacks targeting Windows-based devices, it’s important to have well-engineered security features on your side.

UPOS, WEPOS, THEYPOS…

POS people have a plethora of peripherals that are somewhat specific to their domain. Retail users require hardware such as cash drawers, special keyboards, retail receipt printers, scanners, and scales. These peripherals increasingly are communicating using a standard called Unified Point of Service (UPOS), which encompasses two older standards – JavaPOS and OPOS. With built-in support for UPOS, most retail peripherals become plug-and-play if you’re using WEPOS. As a system developer, that means less work for you and more capability and flexibility for your customers.

The target devices for WEPOS are point-of-sale terminals like those used in retail establishments or self-service checkout systems, as well as in information kiosks. It isn’t set up for regular desktop use (Windows XP would be appropriate for that) or for use in very small-footprint devices like mobile POS platforms (Windows CE would be the Microsoft-based choice for those). It comes with support for network-based updates via Windows Update and other update services, as well as remote management, reboot, and installation.

Since its introduction, Microsoft has seen extremely aggressive adoption of WEPOS in the retail and service industries. Beyond cash registers, WEPOS has made significant inroads into information kiosks, where capabilities like audio and video delivery are much more in demand than in, say, ringing up that pound of broccoli at the grocers. With an almost-general-purpose OS like WEPOS, a kiosk can be quickly created that is extensible almost indefinitely, allowing a longer time in the field and more flexibility responding to changing customer demands without designing new hardware.

As the embedded systems market continues to mature, we are likely to see a continuation of the trend toward market-specific, off-the-shelf components like WEPOS. Already, there are specific embedded operating systems for military and aerospace applications, automotive applications, and a host of other high-value, high-demand areas. Unlike the desktop computing space, where interoperability of applications is the trump card that beats almost all other OS considerations, embedded devices are almost always task-specific from the start.

This less generic market makes more generic operating systems a much lower priority and gives an opportunity for product and market diversity that the desktop and enterprise environments will never see – all the more reason that it’s almost ironic to see Microsoft cleverly leveraging their dominance in the generic space to succeed so convincingly in a narrow one.