Google reinvents network computing
Where Netscape feared to tread
Google has launched itself into the network computing debate by announcing its own operating system, proposing that the browser is the platform and the network is the computer.
We've been here before, several times, most notably during the first dot com boom, when Netscape (the first commercial browser developer) was seriously considered a threat to Microsoft's revenue stream with its vision of putting everything in the browser.
Observant readers will have noticed that Netscape wasn't entirely successful in realising its vision, despite unmitigated support from Oracle, Sun, and IBM to name a few - so what exactly does Google think has changed this time around?
Back then, Netscape didn't have an operating system. But Sun did, and there were plenty of alternatives. Sun's JavaOS was, supposedly, developed completely in Java (when pushed, Sun executives would admit such a thing would be impossible, but it didn't stop them making the claim for both JavaOS and the HotJava browser*), but enough of it was developed in Java to ensure that it suffered from intolerable performance issues that were never resolved.
Java was a key enabler of the Network Computing vision - the ability to download applications for execution in the browser, or browsing application, was heralded as the driver of the next information revolution. Users would subscribe to applications, local storage would become a misnomer and Microsoft would disappear into the history books.
The NC project with which your author was involved was based on the VXWorks OS, using that company's implementation of Java to create a box for Swiss Telecom (now Swisscom). It worked well enough, and executed applications written by developers who had never seen the hardware - something of a revelation at the time - but it was not the only NC project to flounder before making it to the market.
Makers of dumb terminals leapt into the network-computing fray, seeing it as a final revenge on the desktop computers which had so damaged their business, and Oracle put its considerable weight behind the idea with visions of huge databases powering the
cloud network computing model. But with only the Java mantra on which to base designs there was a great deal of floundering, not least down to the lack of a web browser written in Java. (HotJava rapidly became a bloated joke, while the ICEBrowser was fast enough to get bought by Sun and never really emerged again.) This rendered most of the prototypes unable to use the network that was supposed to be their platform.
Google is planning a standard platform, a stripped down Linux that it will develop over the next 12 months along with a Chrome port. Java is no longer in the picture, having been supplanted by AJAX, which (if one is brutally honest) offers much the same functionality as Java 1.1.7 offered last time around - more if one includes Google Gears and perhaps some more extensions such as those envisioned by the mobile industry.
Google also has more money than Netscape ever dreamed of - lots more money. The company isn't pitching Google OS as a Microsoft-killer - not this time around - but the subtext isn't hard to spot. "[Today] the operating systems that browsers run on were designed in an era where there was no web" claims the official release, which is clearly bollocks, but engaging bollocks none the less.
Google's decision not to use Android as the underlying OS is understandable - the aims are different - but maintaining different OS versions will be a constant drain on the chocolate factory's coffers.
That drain could be offset by getting Oracle involved: as recently as last month Larry Ellison was waxing lyrical about the possibility of Oracle pushing out a network computer, but it's far from obvious that Google would want Oracle involved, especially as Mr. Ellison would almost certainly want to see Java back in the mix for any thin client.
The arguments for local storage and processing remain the same: vendors make money selling applications, users want control over their environment, network-dependent systems are only as reliable as the network on which they are based. But this time the networks are more reliable and users are happier than ever to use networked storage and backup systems, so perhaps Chrome OS has come at just the right time.
If so, then Microsoft should be worried. The network computer has been a long time coming, but with Google's backing it could yet be the platform that finally challenges Redmond. But a note of caution - your reporter fell for that last time, and the conservative nature of IT users and admins shouldn't be underestimated. ®
* Which could mysteriously print, amongst other things, despite the lack of suitable APIs at the time.