Top three myths around the cloud

Advice 2010-09-07 12:21

Cloud technology is frequently plagued by misconceptions and inaccuracies, and as the technology continues to mature, the terminology is continuously changing. Jim Stikeleather, chief innovation officer of Dell Services, addresses three widespread myths associated with cloud computing.

Myth #1: There is ‘The Cloud’
There is no universal cloud. In fact, the term cloud has been hijacked by marketing departments and thereby has caused confusion as to what cloud really means. Describing a service offering as a ‘cloud’ is like describing a 47-year-old female astronaut with a PhD in astrobiology and an MD with a specialty in cardiovascular surgery as a mammal. The amounts of information conveyed by the two descriptions are radically different.

In fact, there is no such thing as ‘cloud’ technology. Very different technologies are used by the likes of Amazon, Google, Salesforce.com, Microsoft, Rackspace, OpSource, Joyent and Dell Services. In the same way a mammal is described as an animal (exhibits locomotion), is warm blooded (endothermic), and bears live young (with exceptions – duck bill platypus), a ‘cloud’ exhibits the following properties regardless of the underlying technology:

• On demand services: There is no need for upfront ‘order taking’, and there is minimal delay between requesting services and acquiring service.

• Metered: Payment (or chargeback) is based upon some form of usage metric (consumption over a time period) that is dynamic (can vary over the time period) and reasonably fine-grained (not a major step function).

• Scalable: The consumption and delivery pattern can grow or shrink dynamically (usually tied to the specific metric of the offering).

Myth #2: Any application can run in ‘The Cloud’
The truth about whether any application can successfully work in the cloud is both ‘yes’ and ‘no’. It depends upon the application and the technologies involved in the specific cloud.

Regardless of the implementing technologies and the marketing communications concerning a specific cloud, there is an underlying goal of the movement toward utility computing, i.e., making IT function similarly to the electric grid – plug into the wall and go. Most applications today cannot take full (or even minimal) advantage of the capabilities of these utility technologies without being re-written. Specifically, the applications have to be written using either a stateless or RESTful (Representational State Transfer) architecture. Only then will applications be dynamically scalable, workload migratable, location independent and inherently fault tolerant.