The Great Cloud Disruption of 2011 has been called many things. For some it’s been a wake-up call and a warning for leery IT managers and prospective purchasers of cloud services. For others, and it seems more opinions favour this idea, it has been a shining moment, demonstrating the resiliency and efficacy of well-built cloud infrastructure. Now that we have some distance from it, and have had a chance to review the full Amazon post-mortem and apology, let’s look at few things business owners, managers, and decision makers should be thinking about as cloud computing continues to emerge as a driver of tech growth.
It’s all about change
The cloud is changing the way technology management works, including the way software developers and engineers build and deploy their apps. Everything from enterprise back-office systems, to mobile and social gaming applications need to be cloud ready to take advantage of the scale and resiliency that cloud technology offers, while being prepared for (and embracing) failure of any given component at any given time.
I’ve found that developers and managers from across the spectrum of Fortune 500 companies to startups are not equipped with basic change management systems and software that enables cloud systems to scale, heal, and repair themselves. Developers in particular are resistant to adapt their style or systems to “infrastructure”, however, in this case it is time well spent. Make sure your company has a solid strategy for software configuration management (SCM), and if you don’t, take a look at open-source tools like Git, originally designed for the not-so-small task of managing the development of the Linux operating system that powers much of the web today.
Scale or fail
Once you have a way to manage your application’s code and deploy effectively to a cloud environment, you can get it ready to scale. If you’ve designed your infrastructure correctly, you should be able to auto-scale on demand, or based on some sort of predetermined business or performance event. (Hint: if your cloud consultant tells you the system they built is ready to scale, ask them to prove it you). Is that new viral marketing campaign actually going viral? Cloud scaling gives you a way to keep up with demand.
But scaling isn’t just about being ready when the next Internet Pet Rock goes live. It’s also about being able to recover from failure. Failure can happen at any level – by your software or configuration, or the underlying cloud provider. Make sure that you have the engineering in place to take advantage of whatever data centre separation your provider has (in Amazon ‘s case, you have the choice of many global regions, and within those regions, many “availability zones”).
With a cloud-ready change management process and a design for thoughtful, engineered scaling, your investment in cloud technology should be as disruption proof as possible. Want even more control and more certainty? Start looking for global high-availability, provider diversity, and hybrid cloud options. But if you start with some fundamentals, you can stay one step ahead of Murphy’s Law.