Focusing on open APIs for enterprise applications

Open Web Magazine

Subscribe to Open Web Magazine: eMailAlerts Email Alerts Weekly Newsletters
Get Open Web Magazine: Homepage mobile Mobile RSS Facebook twitter Twitter LinkedIn

Open Web Authors: Jnan Dash, Jayaram Krishnaswamy, Bob Gourley, Kevin Benedict, Pat Romanski

Cloud Computing: Article

Is Nick Carr Right About Cloud Computing?

Will traditional corporate IT really move dramatically to the cloud, as Carr predicts?

By Chris Fleck	Article Rating:
December 8, 2008 01:45 AM EST	Reads:	32,597

Related
Print
Email
Feedback
Add This
Blog This

Chris Fleck's Blog

There is a lot of hype and reality in the world of Cloud computing today. On the hype side there are numerous data points; VC investments, M&A, data centers under construction, new conferences and events. On the other hand there are now many examples of reality as well. SaaS offerings like Salesforce.com and many HR applications plus web site hosting have established themselves as a legitimate part of the IT portfolio of many corporations large and small.

Adding fuel to the fire is extensive press, analyst and blog coverage (like this!) plus books like Nicholas Carr's book The Big Switch. It definitely feels like we are in the Hype cycle.

Will traditional corporate IT really move dramatically to the cloud as Carr predicts?

I think a lot of the answer comes down to control and customization .... or lack of it. SaaS Cloud offerings can work great if the application offered aligns with the business requirement without significant customization or integration. If IT is comfortable without the hands-on control they may even manage and support the project because it's good example of doing more with less. However in most cases there are too many custom requirements and perhaps regulatory controls that make premise based solutions the norm and leave SaaS offerings relegated to specific point solutions.

As business requirements continue to grow, premised-based solutions grow accordingly and sometimes disproportionately when centralizing into limited data centers. A resulting problem that most IT shops now face is a lack of space and/or power. Once the local premise is at capacity a typical next step is to move some or even all the equipment to a dedicated Co-location facility. For many companies this is more of an incremental step because they have already centralized their Application Delivery Infrastructure and may already be using Co-Lo's for back up and Disaster Recovery operations. The definition of "premise" gets blurry, but IT remains in control. The facility is likely owned and operated by a service provider, but everything inside the " cage " is owned and operated remotely by corporate IT.

For companies that have taken this step of moving some or all of their infrastructure to the "Cage" the transition seems to have worked well. Co-Lo Data Centers typically have great bandwidth, power, and room to grow. Plus the high cost of specialized facilities are spread across many Co-Lo tenants often lowering the cost compared to providing company managed dedicated Data Centers. IT is remotely managing the infrastructure with perhaps some on site assistance or occasional trips to the Co-Lo. As long as the App Delivery Infrastructure is robust users don't know or care were the servers are, management is happy because it reduces Real Estate dependencies and facility upgrades like added power, AC and generators.

So, if IT is satisfied remotely managing their infrastructure ( except the trips to the cage... ) new applications get implemented, security is enforced , new users provisioned, etc, do they really care about the hardware the server images and storage is running on as long as it stays up with predictable performance ? I suspect many IT Pros would give up the responsibility, time and burden of acquiring, deploying, cabling, and configuring servers and storage (and trips to the Cage...). As long as they could still architect the solution they need, deploy server images, reboot images, network servers together and generally administer the functions they need to deliver apps and store data, there is a good chance they don't view the underlying hardware as something they really need to own and manage, just like the datacenter real estate.

Perhaps the movement to the Cloud for companies that have already moved to the Cage will not be such a disruptive event. As long as there are Cloud providers that can provide the underlying infrastructure including servers, storage and networking with SLAs, Security, and the appropriate level of administrative control. And if IT can still customize and "control" at the level they need to, they will likely be compelled to make the next transition to the Cloud.

The benefits of Virtualization for rapid deployment and flexibility plus the intrinsic cost benefits of muti-tenancy will win-over converts that are faced with demanding time to deployment challenges and capital cost constraints ( isn't that most of IT ? ). These Cloud service provider offerings will need to overcome some of the SLA challenges that have faced the early adopters like Web 2.0 startups on EC2, plus address security, sustainability and redundancy fears, but these are all curable issues. This will take some time and there are likely scenarios of hybrid models of premise plus Cloud solutions, but the writing is on the wall...the Cage is just a step to the Cloud.