The Data Center: Head in the Clouds - Systems Management News On The Web

Sort By :DateRelevance

DEPARTMENTS

HOME

TOP STORIES

ON THE WEB

BZ MEDIA

Printable version

The Data Center: Head in the Clouds
By John Rath

May 15, 2008 —

When I decided to write about cloud computing, the research took longer to read than it took to write the column! Cloud computing is everywhere in the news and the concept marks an important evolutionary step to the next wave of computing.

Shifts like this in computing do not happen overnight. You won’t see major corporations moving their entire infrastructure to a cloud or cloud provider any time soon. In my opinion, cloud computing is essentially utility computing, delivered in multiple data centers across different regions of the world. It can be a service that a provider with a large data center portfolio can offer, or it can be a model used by a corporation to provision their infrastructure in a more efficient and effective way. Utility computing is “infrastructure as a service,” or computing resources that are consumed in a pay-as-you-go model.

The environmental revolution under way in the data center will be a catalyst for clouds as well, by pushing infrastructure out of smaller, individual data centers to larger, more efficient facilities owned by cloud providers.

However, much of the resistance to such a model is because of the issues of control and regulation, and the idea that you need to be able to touch and see the IT equipment that is being paid for.

The architecture of the cloud can take many forms and be beneficial to a variety of needs and sizes of companies. To the small startup company, it means they don’t have to invest in infrastructure or data center services to get their Web site or Internet application going. They simply buy compute power as a service from a cloud provider that lets them grow dynamically, only pay for what they use, and handle peaks and valleys of interest from their audience with ease.

Amazon’s EC2 (Elastic Compute Cloud) has enabled many such companies to deliver such sites and applications at a fraction of the cost of using traditional methods. My original assumption of EC2 (or any cloud service) was that it meant the infrastructure was spread across the different data centers they owned. It was only recently that Amazon announced the ability to place instances in multiple locations and have elastic IP addresses.

Software has emerged to enable applications to be ported to the cloud and allow companies to design and manage their utility computing infrastructure. Such cloud enablers include companies like 3Tera and Elastra.

Large corporations will be slow to adopt overall but will experiment with projects, as seen by the New York Times, which needed to convert old articles from scanned images to Acrobat files. They used Amazon EC2 to churn through 11 million articles in just less than 24 hours, using 100 EC2 instances and 1.5 Terabytes of storage on the Amazon S3 storage service.

Hybrid examples typically point to event- or sport-based Web sites that need extra capacity at certain peak times of the year. They will use their own equipment to service their regular audience, but turn to a provider to supplement capacity by renting infrastructure on an as needed basis.

To the IT manager, it is an attractive proposition: use 100 servers and pay nothing more than 24 hours worth of compute time, bandwidth and storage it takes to complete the job at hand. The alternative is trying to forecast the infrastructure needs and paying not only for the hardware, but all software, services, engineering labor and other hidden costs. Imagine doing capacity management by trending actual use instead of trying to forecast and provide infrastructure to meet demand.

An alternative definition for cloud computing can be thought of as infrastructure economies of scale. It is a service that, with the help of massive data centers, can be produced on a larger scale with less input costs. Along these same lines, there are opportunities for vertical industries such as health care, where a cloud solution can be designed and managed for particular needs. Generic cloud service providers don’t care if you are a Web startup or the New York Times; the service is the same. Vertical clouds could address particular needs like security and privacy to cater to a particular set of requirements.

Another area of cloud computing that I can see emerging is data and payload portability. Avoiding vendor lock-in is always a factor when evaluating services, and customer demand will drive standards and methods for moving from one cloud to another. Companies that provide cloud design, deployment and management could evolve to offer the ability to move from one cloud provider to another.

A dictionary definition of the word elastic is “spontaneously expansive.” The definition of the word utility is “the capacity of a commodity or a service to satisfy some human want.” I think of a cloud as an elastic utility computing service, delivered over multiple data centers that are in different geographical regions of the world.

John Rath is an independent consultant and blogger at Datacenterlinks.com. He can be reached at johnsr4@gmail.com.

Related Search Term(s): Cloud computing

Share this link: http://www.sysmannews.com/link/32187

EMAIL THIS ARTICLE

SEND FEEDBACK

MORE COLUMNS