There is no such thing as a cloud operating system — and we're some way off even needing one, says Lori MacVittie.
Google's foray into the operating-system business was barely public before it was being hailed as the latest cloud operating system, with some industry watchers saying this product could be pivotal in the cloud revolution.
To be fair, Google's Chrome OS is not the only operating system to which the cloud handle has been attached. It is merely the latest in a long line of attempts to capitalise on the growing interest and hype surrounding cloud computing.
Novell, Dell, Microsoft — in fact, anyone who is anyone with a stake in operating systems has been mentioned at least once in conjunction with a cloud operating system.
There is no such thing. It is a myth existing entirely in the minds of those who cannot seem to get enough cloud in their daily technology diets. And the problem in perpetuating that myth is that it continues to confuse an already confused market.
Marketing jargon
Millions of IT professionals who should be able to separate marketing jargon from technical definitions are confused and unable to articulate a definition for the cloud. It borders on exploitation to label something a cloud product just to make it appear as if participating in the cloud involves special requirements.
A typical IT department spends 80 percent of its budget keeping the lights on, managing, patching, cooling and otherwise caring for servers and applications in their own datacentres, according to analyst firm Gartner. It makes sense to move some of this burden to the cloud, so it is shocking to see vendors perpetuate confusion for their own gain.
A cloud operating system is unnecessary at this point in the evolutionary cycle of the technology. Part of the allure of the cloud is its reliance on the internet, which, as everyone knows, is built on established, well-supported standards such as TCP/IP.
Applications served via the internet are almost always transported via HTTP, which completes the standards-based trifecta that has long supported the web, and now makes the cloud realistic.
Nothing special is required for an operating system to support the cloud. No new standards, no new capabilities, and no new functionality at the operating-system layer.
Cloud computing, as implemented today, is evolutionary — not revolutionary — and combines the best of hosted services with application management at the provisioning layer. There is very little about the operating system that needs modification even on the server side — that is, at the cloud provider — and absolutely nothing that needs changing in the deep underbelly of the client operating system.
To be fair, so-called cloud operating systems may be delivered with cloud-specific management and development tools as part and parcel of the distribution, but these are applications and are almost certainly not a part of the operating system itself.
Core operations
Operating systems focus on I/O, file, CPU and memory management, and on scheduling the core operations into which all applications ultimately break down. Operating systems are granular, tightly focused systems that know nothing about clouds or networks or even applications, except as they relate to the next machine-code instruction that must be executed.
As cloud computing moves towards what purists might deem real cloud computing — the provisioning at the compute-resource level or distribution of applications at the functional level without modification — it may be that operating systems will need to change and evolve and adapt to the specific needs of such an environment.
When the operating system takes into consideration other CPUs and banks of memory on nearby machines when it schedules an instruction, then we will need a cloud operating system.
Lori MacVittie is responsible for application services education and evangelism at application delivery firm F5 Networks. Her role includes producing technical materials and participating in community-based forums and industry standards organisations. MacVittie has extensive programming experience as an application architect, as well as in network and systems development and administration.
Talkback
Gael Duval, one of the founders of Mandrake/Mandriva Linux, has established, what I would consider, an actual cloud experience. Located at http://www.ulteo.com/
It is a complete operating system and storage online for use by anyone. There is some free parts, or you can subscribe at different levels. This is what i envisioned the internet to be, back in the early nineties.
Finally some rain on the cloud BS!
Remember how virtualization was going to solve all our problems 2 years ago?
Remember how multi-core cpus were going to solve all our problems 3/4 years ago?
Multi-/Hyperthreading etc?
2GHz CPUs? Pentium 4? Itanium.....etc etc.
I suppose the next BIG COMPUTER TECHNOLOGY will be tectonic plates or Crustal Dynamics or some such other drivel.
Far be it from me to sound like a tired cynic (moi?) but isn't this always the way? The industry excels at producing 'solutions' that fix problems generated by the lat generation of fixes. It will never end...
(I'm sure there's a book in there somewhere).
It's in its infancy yet a good few years will haft to pass before any one know for sure where this will all end up, they will be some uses for them and them uses will probably turn out to be in the least expected area's.
Not sure the cloud operating system is a myth. Since myths are stories that often have some basis in fact but are actually exaggerated for dramatic effect. Actually it would be more accurate to describe the cloud operating system as an ambition or a goal.
And why is that a goal worth striving for? Since its development will actually have massive economic benefit.
Take an example. A retailer may choose to outsource eCommerce order fulfilment to a service provider who already has the warehousing, logistics relationships, etc, in order to avoid the massive capital expenditure requirement for setting up a warehouse. The warehouse operator's clients share in the benefit of that investment.
In the same way many technology rollouts are massively capital-intensive. With a functional, reliable cloud (which includes a reliable, standards-based cloud operating system) that investment will be reduced and the resources of the world better used.
The cloud operating system is an ambition not a myth, and a worthy one.
"With a functional, reliable cloud (which includes a reliable, standards-based cloud operating system)"
This is just one major stumbling block. Already, various vendors are pushing their own ring-fenced non-compatible offerings. I have no doubt at all that this will devolve into the usual over-priced, anti-competitive mess.
The other 'Elephant' is the security risk of having your private, commercially sensitive data held by a company you have no control over, in a country which may suddenly decide that to protect its own security it needs access to any data held on its territory - or even the territory of its 'partners'. Sound familiar?