The AWS S3 system is a public Cloud hosting platform. It hosts websites, Internet accessible applications and sharing sites for data and images. Because of the outage, you may not have been able to reach a website, or the site may not have been displaying images. Nest, the popular smart thermostat, smoke/carbon monoxide detector and webcam was impacted. If you have one or more of these devices, you would have been unable to manage them, get alerts, etc.
What’s worse is that Amazon’s status page, where they report on the real-time health of the AWS services was also broken. This meant that companies that have systems hosted on S3 with AWS were unable to see the status of the system and more importantly, updates on Amazon’s progress resolving the issue. That last point being the most important. Once you know you have a problem, the most important thing you can do is keep your customers updated on the status and estimated time of restoration of services.
What this clearly points out is the weakness of the public Cloud. If you put all of your eggs in one basket and only run your systems in one public Cloud, an outage like this will cripple your business, at least for the time of the outage. The impact to your brand and success of your product and/or service could be immeasurable.
The only way to safeguard against this risk is with multi-provider redundancy. AWS and Microsoft’s Azure have established themselves as the market leaders. What if you could replicate your systems between both platforms, so that one could be your active, or primary, service and the other could be your passive, or secondary, service? If you build out network redundancy within just one provider, if that provider has an outage, both your primary and secondary systems could end up offline. If you ran each on a different service platform, you should be able to stay up and running through an outage at just a single provider.
With Internet enabled apps and services becoming more the norm than the exception, building out this type of redundancy becomes critical, quickly. The industry recognizes this and has coined the phrase Multi-Cloud Management. Basically, this means being able to configure, control and maintain mirrored infrastructures between two Cloud providers.
This concept is not dissimilar from redundancy within traditional on-premise infrastructures. In this example, companies would establish a primary server closet or data center and a redundant one elsewhere in the building or nearby. The point here is simply that the company controls the infrastructure and therefore the redundancy.
Multi-Cloud Management extends this to the public Cloud. As outlined above, this concept allows companies to establish their Cloud based infrastructures with market leaders like AWS or Azure and then properly maintain it. For those that find on-premise redundancy too expensive, this may provide a highly attractive alternative.
Of course, this all depends on having the right applications and/or services that are able to be moved to the public Cloud. If they are, how will you manage them in an efficient manner? All considerations for which you must develop a strategy.
The Cloud holds tremendous opportunity and if you were impacted by this week’s AWS outage, you understand how important redundancy is, both on-premise and in the Cloud.