- Pros: It will educate you about utility computing reasons better than anything else
- Cons: It foretells the end of your storage world
- Buying advice: The best book you can find about the possible move to utility computing.
It's that stark. Carr's pitch is that businesses, and consumers, will increasingly switch over from present day DIY computing to utility-supplied computing, with information processing and storage functions carried out in 'the cloud.' Most of us are familiar with the concept. Google searches, Hotmail and Salesforce.com are all cloud computing applications. Amazon's S3 and Google's office-like web-based functionality are too.
Sun has very recently decided to move away entirely from in-house data centres to utility-supplied business computing functions and intends, simply, to have no business data centres by 2015.
Carr's book is persuasive, well-researched, authoritative and convincing. He's reasonable in his conclusions and moderate in his extrapolations. This is an exceedingly good book.
The movement Carr describes is one where businesses and consumers increasingly make a decision based on where a computing application is carried out on the basis of cost, trust and convenience, and choose the web-supplied, grid-like, cloud computing alternative.
Consumers may do it more on convenience grounds than business, which will be more hard-nosed about costs, and both will be concerned about trust. That's trust in the network being there when we want it and trust in the remote supplier being reliable and safe. Familiarity will breed content.
If Carr is right, and he paints a very reliable picture, then we can make some assumptions.
First, every time a business decides to cease carrying out an application in its own (in-house or out-sourced) datacentre or on its staff's PCs, then that means a storage resource is not bought by that business.
Secondly, as these decisions mount up then the overall business storage market will increase at a slower rate and then turn down. The cloud computing storage market will grow very, very quickly indeed. The bigger cloud computing concerns, like Google, will roll their own storage and have a heavy JBOD orientation.
The utility computing suppliers will be able to offer services at a greater and greater discount to the cost of businesses having their own equivalent computing function because they will be able to buy kit more cheaply and run it at higher utilisation levels. This will increase the rate at which businesses switch over to utility computing.
It's likely that SME businesses will move over faster than enterprises because they can save money and because thy are more nimble. Enterprises have a huge data centre infrastructure establishment and will find it hard to abandon it. Carr thinks though, that they will eventually do that.
If the storage market experiences a downturn then suppliers will consolidate. What has happened in the disk market will happen in the SAN array, the clustered file and the NAS array market. Storage suppliers will find their markets are shrinking in terms of customer numbers too.
They will focus on the large customers, both traditional enterprise, which will be declining, and new cloud computing suppliers, which will be growing rapidly.
Storage product development will start being skewed towards the needs of cloud computing suppliers. The storage bought by cloud computing suppliers will surpass that bought by business in general. Business IT departments, faced with the same scenario for both servers and storage, will start downsizing.
Utility computing becomes even more attractive and the general business storage market fades away. You, unless you are concerned with storage in connection with utility storage suppliers or some enduring enterprise storage niche, are out of a job.