While many healthcare organizations are still determining how best to access the cloud for the first time, others are already confronting the complexities of managing the multiple clouds they’ve engaged for their IT systems.
As tech writer David Linthicum points out, “Since each cloud provider and platform . . . offers different services for storage, compute, security and governance, multi-cloud enables enterprises to take advantage of the best offerings to meet their needs and budget. Unfortunately, managing those different services across multiple cloud platforms can become overwhelming.”
One common way of managing multiple clouds is to use a cloud management platform specifically designed to support multiple Infrastructure-as-a-Service providers, but Linthicum cautions that it’s critical “to understand the upsides and downsides of multicloud management tools.”
For example, beware what he calls “the least common denominator effect.” Some enterprises, he explains, “place an abstraction layer between the consumers of cloud services -- such as application workloads or end users -- and the services themselves, using a generic API for cloud storage and compute. That API can only support a subset of common features from each cloud provider.” If, for efficiency’s sake, an organization uses an API that is common to all the clouds in use, it runs the risk of limiting how much of each platform’s capabilities it can use.
Another limitation, Linthicum says, is “the ability of cloud management and brokerage tools to keep up with the changing nature of public cloud services. For example, some tools might have limited support for new IaaS services, such as those related to serverless computing and containers. If you can't find certain services in the platform's directory, you can't use them.”
Finally, Linthicum points out that “most cloud management platforms provide a layer of abstraction intended to help admins manage multiple clouds through a single interface,” but that also complicates the cloud’s overall architecture, adding “another potential point of failure and another potential attack point for hackers.”