While providers and analysts define it differently, multicloud computing generally refers to the use of multiple providers for public cloud infrastructure, such as pairing Google Compute Engine and Amazon Web Services or pairing Microsoft and IBM.
There are different reasons to use a multicloud model, although most relate to platform requirements. For example, organizations might meet their big data needs with Google, but prefer the compute and storage options on Amazon Web Services (AWS).
The latest multicloud discussions focus on active/active cloud redundancy. This means organizations place their primary systems on one cloud platform, such as AWS, and their backup or secondary systems on another cloud platform, such as Microsoft Azure. IT teams then continuously update those redundant systems with the same data on both cloud instances. If the primary provider experiences an outage, workloads can continue to run with the backup cloud provider.
There are three key steps to take when forming a cloud redundancy plan that uses more than one provider.
First, understand your application and data requirements. Remember that the application and data need to live on both cloud platforms at the same time. They should run the same way and have the same operations plans, but those plans should be specific to each cloud platform. For example, an organization might have an inventory control system it deems mission-critical. It could run one instance of the application and database on AWS as the primary cloud platform and another instance on Azure as the secondary platform.
While each cloud and application instance functions identically, there are different operational procedures to consider. For instance, the security mechanisms between each cloud would differ, as would usage-based accounting and other factors. IT teams must be aware of the different procedures and be trained on the differences between each cloud platform; this includes ensuring that the data is in sync.
When building a cloud redundancy strategy, ensure data won't be lost during a switch from one cloud to another.
Second, test each application and data instance on each cloud platform. Test for performance, data accuracy and usability. It's acceptable for the secondary or backup cloud platform to fall a bit short in performance since it's intended for temporary use, but data accuracy must be on the money.
When building a cloud redundancy strategy, ensure data won't be lost during a switch from one cloud to another. To do this, think through your data sync operations. Each application and data instance that runs on both a primary and secondary cloud has to maintain synchronized data in case the secondary application is needed. IT teams should mirror updates to the database between cloud platforms with as little latency as possible to ensure data is not lost during the transfer between clouds.
Backup and archive with cloud storage applications
Discover how cloud storage applications and object storage can benefit your data center in this video Tech Talk from analyst George Crump.
Finally, build a strategy to automatically switch from your primary cloud to your backup when the primary cloud becomes unavailable. There are several ways to do this. For example, you can redirect a network connection to switch users from one cloud to another. Establish policies that reconnect to the secondary server when the primary one is unavailable. IT teams can also use software independent from the cloud platforms to broker cloud applications and automatically switch to the secondary cloud in the event of an outage.
Set policies in your cloud redundancy plan that define when you make that switch; outages on your primary cloud that last 20 minutes, for example, should not qualify for a switch, but an outage that lasts an hour or more should.
The likelihood that your primary cloud host will go down and force you to go onto a secondary, redundant system is low. Most providers furnish cloud redundancy into their systems and would not recommend that you use another cloud as backup.
However, nobody is perfect, and we could see a major cloud outage in the next few years. The outage could be more than a day or two and, in that case, a cloud redundancy plan using multiple providers would be critical. The cost of this plan is also low; you only pay a small fee to host your active code and data on the secondary cloud. Then, during an outage, your bill increases as user load switches to the backup system.
