Comments on: Virtualization: Changing the OS game, or not?

By: Alexander Manfrin

Alexander Manfrin — Wed, 13 Feb 2008 13:56:27 +0000

I respect the view of Mr. Linus Torvalds, he is a great programmer and who work with your kernel systems may respect this guy. But, the history show some phrases of other great men providing the future, like:

“There is no reason anyone would want a computer in their home.”
Ken Olson, Founder of DEC, 1977.

I think there is a world market for maybe five computers”
Thomas Watson, president of administration consul of IBM, 1943

Bill Gates in 1981 said:
“640K of memory should be enough for anybody”

By: R K W

R K W — Tue, 12 Feb 2008 17:34:21 +0000

Once everyone realizes and truly grasps that putting 14 OS’s on one piece of hardware causes 14 points of failure with hardware that is only partially redudant one realizes that we are merly making an OS (Virtual)the same as an application sitting on top of anyones computer. IF your laptop crashed hard.. you get how many application failures.. Virtualization is not so much about innovation as it is pure and utter hip shotting by CIO’s that think Virtuality means saving money..
At what? Expense of the customer? Its a niche and as a selling point to clueless IT depts that get hooked up on every bandwagon.. Ive seen it Ive watched it destroy and never improve the bottom line. IM still waiting though Im sure that somehow Titanium virtualization will save us all.. Its no different than having MS word on my machine woops the main memory went out on the Master Computer woops there goes 36 depts down… ohh no big deal (SLA’S go to hell)Virtualization its not about pushing us ahead its about a political game. technology wise there is no true innovation.

By: Rick Vanover

Rick Vanover — Fri, 08 Feb 2008 23:32:14 +0000

I think virtualization will actually accelerate the operating system. I know that in my recent jump into the virtualization space, my number of operating systems as well as quantity of installed systems (including virtual) has increased by about 10%. The hardware footprint, however, has decreased.

I think virtualization is a killer to the hardware sector for the departmental or application server role. Even for the small business, I think the free tools will accelerate so that even they can cash in from the savings perspective on virtualization.

Virtualization to me provides an isolation environment as needed for the service model. And that doesn’t remove the OS footprint in my situation.

By: Anil Desai

Anil Desai — Fri, 08 Feb 2008 16:48:46 +0000

I think this is definitely an interesting topic. One of the biggest advantages of virtualization is a layer of abstratction taht effectively decouples workloads (applications, services, and OS’s) from the underlying hardware platform. Had current operating systems been designed to “share” the hardware and to run isolated from the others, we really wouldn’t need the idea of virtualization (you could argue, of course, that the same would be true if the entire world ran the same OS).

What hasn’t changed is that we really do need the services that operating systems provide. We need a network layer, we need a user interface, and we need security. We need hardware management and the ability to support efficient memory and CPU utilization. Add to that common use cases such as web, file and print. Clearly, this functionaltiy has to be delivered somehow.

So how will all of this change the role of the OS? It’s hard to predict, but I think the general trend will be one that equalizes the playing field. Perhaps we’ll see standards emerge that allow common services to run on any platform or OS. That would pressure larger, monolithic OS “stacks” to become more modular. Of course, we’ve heard (and have seen fail) many such promises. Java on the desktop comes to mind. The end goal for most environments is to avoid complexity and to treat hardware resources as a seemless pool of capacity. For the foresseable future, that will be delivered through “standard” oeprating systems.

By: Schley Andrew Kutz

Schley Andrew Kutz — Fri, 08 Feb 2008 16:35:25 +0000

Wow it may seem arrogant to contradict Linus, I would have to say that he is comparing apples and oranges. Virtualization on the big-iron is not the same at all x86 server virtualization, there are many reasons for this, but one that is so immediately apparent is the initial acquisition cost of an x86 server versus a mainframe.

Saying x86 server virtualization won’t make a huge difference because it has failed to so in another form is akin to saying that the DVD format was destined to fail because LaserDiscs did. The idea is the same, but the implementation and associated costs are completely different.

All in all, I respectfully and humbly disagree with Linus on this (and that is not my usual tact WRT Linus’ arguments). I think that his argument is based on a weak assumption. That said, he may be right in the end. In another 10 years I’ll post a follow-up on this blog with the answer : )