Amazon suffered another Web Services outage last week, taking down popular websites and throwing the cloud community in a bit of tizzy. Since then, experts have weighed in on what people should learn from these outages, and I’ve aggregated a few of the lessons here. Why? Because we’re increasingly going to rely on cloud services in the future to power business, government, consumer services and more. And if we’re going to create more applications in the cloud, we’d better understand the foundation we’re building on.
- The cloud isn’t magic. Just because a company offloads its infrastructure into the cloud doesn’t mean best practices should go out the window. Many, many commenters in online forums have stressed the importance of distributing web services across multiple geographies. That way if there’s a power outage or other problem in one region, fail-over to another region can help prevent serious disruption.
- Beware the “fog of virtualization.” The phrase comes from Craig Labovitz of DeepField Networks. What it means is that as we move more stuff into the cloud, it’s going to become harder and harder to understand where supply chain dependencies and potential weak points lie. I’m betting that some of the people affected by last week’s event didn’t even realize they were vulnerable to an Amazon outage. That dependency can be disguised by layers of application and/or infrastructure abstraction.
- There’s always a balance between cost and availability. Ensuring day-to-day performance and planning for effective disaster recovery both cost money. That doesn’t change with a shift to the cloud, and companies have to plan accordingly. You can certainly over-invest in availability, but without a sufficient back-up plan, the risks for disaster are high. As cloud platforms start to proliferate, the option to fail over to a secondary cloud service will start to become more viable. That’s a good thing, but it will still cost money. Even in the cloud, there’s no such thing as a free lunch.