Enterprise Unix Roundup: Google's Linux Leap
|Main||In Other News|
The Spring of 2006 will certainly be remembered as a banner season for desktop Linux.
Too much hyperbole? Well, perhaps. But it's not as bad as the endless predictions of "200x will be the Year of Desktop Linux." Those predictions are really over the top and, we fear, are never really going to come true.
Don't get us wrong, we're hip to the whole prospect of a successful desktop Linux, which we do think is an eventuality. But we don't think anyone will be able to point to a given day, month, or year and say "that was when Linux on the desktop took off." By the same token, it will be difficult to ascertain just when Linux will be lauded as the successful operating system in the enterprise space. Although, because of the scarcity of enterprise companies, it will be easier to track that than overall desktop deployments.
But we digress. The reason we think this will be remembered as an "up" season for Linux is due to the efforts of one company to port some of its popular applications to the Linux desktop. We speak, of course, of Google, which announced the Linux version of its photo-sharing application, Picasa on May 25, and then quietly announced the beta Linux version of Google Earth 4 earlier this week.
Now, Picasa and Earth are very cool and rank high on the old Geek-o-Meter, but what do they have to do with enterprise Linux? Right now, nothing. But the porting of these applications to desktop Linux supports a trend that must happen if Linux is to be successful in the enterprise.
If desktop and enterprise Linux are to be succeed, then independent software vendors are going to have to get their applications ported to the operating system. Google, we would argue, holds the key to this adoption of Linux as a development target, since its size and popularity lends the concept of Linux applications a great deal of validity.
Let's look at it from a business standpoint. We talked with Google's open source guru Chris DiBona last fall. At the time, he explained Google's plan to develop apps for Windows first is simply a matter of numbers. There are more Windows installs, so developing for Windows first is more cost-effective. We certainly will not argue with that, and clearly other companies feel the same way. But although the number of Linux desktop installs is lower than than Windows desktops, porting Google's applications to Linux gives Linux a huge bootstrap up.
Before we get all gung-ho on this approach, we should mention there is another argument that begs the question, why do we really need native Linux applications anyway? The advent of more powerful thin-client technology means all we will need are applications on the server, a good thin-client browser, and a nice network pipe between them all of which the enterprise has. In this setup, no one will care what the client operating system is it's transparent.
This is all true, and we don't think these two scenarios have to be at odds with each other. Whether vendors take the thick- or thin-client approach, apps still must be ported to Linux (either on the server or the client), and there are going to be some applications where the thin-client approach might not be the best way to go. Vendors must still realize the benefits of Linux as a platform, and the involvement of high-profile vendors like Google will only make Linux that much more attractive.
Desktop Linux is not the only way Linux will get into the enterprise. It's server and clustering capabilities alone can do that. But desktop Linux, thick or thin, is more visible and will directly affect far more users than will running Linux in the backroom. Can it be done? We hope so, but if it happens, it will be somewhere where no Unix has gone before.
We may have just witnessed Google taking the first step into a whole new territory, one where Linux in the enterprise will be demonstrated on all of a company's machines.
» News flash: Sun is not wimpy.
We have no idea if its blades are, though. Seeing as the systems vendor claims the proud honors of being third to market, it's got a tough row to hoe. HP and IBM command more than two thirds of the space, and notables, like scrappy first-to-market RLX, have already been swallowed up.
The Sun offering, scheduled to ship later this summer, will use AMD Opteron processors and Sun SPARC chips. It will conform to the PCI-SIG's PCIe ExpressModule standard, which allows true hot-swapping of I/O components from multiple vendors.
In another corner of the Sun campus, the open source community hit the one-year milestone. Like any proud papa, Sun is boastful of Open Solaris' prowess: 33,000 downloads, 14,000 members (1,500 of whom are Sun employees), and 170 code contributions (of which 111 have been integrated into OpenSolaris),
Sun was less boastful about how community contributions are finding their way into OpenSolaris. Director of Marketing for System Software Chris Ratcliffe said it represents one of the largest challenges that the project faces.
And when asked, in a briefing last week, about the community's contribution to Sun's bottom line, Ratcliffe was less forthcoming, telling ServerWatch that cold calls are being made to the 4.8 million registered Solaris 10 users.
» Oracle's Linux embrace got a little tighter on Tuesday when it rolled out the Oracle Validated Configurations program. Although Oracle isn't packaging it as an official stack, when used together the applications offer an end-to-end configuration solution of hardware, storage, networking components with a Linux operating system and application software.
Oracle has validated and pre-tested the configurations and posted them, along with best practices, online for all to view.
Current partners in the program include Dell, EMC, HP, IBM, Network Appliance, Intel, AMD, Red Hat, Novell, Emulex, and QLogic.
» The Apache Software Foundation got a new benefactor this week. IBM will donate some IT management wisdom in the form of the WSDM (Web Services for Distributed Management, pronounced "wisdom") spec to Apache WebServices - Muse.
Ideally, WSDM will become a standard for the management interfaces of servers, routers, switches, and other IT hardware and software. WSDM began in 2003, when the OASIS Web Services Distributed Management (WSDM) Technical Committee was created. It was approved in March of 2005.
Apache Muse is the ASF's implementation of the WSDM standard. IBM worked on its own implementation of the WSDM standard and will now contribute code to enhance and augment Apache Muse.
IBM will offer support for the latest WSDM specification version, Release 1.1, supply pre-built code for all the WSDM-defined capabilities called "Helper classes," and offer better code portability to enable WSDM implementations to run on different Web Services runtimes.
IBM's WSDM contributions are expected to be integrated into an Apache Muse build soon. It will be available by the end of June, with refreshes every six to eight weeks after that.
>> To Main