ExtremeLabs, Inc

This is going to be easy.

#1: ARM gives Intel/AMD/VIA a run for its money

Yes, lowly RISC processors are coming in huge densities, all using much less power. Seems counter-intuitive, but there’ll be multiple OS changes to accommodate the density possibilities, mostly in interesting server combinations.

#2: Social Sickness

Twitter, Facebook, G+, all interesting. A unified user interface that just does it will become popular, and will separate the truly social from the casual. The client piece will be fought tooth and nail because it disrupts the business models of the social site vendors.

#3: The Great Server Makers Will Fall and NOCs Will Change Forever (for the better)

Servers aren’t what they used to be. High density servers will steal the market. NOCs will start to look like clouds, and many skies of clouds. Traditional server makers will have some problems getting to market and learning how to make margins based on differing densities.

#4: Client Hardware Will Become: Anything

I want my (fill in this blank) to work on the darn network. Make it so. Desktop hardware doesn’t go away, there is no death of anything. Instead, accessibility will become the huge challenge for large organizations, who will become driven by user pressures. This will lead to

#5: Security Rethought

Unified security will become the norm with (gasp!) interoperability. Lacking that, we have chaos. Maybe the government will wake up, too, but I doubt it.

And finally, Obama wins, and someone will owe me yet another beer to pay me off.

ChromeOS: Not Yet For Civilians

For now, ChromeOS is pretty boring: after a bit of work, you get a fast loading browser, the ability to change a few themes, and if you’re a developer, access to a new platform. Google’s new ChromeOS is an operating system, but it’s strongly browser-based, and therefore is different than Mac’s MacOS and Windows– even from Linux. It’s designed to be extremely lightweight. Some call it a web appliance, and that’s partially true. It uses no hard disk, but that’s a red herring, as it uses and needs storage– but not much unless the usage profile of ChromeOS changes. For now, most people can ignore it. There’s nothing revolutionary about it.

There’s also a lot of confusion from Google, as Chrome-the-browser and Chrome-the-OS are used almost interchangeably. We even found this to be true when downloading and compiling ChromeOS for use. The two are separate entities, one the browser application, and the other is the appliance/OS substrate needed to anneal Chrome-the-browser to hardware and give it tools and native apps.

Initially, Google recommends running the ChromeOS guts as a job under Ubuntu Linux. Eventually, ChromeOS fits into a category called a semi-persistent instance. It uses no traditional hard drive for storage, and essentially launches as a completely self-contained browser-only instance. We downloaded Ubuntu 9.10 and installed it as a virtual machine. Indeed ChromeOS eventually becomes an appliance. Handily, Java and Flash are supported. Only the oracles know if Microsoft’s Silverlite components will run.

One of the closest conceptual operating system schemes to ChromeOS is Sun’s Containers system that works with Oracle/Sun’s Solaris, or perhaps Parallel’s Virtuozzo. Both of these ‘sandbox’ applications into their own metaphorically walled areas. Core operating system resources are available, but the applications are highly isolated from the operating system and its components– as well as other running applications.

ChromeOS seems poised towards very small systems (think netbook or notebook or OEM refrigerator display-type) web application running machine. It lends itself towards use of Google’s own web applications, but could conceivably be poised as an Android (Google’s mobile/cell phone OS) for larger platforms. Common interoperability between ChromeOS and Android are unknown at this moment, and probably is doubtful. Both of these operating systems are community/open operating system development platforms. None of them will pay developers the first dime for at least a year– which is Google’s ostensible target release date for ChromeOS.

The much pre-announced ChromeOS is now available for download as a kit rather than something consumers can munch on. If you’re a civilian, don’t even try, as the OS has to be compiled, and even for an operating system, it’s not a pretty picture. It’s do-able if you’ve ever compiled Linux or BSD from a CVS (not the drug store).

What ChromeOS Does

ChromeOS is a cloud operating system in the sense that most applications and storage will be used on hosted computers somewhere else, the kind that comprise corporate data centers and hosted/cloud environments. The light weight part of it has been done as well. FastScale, now owned by EMC, has a black belt in stripping the unused pieces from current operating systems into lightweight server operating systems in virtual environments. Nlight does the same thing for Windows clients, and is used frequently for virtual desktop interface environments that use ‘terminal server’ like virtual machines for remote users.

What Google might be trying to attempt is to bring one of Scott McNealy and Larry Ellison’s ideas to fore: the network as the computer. Of course the fact that Google’s Eric Schmidt once worked at Sun and Microsoft rival Novell (now owner of popular Linux distro Novell/SUSE) probably has nothing to do with this invention in Chrome. After all, be it ever so crumble, there’s no plate like Chrome……

Everyone’s regrouping.

In one corner, Microsoft may have a winner after the disaster that’s Vista. We’ve looked at Windows 7 at http://www.networkworld.com/slideshows/2009/011509-win-love-hate.html#slide1, and then again this week at http://www.networkworld.com/news/2009/030209-microsoft-windows7.html?hpg1=bn. It has a chance.

But Apple’s also releasing Snow Leopard, the Intel-only update to Mac OS, shortly. Microsoft’s most feared competitor only works on Apple hardware, and that leaves HP, IBM/Lenovo, Dell, and a raft of hardware makers happy. These hardware makers have been deeply disappointed in Vista sales, and they’re occasionally turned to variations of Linux.

One variation soon to arrive will be Novell SUSE Linux Enterprise Desktop. The glory of this version is all of the extra stuff Novell has written to do enterprise management of Linux desktops. If rumors and advanced features seen elsewhere are an indication, SLED (as it’s called), might be a low cost, high-feature way of beating Microsoft at their own game. Microsoft typically doesn’t like that, but Novell’s an arms-length partner of Microsoft in a number of lucrative cross-support endeavors.

And all of this might not make much difference, and here’s why. Application virtualization is becoming both stylish and is starting to meet the test of audit, security, and even availability metrics among the enterprise cognoscenti. An app for your desktop, no matter what it is, can be loaded out of the Internet as though the Internet was your local hard disk. The plumbing for this can be simple or complicated, but also entirely masked from users- and the guts live in ‘the clouds’. The clouds in this case are a rising number of Internet hosting organizations, now with a la carte offerings of varying server system hosts online. It’s getting very easy.

Even the desktop resources one uses are now becoming virtualized as well. No need for a fat hard disk and lots of processing power, when your netbook or notebook is getting the desktop and even application hosting on some way-fast machine sitting on the Internet. Accessibility goes way high. So do concerns regarding availability, security, and little things like audio and video virtualization.

Virtualized desktops tend to make operating systems somewhat irrelevant, as the one-operating-system-on-one-machine metaphor slowly starts to disintegrate. The wars are here. May the all-important users win.

After a revealing guess based on the Windows 7 beta, the press and/or other pressures brought to bear the answer to the question: how many versions of Windows 7 will there be? The answer is just six, down from 10+ of Microsoft’s failing desktop product, Vista.

Much insanity ensued after Vista’s release. The problems are well documented and are now a matter of history. Also a part of history is the fact that Vista hasn’t sold well, and customers have stayed away from it in droves. Indeed the sales of Vista to date are largely the result of OEM agreements that Microsoft has with the hardware computer community.

The lack of sales has given rise to competitive operating system sales advances, although these advances aren’t going to shake Microsoft’s market dominance– for better or worse. But what has potentially happened is something that no one planned on– a vision of the drop of perceived cost versus value in operating systems.

When people buy a computer system at retail, the cost of the operating system is hidden. It’s not broken down like HP Notebook (add $20 for Windows, $10 for Linux, and of course Mac OS is missing from the availability list). The value of Windows is somewhat known, but it’s fuzzy and vendors have made it fuzzy because of their agreements with Microsoft.

No one buys a notebook, takes it home, then makes a selection to load an operating system choice. It’s not done. A truly barebones system would require end-users to somehow purchase, download (or enable) a selected operating system. Most civilian users don’t have time for that– the size of the operating system needs to be resident on the user’s hard disk at purchase time– not as a download.

What this does is to devalue the perceived cost of the operating system. Familiarity for many user purchasers would be with Windows, and they might wonder why it was missing on their computer. Simply installing Windows on a new computer system takes a while, for reasons that are largely unclear. The state of a Windows machine under brand-new-condition has changed quite a bit when the purchased machine is ‘bundled’ with Windows. It takes less and less time to install the OEM/bundled version of Windows, but it’s still time-consuming and potentially fraught with user errors from incorrect choices. As users become more savvy, these problems are reduced, but civilian users are still subject to many potential problems with new machines. The state of such OEM products is largely outside of Microsoft’s control, yet Microsoft and the OEM vendor must support the satisfaction of the initial install of the operating system at purchase-setup time. These influence the choices we see today.

The value of a captive OEM operating system is ill-defined. It’s no wonder that establishing value for them is often only done in terms of upgrade price. Without compelling reasons to upgrade, any OS vendor won’t get upgrade revenues because value defines demand.

One wonders. After watching Windows from 2.0 through 3.0, 3.1, 3.11, WfW, Windows 95, Windows 98, Windows 98 Second Edition, Windows ME, Windows NT 3.x, Windows 4.0, Windows 5.0 (I have the T-shirt), Windows 2000, Windows XP (various flavors), Windows 2003, Windows 2008, Windows Vista, Windows Mobile, and variations, service packs, and so on, I watch with no lust for Windows 7.

When it was announced, Windows Vista was going to be the result of a project to end all Windows projects, with billions of dollars spent to develop what was to be a historic milestone in operating systems. It was and is a historic flop.

The basic architecture of Windows has seen only three generations. The first generation was delivered as an application that produced a graphical user interface and common application programming/use model to what was then known as DOS.

The second generation actually was an operating system that lived as the host operating system, first seen in Windows NT– itself a variation on the beginnings of OS/2. Up until now, both generations used a model that had configuration files, first only on disk, then disk/memory to permit various facts/settings to be stored. This methodology allowed programs or even the user to modify often cryptic settings that could be allowed to control the machine.

The third generation of Windows separated the user mode accessibility from these settings, and prevented viruses and malware (unless a machine’s user specifically permitted this) from controlling the operation of the machine. This allowed, in Windows XP SP2 (and Vista versions) to become vastly more immune from threats than before. Users of Windows XP SP2+ and Vista editions were role-demoted, separating them from their capability from making crucial systems changes unless under specific conditions where the user was made aware that this was about to happen. Windows was comparatively safe for the first time in almost a decade.

Windows Vista has flopped for many reasons, most of them related to demand and the perception of value. The numerous ‘Vista-Ready’ programs were found to be disingenuous and hardware upgrades were needed to make some version of Vista work, while other versions were ‘slackened’. The problem with too few device drivers (meaning consumer peripheral choices) also darkened Vista’s skies.

Another perception was that if things weren’t broken, why ‘fix’ them with Vista, when Vista represented possible application problems (including very old ‘DOS’ games), things that wouldn’t work, additional and onerous digital rights management/piracy prevention, as well as a staggering variety of the Vista versions themselves– all poorly differentiated.

Windows 7 must use most things that made Vista and Windows XP SP2+ versions safe. Some of the drudgery therefore must continue. Microsoft can’t go backwards in some of the safety components that it introduced into Windows to fix long standing problems. What they can add or subtract or make right with a new Windows consumer version (ostensibly the ‘Windows 7′ now in development) is unknown. Will they add more of the products in for ‘free’ that they now charge for? How can you add value when things are perceived to work well with existing hardware? Has market saturation occurred for Windows?

For the past dozen+ years, Microsoft has breathlessly teased and taunted developers and the public for each new version. Now, the direction isn’t clear. The world has changed in those dozen years, and the vision (or threat) of a new Windows version is now held against the metric of the release of competitive operating systems that don’t carry the burden/legacy of Windows versions that didn’t work or were plainly unsafe and poorly designed.

This coming version, whatever it’ll be called, will likely be one of the most scrutinized; skepticism regarding value is at an all-time high, world economic malaise or no.