Virtualization? Yes. The Cloud? Maybe Not Quite Yet

Arthur Cole

If virtualization is one of the underpinnings of the cloud, then how do you explain the relatively broad adoption of the former and the relatively slow adoption of the latter?


According to a new report from the Data Center Institute, less than 15 percent of the group's members say they have deployed any form of cloud computing, numbers that roughly match the IT industry at large. And while the reasons for holding off include the usual concerns of security, availability and overall reliability, it is also possible that there is another, more fundamental, deterrent to cloud deployment: It just isn't as easy as some of the top proponents say it is.


This is especially true with private clouds. On the one hand, you gain a certain level of control that gives you more sway over security, availability and reliability, but as analyst Bill Claybrook pointed out recently, private clouds come with a fair amount of baggage, including cost, integration with both internal and external environments, and constant pressure to maintain state-of-the-art infrastructure. And in the end, the processes through which to transition from a traditional data center to the cloud will be as varied as the IT industry itself.


If there is any advice to be given when it comes to launching a cloud, it would probably be to start with baby steps. For most, that would mean a small test environment that allows existing personnel to gain experience in the fluid nature of the cloud. Fortunately, there are a number of simplified platforms geared toward low-cost cloud experimentation. Platform Computing, for instance, has the Platform ISF Starter Pack, an end-to-end software stack to let you set up a cloud sandbox in less than 30 minutes. The system offers application workload management across multiple virtual machines and includes features like self-service and auto-provisioning for platforms ranging from ESX and Hyper-V to Red Hat Satellite and IBM xCAT.


From there, though, it takes quite a leap of faith to jump from an internal to an external cloud. But at least a new generation of software will make the operational transfer a little easier. A company called Makara has developed a means to deploy private clouds on public platforms like Google App Engine and Microsoft Azure. The company recently developed a new abstract layer for Amazon's EC2 that essentially brings it under control of private cloud management structures like Eucalyptus, Ubuntu Enterprise Cloud and even vSphere. In this way, cloud management becomes more about application and service levels rather than the underlying infrastructure.


These kinds of tools certainly take some of the sting out of launching cloud services. But the fact remains that we're talking about a fundamental shift in the way IT services have been provisioned and deployed for the better part of the past three decades. That kind of change does not come easily. And the cloud industry will just have to accept the fact that many enterprises are determined to take it slowly.



Add Comment      Leave a comment on this blog post

Aug 27, 2010 4:11 AM Edward Newman Edward Newman  says:

This post captures what I'm seeing from EMC customers. They are adopting the baby steps approach, analyzing their application portfolio to determine cloud and virtualization candidates while developing a catalog of consumer and operator oriented services published via a portal. They then have a prioritized approach to dealing with virtualizing and replatforming their applications while getting a new cloud-like front end to service requests. In this way, they can execute according to cost and risk and demand management for the new services to drive the virtualization, automation and orchestration of the environment to implement the private cloud.

Edward Newman

Director, Cloud and Virtual Data Center

EMC Consulting

Reply
Oct 29, 2010 8:24 AM Luke Vorster Luke Vorster  says:

I agree. I believe it boils down to human capital boundaries and limits, more than anything. Is it virtual enough? Will the code even utilise the could as 'it' scales? How much does it cost to prototype collaboration portals on cloud platforms? How can the total cost of ownership for a private cloud be negligible on the risk-management matrix? Is the private cloud an asset or a liability? Where is the human capital capable of developing assets on cloud platforms? Why don't they work for us?, etc.

The most interesting cloud technology I have come across recently is AppScale, an OpenSource implementation of Google's GoogleApp platform that can allow the GoogleApp solution to be cloud-i-fied without code change (apparently). It is now a Google Code project:  

http://appscale.cs.ucsb.edu/

http://code.google.com/p/appscale/

Reply
Oct 29, 2010 8:36 AM Luke Vorster Luke Vorster  says: in response to Luke Vorster

Did I mention that AppScale allows you to take your GoogleApp, without code change, and deploy it on your own private cloud?

Or you could build it on AppScale from scratch... maybe deploy it on GoogleApp for cleanroom testing...

Java/Python APIs supported...

It executes automatically over Amazon EC2 and Eucalyptus as well as Xen and KVM...

Supported Datastores

   

  • HBase

   

  • Hypertable

   

  • MySQL Cluster

   

  • Cassandra

   

  • Voldemort

   

  • MongoDB

   

  • MemcacheDB

Anyway, I can finally see a roadmap for my own private AI research cloud, and hope to mitigate the breadth and depth of the current skill matrix required in human capital to deliver cloud platforms as assets and apps with longevity...

(so romantic!)

Reply

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.


 

Resource centers

Business Intelligence

Business performance information for strategic and operational decision-making

SOA

SOA uses interoperable services grouped around business processes to ease data integration

Data Warehousing

Data warehousing helps companies make sense of their operational data