It wasn’t all that long ago that we were beefing up our on-premises server rooms with air conditioning units and uninterruptible power supplies to improve business continuity. Invariably, internet connections would go down, redundant circuits failed to take over, or when IT was still relegated to the basement, the server room got flooded. It became clear on-prem was not the best way to go.
Collocation was a saving grace. Someone else taking care of all the underlying infrastructure was a dream come true. But it was still expensive owning all your own kit and we were still susceptible to hardware failures and subject to painful hardware refresh cycles.
Next, we had the private cloud buzz. Then public cloud. Then hybrid cloud. Now the latest craze: multi-cloud.
For some of us, this latest trend has been against our will: the result of shadow IT from some pioneering line of the business owner, running off their own personal credit. But for others, selecting a multi-cloud direction has been a more deliberate choice to reduce business vulnerability and operational risk.
Regardless of which camp you fall into, here are 5 areas to consider as you implement a multi-cloud strategy to make sure you get the best outcome for your business:
1. Price & Performance
Choose the cloud platform that will deliver the best performance for your workload. But don’t forget about your workforce e.g. choosing Google Apps over Microsoft Office365 for improved collaboration could mean lost productivity while your users come to terms with a different mindset for working.
2. Software Development Life Cycle (SDLC)
Perhaps you’ve selected to run your production workloads in AWS for performance, but that doesn’t mean you have to run all your platforms there. Maybe keep UAT in AWS too (thereby keeping UAT like-for-like with production), but you should still consider looking at other cloud options (like Azure Dev and Test Labs) for running your development and system integration test environments.
Whatever your SDLC framework might be, there may be cost benefits, reduced speed to provision environments, or out-of-the-box dynamic data masking to improve your security posture, but more importantly, it encourages your team to develop deployment practices that are not platform specific so if you want to change later, you can.
3. Legacy Applications
Whether legacy apps spawn from a lack of investment over time or have been acquired through acquisition, sometimes it’s just not worth the effort (or even possible) to migrate systems to a different cloud platform. Opting for a multi-cloud strategy gives you the flexibility of leaving them where they are.
4. Backups & Disaster Recovery
It’s unlikely one of the public cloud big boys will go bust, but it’s not unheard of for them to experience downtime, outages or lose data. Give yourself a fall-back position by splitting your backups or DR environments across two or more different public cloud provider. E.g. If you are leveraging Office365, you could use SkyKick to back up your email to AWS.
5. Vendor Lock-in
No one likes to be held to ransom, so follow the adage “don’t put all your eggs in one basket” just out of principle. Split your workloads so that you aren’t always going to be tied to one cloud provider. It doesn’t necessarily have to cost you much in the short term, but it could save you a packet (pun intended) down the road.
Your biggest challenge is likely to be ensuring your organisation has enough knowledge about all the various dedicated, public, private or hybrid cloud offerings to design the best multi-cloud strategy for your business. If not get in touch today to discuss how Starboard IT can help.