Many say the choice of using compute and storage cloud services is a choice of capex versus opex. Some see the choice as no different that choosing to buy a car with cash up front or to finance that purchase over 48 months. The fact is, nothing could be further from the truth.
The cost of compute and storage is not the cost of the box – it’s the cost of managing the compute and storage. It’s the cost of managing the security, availability, performance and change of the compute and storage. A rule of thumb for estimating the annual cost of managing compute and storage is to multiply the purchase price of the compute and storage box by four. This means in 48-months you’ll spend 16x the purchase price of the hardware to manage it. So the hardware could be free and it would make little difference to the true cost of computing. It’s not about capex versus opex.
So, why does it cost so much? Actually, it doesn’t have to cost so much. During the 48 months, apply no security patches, no updates to any of the operating system software or storage management software and just keep the computers plugged in. Furthermore, if you don’t want expensive compute and storage don’t spend any money on staff to manage the performance or availability and of course don’t purchase any performance or availability management software. In reality, though, whether you’re using VMware, OpenStack, Postgres, Oracle, Cisco or your favourite SDN software there will always be a constant barrage of patches. Some of those might be security patches. In one particular enterprise application we studied there were 137 security patches released from just four suppliers in a 12-month period. If you’re the CIO, CISO or Director of IT, pick one of your enterprise applications and see how many patches your suppliers issued last year. You might be surprised.
Remember, in traditional IT shops you’ve employed people to hopefully study each of these patches, decide which ones to apply, apply the patches, test the application, roll it into production with a plan to rollback the change should it not work in the production system. The source of the 4x costs is people cost, and while we can move management to low cost labour markets, with no standardisation and no automation there will be a floor in both cost and reliability, as the single largest source of systems failure is human error.
So what’s the answer to reducing cost, true cost, and improving both security, performance and availability? Compute and storage cloud computing. Cloud computing is not about the location or ownership of the data centre, or even whose control a computer is under, rather it’s about the standardisation and automation of security, performance, availability and change management of the compute and storage. A high quality cloud service provider will standardise these processes within a standard environment and as a result deliver you not only a significantly lower cost compute & storage infrastructure, but also a better one.
For more information, check out book 1 of Timothy Chou’s cloud computing trilogy.
What opportunities do you see in changes to cloud computing? Let us know in the comments below.