Google Chromium OS – the Open Source Desktop Operating System for the Cloud?

In much of our analysis of  Cloud computing we are concerned with technologies from the center of the Cloud, The actual user interface to the cloud – the device through which cloud services are consumed – becomes almost an afterthought. And yet there is no doubt that for organizations that can provide both services in the cloud, and own the control points in the devices that access those services, the prize to be gained by delivery of services, and intermediation of third-party services (payment systems, advertising, product promotion etc.) will dwarf those of the traditional software or hardware vendor.

It’s a different way of thinking about the Cloud, where the starting point is not the DataCenter, or the IT service, or the “user”, but the direct delivery of consumer services. Into this vision plays Chromium OS (released to Open Source on November 19th).  Google devices delivering Google services (and other services intermediated by Google) from the Cloud to consumers.

This concept of a closed-box online service delivery actually pre-dates the widespread adoption of the internet. There is of course the Sabre booking system for the travel industry from way back in 1960 (currently revenues exceed $2.5Bn). Readers may also remember that the telcos wanted this model for the “information superhighway” in the early 1990s.  It looked like it was going to happen with America Online and also the Microsoft Blackbird project – delivering OLE objects into proprietary “browsers” across the proprietary Microsoft Network – cancelled in 1996. The closed-box models are now mainly in the entertainment (satellite/cable TV) and mobile spaces,  for example locked-down proprietary set-top-boxes with delivering subscription services and locked-down proprietary iPhone devices delivering applications from the Apple AppStore.

Before we continue here it is worth pointing out that there are two separate Operating System products.  Chromium OS, the Open Source software, and Chrome OS, the Google branded version of the Open Source software with additional proprietary elements (You can’t actually get this yet).  Then there are Chromium and Chrome (without the “OS” bit) the Open Source and proprietary versions of the browsers.  These are designed to run on both Chrome/Chromium OS and other operating systems. Chromium’s multi-process architecture and “task manager” feature starts to make more sense when you realize it is supposed to run standalone inside Chromium OS.

Chrome OS is probably best understood in terms of the smartphone – a sort of a Google Android for the PC. The first thing people ask about Chrome OS is “what happens if my PC is not connected to the internet?” The answer of course is that it doesn’t really do much. For most PC users that is a bit perplexing, but if a smartphone user asks “what happens if I don’t get signal?” they get the same answer and nobody seems perplexed at all.  Then people start asking about Word Processors and Spreadsheets or Graphics programs on Chrome OS, which also don’t work, at least not in the conventional way, but perhaps we will learn to live without them, just as most of us do on our mobile devices.

What we are essentially dealing with is the Thin Client device (Wyse WinTerm etc) we typically use to access Presentation Virtualization systems, packaged up as a consumer product. I actually have a Thin Client device in my home, but even I only tend to use it as a doorstop. The diskless workstation with only a web browser is also an old concept. It seems to be a recurring theme for those who don’t like Microsoft.   Larry Ellison of Oracle was talking about this as the Network Computer in the 1990s.  Dilbert had some fun with it at the time, comparing it to an “etch-a-sketch”.

So, if the closed-box service delivery model and the Network Computer weren’t a good idea in the 1990s, what makes the difference now?  Perhaps phones have gone this way, but surely not PCs?  The question here is partly technology, partly about user acceptability and partly about business.

As an enterprise, it is one thing to deliver your services through the public Internet, and quite another to deliver it via the iPhone.  In one case you are dealing directly with the customer, in the other you have introduced Apple as an intermediary, and your relationship with Apple determines the level of visibility you have on the AppStore.  Google may present itself as a benign intermediary, but to what extent are services delivered via Chrome OS to be under its control?  Is the Internet to become a giant Disney World, where a seemingly-friendly corporate entity determines which products are sold and takes a profit from every apparently-independent hotel, cafe and shop?

It’s not something consumers want, nor business. One thing that is supposed to make us comfortable with this is Open Source, and this is why the release of Chromium OS to the Open Source is relevant.  Certainly if this were Windows 7, it would be a very important day.

Actually Chrome/Chromium OS is just a Linux distribution (which we believe is Ubuntu-based) stripped down to bare bones.  It contains bootloader, kernel, basic GNU packages to support hardware, system/process handling and to support the network.

And you can add stuff back, so you could have a Chromium OS-based distribution that ran OpenOffice from local storage.  But you wouldn’t, that would be mad.  The Open Source piece here is actually about responsiveness to OEM hardware manufacturers. They can more easily build their own device drivers and suggest customizations in support of specific hardware features if they have access to the source code.

This is not about Google creating a competitor for itself. At the end of the day, there are unlikely to be other Chromium OS-derived Operating Systems set up to deliver Google-competitive services.

In amongst the orchestrated hype, there has been some negative stuff about Chrome OS. Indeed there is little here that moves us on from the Network Computer, the Thin Client device I use as a doorstop, or indeed Dilbert’s “etch-a-sketch”.  If it succeeds it will be initially driven by a complex set of market dynamics.  For example, it may get traction in cybercafés and web kiosks, (Microsoft has helpfully just decided to charge an additional fee for every existing “rented” Windows Licence), or there may be one combination of Service and Device (like a Google version of the Kindle) that drives people to buy it rather than a PC.