How Come We Don't Have a Standard Handset Architecture?

Should the handset be, literally, a tiny PC?

We wrapped up the Mobile Business Conference at Interop last week with three excellent sessions from Lisa Phifer on mobile management, both technical and operational. Among many other topics, Lisa mentioned the concept of personal liability, wherein employees use their personal cell phones for company business. This is already a common phenomenon, I know, but there are a good number of operational, practical, and financial concerns inherent here. I've just wrapped up a Farpoint Group White Paper on this topic, and I will have more on this for you next week. In the meantime, suffice it to say that my interest in the management side of things has increased dramatically in recent years, primarily because an effective wireless network management strategy and implementation is the key to lowering operational expense - which is usually a much bigger number than that reflecting the capital cost to buy the required equipment in the first place. As I often remind clients, capital cost is on an engineering/manufacturing price curve, which always declines. Operations, on the other hand, is on a human cost curve, which always increases. You do the math, but many elements of operational network management (including the unification of wire and wireless, which I'll return to in 2010) are regardless going to be huge themes for next year.

But what I really want to talk about here is an idea that came up in my panel session on operating systems for handsets. Why, I ask, don't we have a standard hardware definition for a cellular handset, along the lines of the PC? PCs are basically all the same today, and can run all manner of Windows, Linux, and even Macintosh operating systems - the last one often being called a Hackintosh, and which while formally not allowed under Apple's OS X licensing terms, is possible nonetheless. Handsets, on the other hand, are balkanized, with a tight coupling between OS and a specific hardware implementation. Sure, anything can be hacked, but I'm wondering why someone hasn't proposed at the very least a standardized hardware abstraction layer (HAL) for handsets, if not a full standard hardware definition. This would allow a variety of OSes to be ported, and offer consumers more choice and perhaps a longer useful life for their purchases via easier upgrades, or, for that matter, any opportunity for upgrades at all - often impossible depending upon the whims of a particular carrier. The hegemony of carriers over handsets must be broken, and this could be one way to do it.

I'm also wondering when virtual machines will appear on handsets, an effective way to separate business and personal use and thus enable more efficient mobile device management and personal liability strategies (with, to be fair, a few additional security concerns introduced in the process). OK, I don't expect to see either VMs or HALs become common on handsets in 2010. But these have been major factors in the definition of the path of the PC in recent years, and, if we look at the handset as a little PC (maybe not a good idea if you, like me, favor the thin-client approach), then both of these may become factors in handsets in the not-too-distant future regardless.

• Wireless / Mobile
• cell phones
• handsets
• Interop
• mobile operating systems
• virtualization