I've been listening to a lot of LastWeekInAWS podcasts and on Twitter. One of the regular conversations that happens in both is Multi-Cloud, meaning should one use more than one cloud provider for the purposes of making your service more available and/or reducing vendor lock-in. The answer here is almost always 'Dear lord, NO!'.
But, multi-cloud does make sense with some limited use cases. Let's take a look.
Use Case #1: Serving Government Users
AWS, Azure, and GCP all offer some type of segmented services specifically for government workloads. If your company's application portfolio includes serving governments and businesses, then deploying to multiple clouds is almost a requirement.
Services deployed to the government cloud are usually separate release schedules than the worldwide GA releases. Updates to data engines, connecting services, and security features all require a more indepth review in order to be certified for use. Therefore, deciding on the best cloud fit for public cloud customers might not be the same as the cloud for government.
One could make the case that a B2G application should be considered its own line of business, with its own code base (or at least feature branch), but that's a topic for some organizational effectiveness teams to figure out.
Use Case #2: Extremely High Peak usage
If you are concerned that your peak workload will exceed available demand for an Availability Zone or Region, then maybe....maybe....sharing that load to another cloud provider might work.
Think BestBuy, Wal-Mart, and Target fighting for compute, storage, and network resources on Black Friday and Cyber Monday.
Use Case #3: Data Locality Restrictions
I thought about grouping this with Government users, but it's not quite the same. If you are operating in a space where there are regulatory concerns about where the data is stored or with specific data lifecycle requirements, then a multi-cloud strategy may be more effective. This is if, and only if, the local cloud requirements offer better features at the same or less cost. At this point, it's not likely, but it's not an impossibility.
That's it. That's the three use cases. If you think that your technology solution falls into one of these use cases, then congratulations. You've got an interesting and (probably) expensive problem to solve. Hopefully someone is willing to pay handsomely for your services.
Comments
Post a Comment