At this year's AWS re:Invent conference, TheServerSide got an enlightening look at how cloud technology has evolved into its present state of maturity. It's becoming more apparent than ever that businesses need to take a long view when they start thinking about how, why, and when to migrate their application to the cloud. Enterprises have long held back from entrusting more than their public websites to the cloud due to security fears. But today's business decision makers are starting to relax a little about perceived threats. With the proliferation of private and hybrid cloud models, it's become much easier for vendors to allay customer concerns by offering more secure solutions for sensitive enterprise data and apps.
Security requirements are still a critical aspect of the migration conversation, and businesses are far from blasé about staying safe in the cloud. However, a newfound willingness to face and deal with these issues means buy-in is easier to come by these days. According to Peter Roosakos, co-founder of Foghorn Consulting, "We're seeing companies accelerate in their willingness to move workloads to the cloud."
Unfortunately, there are still some significant hurdles to overcome. Legacy technologies are throwing a wrench in the machine, just like they have always done. Some older applications aren't technologically viable in the cloud; others simply won't be supported by today's cloud vendors. This means no app left behind simply isn't realistic as any kind of short term strategy. However, doing a staged migration can provide time to make more apps cloud ready.
Migration isn't supposed to be instantaneous
Making a transition into the cloud will always be a staged process for enterprises. That's not a bug, it's a feature. This approach and the very act of analyzing and classifying apps and workloads to determine what goes 1st, 2nd, and 3rd provides an opportunity to learn important lessons about how an organization is using its resources. If this process doesn't uncover at least a few things the business could be doing better with or without the cloud, those involved probably aren't looking hard enough.
Taking it a little slow also provides an opportunity to eliminate technology risks, maximizing ROI. In the end, deciding what stays in-house and what goes to the cloud is a business decision, not just a technology decision. An enterprise should never lose sight of what its organization has to gain, or lose, in real terms of functionality, flexibility, and profitability with each app migration.
What about reverse migration?
A huge percentage of enterprises are developing an entry strategy with cloud readiness assessments, ROI and cost calculations, and implementation road maps. But few are considering what they might do if the time comes to make like the rain and drop from the cloud. What if a day should come when a business decides the cloud simply isn't working for them?
Peter's take on this is simple. He says not all businesses need an exit strategy. Sure, a government agency might plan for a worst case scenario so bad that the cloud is no longer an option. They've got the resources to do this and might actually face a situation where it would be a reasonable decision. But for most businesses, the likelihood of an event happening that would necessitate migrating out of the cloud and the cost if it did happen are outweighed by the time and expense of actually putting a backup plan in place. It's not always appropriate to prepare for the worst if the money spent on planning and implementation could be put to better use elsewhere, like getting new customers.
Don't migrate out, just shift loyalties
It is reasonable for most businesses to consider what they will do if they need to break ties with a specific cloud vendor. No one wants to be locked in with their entire IT infrastructure, their mission critical apps, or their vital data at the mercy of a single supplier. So, it makes sense to examine what hooks are in place that might anchor a cloud solution too firmly into the guts of the enterprise. Choosing a solution with well-defined extraction layers that make cloud hopping feasible may be a responsible course of action.
At the same time, being too commitment-phobic isn't optimal either. High portability is just a prison of a different sort. As Roosakos says, "When they take that strategy, they have to lock themselves in to the lowest common denominator of features for each of those cloud vendors. They don't really have the opportunity to take advantage of some of those more advanced features." If it's too easy to break up with your cloud vendor, you probably aren't getting the most from the relationship in the first place. You may have to marry the cloud instead of stealing the rain through the fence.
Is an exit strategy as important as a cloud based migration strategy? Let us know.