Private Clouds: Not as Easy as it Looks

Arthur Cole

A strange dichotomy is forming around the emerging market for private clouds.

Despite, or perhaps because of, the relatively slow uptake of the technology so far, vendors continue to introduce ever-more-simplified cloud startup packages. At the same time, though, there seems to be a growing community of integrators, consultants and other experts who insist that the cloud is a lot more complicated than it's made out to be, and getting started is the least of the difficulties.

So which is it? Is the cloud merely the next step in virtual evolution, available with little more trouble than a standard software upgrade? Or is it a major restructuring of all things IT that will require years of architectural re-engineering. Or is it both?

For starters, let's look at some of the latest cloud-in-a-can offerings. The most recent is HP's CloudStart, a mix of hardware, software and services that the company says can get business up and running as cloud providers in less than a month. The package is open source, so it should accommodate a wide range of third-party solutions, such as portals, public cloud services, billing systems and management stacks.

The package relies heavily on HP's Cloud Service Automation toolkit, which covers such functions as provisioning and security, as well as the scaling of services and applications to meet demand. The hardware backbone consists of the HP BladeSystem Matrix augmented by the StorageWorks portfolio, all of which provides clients with tools like one-touch provisioning across various infrastructures, applications and services. The company claims an 80 percent reduction in the provisioning process.

Over at Citrix, the focus is on improving interoperability, scalability and, most importantly, self-service in its OpenCloud platform. To that end, the company has acquired VMLogix a virtualization-management company specializing in cloud environments. The company offers a VM lifecycle-management system that will add a self-service interface to the XenServer platform. This should give private cloud clients the same kind of access and management capabilities found on public cloud services. Citrix is also developing an open management system called OpenStack that should present a unified control point for public and private cloud architectures.

But according to some cloud experts, the thing to keep in mind about the technology is that launching the cloud is, literally, just the beginning. According to Dick Weisinger, chief technologist at Formtek, the real benefit of cloud computing can only be realized if the system is designed and architected from the ground up. And that requires a fair amount of specialized know-how in areas such as auto-provisioning, identity-based security, governance and multi-tenancy.

In truth, according to Computerworld's Bill Claybrook, the complexities involved in setting up the cloud pale in comparison to the challenges of integrating it with both legacy hardware and systems, and the public resources that will undoubtedly form part of the overall data center ecosystem. A transition of this magnitude can only be accomplished over time, most likely in fits and starts, which means the true benefits of the cloud will probably be delayed until the rest of the enterprise catches up.

So which version is real? Probably a little of both. On the one hand, there's enough in the new cloud platforms to simplify the initial deployment phase to at least increase your in-house experience. But we're still talking about a lengthy and complicated transition to a new data environment that ultimately should usher in a new era of portability and flexibility.

That kind of change won't come easily or cheaply.

Add Comment      Leave a comment on this blog post
Oct 29, 2010 7:40 AM Luke Vorster Luke Vorster  says:

I have made various comments RE cloud computing on the articles provided below.

I foresee a radical de-scoping of what cloud computing actually turns out to be-when-its-at-home.

The main points are:

"I foresee a trend of accelerating virtual technologies with 'peripheral' HPC systems...i.e.the cloud is a machine, so we should treat it like one - optimise where we have to and can afford to.There is no way anyone can afford to optimise every aspect of a computation machine for all cases."

"Mission critical is not a generally solvable space - the bar raises as we can do more - it will always be out of scope of general solutions, virtualisation included."

"In my humble opinion, we still haven't overcome the software crisis from the tail end of the previous millenium...It doesn't matter what machine you use - it's how you use it.There are many types of machine, and many types of language (and programming paradigms), because there is simply no silver bullet.The complexity of internet apps, HPC apps, and 'infinitely scalable' cloud-based apps is the real problem.We have the vehicle, but no idea (really) how it differs from its predecessors.We don't even know if it is, in fact, any different!"

"It is not at all surprising that enterprises are wary of public clouds - they have never been otherwise with previous large-scale inter-networked technologies...Let's get real - IT business is about racing for the short-term gain over competitors.It is a primitive biological urge, when a species finds a comfortable environment, to eradicate all other species in that environment.The back-lash, however, is going to sting like a genetically modified killer-bee from the depths of hell itself - the compound total cost of ownership, the inevitable fragmentation (and lack of motivation to produce open standards through collaboration), will breed a new kind of information war, putting the power of technology back into the hands of the custodians of cyber-space to try to patch the problem in per!"

"I think we're just scared because we have no real experience in the reality of public clouds.In-house, private, obscure, and home-grown systems, however, have been around since before, but not after, the brith of the Internet.Development cost of using private clouds is surely much higher than public ones?This is for reasons that I could only explain in volumes of books more dense that Knuth's Art of Programming!"

"I think the Internet is a good model to take ideas is it coping with public and private issues?In the previous 15-20 years, or so, it has been blatantly clear that the US internet industry has only proliferated the way it has due the convenience of inter-state law.When US companies want to deploy successful (in US/Canada) business vehicles in Europe, or other equally large domains, it becomes largely apparent just how inexperienced these companies are in truly public cyber-environments.(I have been in projects where we had to bring in the big five just to help us knock out a temporary legal framework for inter-continental online business and taxation models) In the cases I have been invovled in, most systems required a complete re-analysis, and much money was spent on mentoring these companies in the ways of the rest of the world.I am not slagging the US here, I am just saying they have had it lucky for so long and didn't realise it until the late 90's... Reply

Oct 29, 2010 7:40 AM Luke Vorster Luke Vorster  says:
How is a public cloud different from a public search engine?How is a private cloud different from a national phone/sateliteTV/DVB broadcasting platform?Clouds are not a new concept at the political level.I have one positive point RE public clouds:TOTAL COST OF OWNERSHIP!!!Does anyone really want to bear the TOC of a private cloud?Haven't we learnt (over and over and over), that the long term issues invovlved in sustaining large scale computation resources can only be justified if the resources are seen as liabilities and the work done on them extremely lucrative?Any comments welcome, as I am concerned about the future of the internet.Private clouds sounds like fragmentation in terms of technology and human demographics..."

Full article comments here:

Oct 29, 2010 8:14 AM Luke Vorster Luke Vorster  says: in response to Luke Vorster

The most promising private cloud platform to me is the AppScale project from University of California Santa Barbara.

It is an Open Source implementation of Google AppEngine, called AppScale. Apparently, you don;t have to change your GoogleApp code to run it on AppScale. There is Java and Python support, and, at a 'garage-lab' level, can run on xen-based Ubuntu. On the other hand it can run on AWS/EC2 and Eucalyptus...

Though I am well experienced with clusters, multi-core, GPGPU, I am building an AppScale system in my lab for parallel AI research...


Post a comment





(Maximum characters: 1200). You have 1200 characters left.



Subscribe to our Newsletters

Sign up now and get the best business technology insights direct to your inbox.