The Private Cloud Strikes Back
Having read JP Rangaswami’s argument against private clouds (and the obvious promoting of his version of cloud) I have only to say that he’s looking for oranges in an apple tree. His entire premise is based on the idea that enterprises are wholly concerned with cost and sharing risk when that can’t be farther from the truth. Yes, cost is indeed a factor as is sharing risk but a bigger and more important factor facing the enterprise today is agility and flexibility…something that monolithic leviathan-like enterprise IT systems of today definitely are not. He then jumps from cost to social enterprise as if there is a causal relationship there when, in fact, they are two separate discussions. I don’t doubt that if you are a consumer (not just customer) facing organization, it’s best to get on that social enterprise bandwagon but if your main concern is how to better equip and provide the environment and tools necessary to innovate within your organization, the whole social thing is a red herring for selling you things that you don’t need.
Traditional status quo within IT is deeply encumbered by mostly manual processes—optimized for people carrying out commodity IT tasks such as provisioning servers and OSes—that cannot be optimized any further, therefore a different, much better way had to be found. That way is the private cloud which takes those commodity IT tasks and elevates them to automated and orchestrated, well defined workflows and then utilizes a policy-driven system to carry them out. Whether these workflows are initiated by a human or as a result of a specific set of monitored criteria, the system dynamically creates and recreates itself based on actual business and performance need—something that is almost impossible to translate into the public cloud scenario.
Not that public cloud cannot be leveraged where appropriate, but the enterprise’s requirement is much more granular and specific than any public cloud can or should allow…simply to JP’s point that they must share the risk among many players and that risk is generic by definition within the public cloud. Once you start creating one-off specific environments, the commonality is lost and it loses the cost benefits because now you are simply utilizing a private cloud whose assets are owned by someone else…sound like co-lo?
Finally, I wouldn’t expect someone whose main revenue source is based on the idea that a public cloud is better than a private cloud to say anything different than what JP has said, but I did expect some semblance of clarity as to where his loyalties lie…and it looks like it’s not with the best interests of the enterprise customer.