How complicated is Amazon Web Service’s cloud? So complicated, it had to devote a full conference track to managing cloud resources at last week’s re:Invent conference, including – kid you not – a complete session just to explain AWS billing. This is just one example that for all the talk about the awesomeness of cloud, its complexity is a very real impediment to adoption by business customers.
You’re Doing It Wrong
The billing session came as a bit of a shock. I did not actually attend re:Invent, AWS’s first conference, but I happened to be at the same venue in Las Vegas for next event, the CloudStack Collaboration Conference. Somehow or other, over a breakfast meeting with a bunch of rather tired and post-chemical attendees, the topic of AWS came up. It was only natural, given that some of them had been in Vegas the whole week and had attended re:Invent, too.
Reaction at the table to the conference track, entitled “Resource Management,” was incredulous and laced with invective. One Apache OpenCloud committer who did not wish to be identified summed it up best: “If you have to have a session on billing, you’re doing it wrong.”
It’s a valid argument, because while one should rightfully expect all levels of interest to be addressed at a trade show’s first run, it seems that something like billing for cloud services should be pretty simple and not in need of tutoring. This kind of complexity raises serious questions about the real value of using a public vs. private cloud in the enterprise.
Not every session in the track was like this, of course. “Introduction to AWS CloudFormation,” “Introduction to AWS Elastic Beanstalk” and “Cloud Infrastructure and Application Monitoring with Amazon CloudWatch” appeared to be good technical introductions to some of the 21 tools used within the central AWS Management Console.
Along with the “Decoding Your AWS Bill,” the “Optimizing Costs with AWS” session also seemed to focus on figuring out your bill – and driving it as low as possible. Which is a good idea, of course, but why should it be so damn complicated?
Perhaps I should have gotten my first clue earlier this year when I covered the new Xervmon service. Xervmon’s singular mission is to analyze your cloud computing and data center use and help you find ways to keep costs in line. The fact that such a start-up even exists should have served as notice enough of the issues enterprises are having with cloud computing.
How Complex Does Cloud Need To Be?
The problem is, as complex as cloud seems to be, there doesn’t seem to be much progress toware making it more manageable. Cloud computing is often compared with Linux, and indeed there are parallels in how it is being adopted. But even though early iterations of Linux were pretty complicated to install and use, there was always real progress towards ease of use. With cloud, that movement seems glacial.
That may be either a good thing or a bad thing, depending on your point of view. There’s the argument that cloud computing is doing very complex things and so therefore should never be too simple to implement – otherwise it would be too easy to screw something up.
But IT managers are generally not the type of people who like to feel stupid, and sooner or later cloud computing is going to have to get to the point where those managers can operate these platforms – be they public, private or hybrid – and, more importantly, explain the value and cost benefits to their bosses.
Otherwise, cloud adoption will remain as ephemeral as its namesake.
Image courtesy of Shutterstock.