It is for a lot of people. There are [rules] of compliance. You have European rules that make sure your data is held in European boundaries and the US puts a lot of liability on anyone who holds health care data. These compliance issues make it very difficult for people just to throw their data out [in the public cloud]... because you are not quite sure where the data is going to end up.
People [also] have concerns about who can actually access that data. A lot of public cloud infrastructure providers... don't have real SLAs (service level agreements) so you are not quite sure what is going to happen.
Security is a big concern. I think some people might be over thinking it, but for a lot of people it is a true concern out there.
Is there anything that isn't suited to being in the cloud?
At this point, yes. There are a lot of things that, again for security and compliance reasons, [aren't].
You also have some applications with requirements, in terms of data, that if your computer and data are too far away, it introduces latency that wouldn't really work. I
If you want to have a certain level of customisation or control over the cloud, it also won't work. For example, you never know when [a provider] is going to do maintenance, you are never guaranteed that your machines are going to run here or there, or on this type of hardware or that type of hardware. If you have specific hardware requirements, it is going to be more difficult to find a suitable public cloud infrastructure for you.
Is the cloud better suited to the small and medium business (SMB) market or to large enterprises?
I think it works for both. I think SMBs are going to look for more packaged public cloud [offerings] because they dont' necessarily have the IT, whereas a larger company can actually say for some sort of applications or lines of business 'I am going to use this kind of computing because its cheaper.'
A good example is Intuit. It is a large provider of personal finance software in the US and provides a product called TurboTax, to [do] your taxes.
Here is a company that has huge data liability issues because you are putting your tax details on its network. But, at the same time, they need to do load testing. Once a year it fires up a lot of Amazon instances to simulate 18 million people hitting its network. It then shuts its network and carries on with its own service which is hosted on its own systems.
To do that load testing, it doesn't really make sense to go and buy equipment just to do one load test.
This is a model where large enterprise customers will use a public cloud for a very specific thing at a very specific time. We do the same, we do a lot of our regression testing on the public cloud.
