Some of the traditional assumptions are no longer true today or won’t be true soon. The changes are for the worse – they raise prices and introduce significant additional complexity for companies that operate in a hybrid and multi-cloud environment.
For years, two key assumptions dominated public cloud operations: the price consistently declined, and work was portable from one cloud to another location. Both of these operational trends are now reversing.
What’s Happening to Cause Cloud Changes?
There are three hyperscale cloud providers: AWS, Google Cloud Platform and Microsoft’s Azure. The three providers invest in their cloud services in ways that ensure customers can't shift their workload to other cloud providers.
Cloud hyperscale providers are creating constructs that lock in enterprises and discourage multi-cloud environments.
They do this by partnering with system integrators to push their native platform offerings. Going forward, implementation partners, strategy consulting firms, etc. will take preferred partnership status to push offerings of a cloud vendor.
These offerings make it difficult to migrate work to different clouds through a wide variety of mechanisms ranging from long-term contracts to interdependent functionality with tight integration to management tools and to attractive complementary capabilities.
Microsoft, for example, found a way to hurt the partnership between AWS and VMware by raising prices for customers using non-Microsoft clouds. They increased the licensing cost for running Microsoft workloads in competitors’ cloud platforms.
Not all moves create friction with interoperability. For instance, Google is the third major cloud vendor to tie the knot with VMware.
Following AWS’ partnership with VMware, both Microsoft and Google Cloud Platform mimicked the move by partnering with VMware.
However, for every highly publicized move to appear open and interoperable, there are three designed to constrain interoperability.
Another strategy of hyperscale cloud providers aimed at locking enterprise customers into their platform is to create important data management services, testing services and AI services.
With these services, the hyperscale solution is harder to operate across multiple clouds and increases the practical challenges of moving work from that cloud provider’s platform.
In addition, the three hyperscale cloud providers build walled gardens that contractually lock companies in. They achieve this lock-in by building alliances with other important cloud service providers, which then become necessary services as part of a customer’s ecosystem.
The providers also tie their software tightly to their hyperscale platforms. For instance, there are software tools that only work with AWS and don't work in Azure. And customers can achieve better integration for advertising tool sets in the Google Cloud Platform than others.
This strategy of creating fortress-like behavior in the cloud – allowing providers to lock in customers to their platforms – is not the only change organizations will see in 2020.
Along with this strategy, the hyperscale providers are increasing the cost of cloud. They achieve this by introducing additional costs for data storage or data transfer, for example. And the increased costs create further lock-in.
Complexities and ComplicationsAs the complexity rises in operating in a multi-cloud environment, private cloud will remain relevant for workloads with strong performance and security requirements.
The alliances that the major cloud players forged with VMware will make it increasingly easier to interoperate with private cloud than across multiple public clouds.
The realities of multiple clouds operating at scale create increased complexity and slow agility, thus eroding key benefits of cloud environments.
Unintended consequences of the complications will affect the following two practices:
- The current widely accepted practice of matching an application to the cloud platform best suited to optimize its performance
- The even more widely accepted practice of allowing developer groups to choose the cloud platform on which they prefer to develop as the target cloud environment.
The changes to these practices may result in an evolving siloed environment. If the unintended consequences become too high, they may start to give way to a more systematic set of standards.
Furthermore, companies must be aware of the talent trap associated with this multi-cloud world. As talent skills form around cloud ecosystems, it becomes increasingly clear that moving talent indiscriminately across platforms comes at a high price.
A further manifestation of the talent challenges is the talent provided by service providers. It is common for service providers to claim that their teams have capabilities across all cloud platforms. But customers should be careful not to fall into the trap of assuming that the individuals that the third-party firm provides have the necessary skills and experience in their assigned areas.
It is most important to examine the skills of senior engineers and architects where deep design knowledge can make all the difference.