here is the full text from my Blog post shared as part of the Work-Bench 2019 predictions article.
2019 will be the year that Enterprises realize that fears of cloud lock-in are overblown and miss the existential threat faced by cloud agnostic or mulit-cloud adoption. While it is understandable to have concerns about multiple suppliers for all critical services, a time honored strategy that lasted decades, applying this to multi-cloud results in competition at the lowest common denominator level, infrastructure as a service (IaaS). So multi-cloud has let us hedge our risks of our provider changing terms or raising prices, but large amounts of execution risk remain on the Enterprise’s engineering and operations teams. This multi-cloud strategy will only please the CFO and even then only temporarily, due to the increased operation complexity and risk.
Moving up the value chain does introduce a “preferred cloud provider” which can lead to lock-in, but is lock-in your biggest threat? Most large Enterprises are at greater risk from the operational burdens of legacy software at scale, running highly-available replicated databases and migrating clients/users to modern platforms. The skills to operate these systems in-house have been eroding with the aging population. As those skilled resources become scarce, they will tend to congregate where the money and opportunities are greatest, the major cloud providers.
So how can you handle those hard things in the public cloud without the deep engineering talent? Move up into the “Lock-in land” of Aurora and other Managed services on AWS. Many of these managed services are already multi-region, fault tolerant and “infinitely” scalable. Enterprises that risk “lock-in” by adopting these managed services will discover they have bypassed huge engineering tasks to become cloud scale with a minimum investment. Today’s tooling and cloud vendors enabling adoption modern techniques with minimal costs, makes lock-in a data migration and reimplementation risk to be managed not feared. Recognizing that the pace of change is accelerating and that failing to adopt cloud native patterns including managed services will be fatal. Enterprises will finally learn that software projects are never “done” and that re-implementing the valuable business logic becomes simpler with every new language, platform or service.
Trying to treat all cloud providers as the same in a multi-cloud plan is the same a treating every airplane or automobile as the same. A fundamental error failing to recognize that using the right tool for the job is the difference between driving a Model T or a Tesla. Do what you fear!