The big picture: What’s been happening in the cloud computing world over the last 12-18 months is more than just a simple increase in competitive options. It’s a significant expansion in thinking about how to approach computing in the cloud. With multi-cloud, for example, companies are now embracing, rather than rejecting, the concept of having different types of workloads hosted by different vendors.
Ever since the rise to prominence of cloud computing, we’ve seen businesses grapple with how to best think about and leverage this new means of computing. Some companies, particularly web-focused ones, dove in head-first and now have their entire existence dependent on services like Amazon’s AWS, Microsoft’s Azure, and Google’s Cloud Platform (GCP).
For most traditional businesses, however, the process of moving towards the cloud hasn’t been nearly as clear, nor as easy. Because of large investments in their own physical data centers, thousands of legacy applications, and many other customized software investments that weren’t originally designed with the cloud in mind, the transition to cloud computing has been much slower.
One of the stumbling blocks in moving to the cloud for these traditional vendors is that the shift has often required a monolithic change to an entirely new, distinct type of computing. Needless to say, that’s not easy to do, particularly if the option you’re moving to is seen as a singular choice, with few alternatives. In particular, because AWS was so dominant in the early days of cloud computing, many organizations were afraid of getting locked into this new environment.
As alternative cloud computing offerings from Microsoft, Google, IBM, Oracle, SAP and others started to kick in, however, companies began to see that many different viable alternatives were available.
In a way, we’re seeing cloud computing evolve in a similar path to overall computing trends, but at a much faster pace. The initial AWS offerings, for example, weren’t that conceptually different from mainframe-based efforts, focused around a platform controlled by a single vendor. The combination of new offerings from other vendors as well as different types of supported workloads could be seen as a theoretical equivalent to more heterogenous computing models. The move to containers and microservices across multiple cloud computing providers in some ways mirrors the client-server evolution stage of computing. Finally, the recent development of “serverless” models for cloud computing could be considered roughly analogous to the advancements in edge computing.
In this context, the announcements that IBM made at last week’s Think 2019 conference around their Watson AI services are well timed to meet the evolving cloud computing demands. Specifically, the company said that through their Watson Anywhere initiative they were going to be making Watson AI services available across AWS, Azure, and GCP, in addition to their own IBM Cloud offerings. For situations where companies may want to develop and or run AI-based applications in private clouds or their own data centers, the company is licensing Watson to be able to run locally.
Building on the company’s Cloud Private for Data as a base platform, IBM is offering a choice of Watson APIs or direct access to the Watson Assistant across all the previously mentioned cloud platforms, as well as systems running Red Hat OpenShift or Open Stack across a variety of different environments.
This gives companies the flexibility they are now expecting to access these services across a range of cloud computing offerings. Basically, companies can get the AI computing resources they need, regardless of the type of cloud computing efforts they’ve chosen to make. Whether it’s adding cognitive services capabilities to an existing legacy application that’s been lifted and shifted to the cloud, or architecting an entirely new microservices-based service leveraging cloud-native platforms and protocols, the range of flexibility being offered to companies looking to move more of their efforts to the cloud are growing dramatically.
Vendors who want to address these needs will have to adopt this more flexible type of thinking and adapt or develop services that match not only the reality of the multi-cloud world, but the range of choices that these new alternatives are starting to enable.
The implications of multi-cloud are significantly larger than just having a choice of vendors, or choosing to host certain workloads with one vendor and other workloads with another. Multi-cloud is really enabling companies to think about cloud computing in a more flexible, approachable way. It’s exactly the kind of development the industry needs to take cloud computing into the mainstream.
Bob O’Donnell is the founder and chief analyst of TECHnalysis Research, LLC a technology consulting and market research firm. You can follow him on Twitter @bobodtech. This article was originally published on Tech.pinions.