The Multi-Tool for Multi-Cloud

The initial days of “cloud” introduced automation and user interfaces previously unseen in traditional IT settings. Compute resources could be provisioned rapidly and through an interface accessible to both technical and nontechnical consumers. This provisioning could be initiated by anyone, anywhere, and at any time. 

An underappreciated reason for the rapid speed at which cloud initially spread is that it was not encumbered by organizational change, risk, or security processes. Not only did the cloud ecosystem allow for more rapid provisioning, it did so in a world apart from corporate controls. 

These nontechnical advantages were like oxygen to a fire and made “as a Service” cloud a true disruptor. Many organizations responded in an attempt to regain control, but all were facing new customer expectations of a faster, friendlier experience.  

Opportunities for consulting and managed service providers quickly followed. Few organizations needed assistance managing cloud resources and even fewer needed help procuring them. However, connectivity back into the legacy IT environment required technical skills that few possessed. In fact, many organizations were surprised at the amount of training required to equip their IT staff to progress beyond their initial use cases.

What this IT staff was largely trained on was the tool provided directly from the cloud service provider itself.

It was in this environment that the first cross-platform tools were born. Initial efforts were to address single gaps in vendor tools, but most grew into what is now known as a Cloud Management Platform (CMP). These solutions are actually multiple tools bundled together to solve common problems. They typically contain a provisioning interface, hypervisor interface, automation, and reporting. 

Initially, these four components came from four different solutions. Few organizations operated at a scale large enough to justify the challenge of integrating these tools. The complexity of such an undertaking was monumental and often involved building middleware to tie everything together. 

As the market grew, consolidation took place. Single tools began to provide more and more of the capabilities required by customers. Eventually, single vendors could provide a full CMP sold as one tool for consumption by customers of any size.

Market solutions are now robust enough for an organization to turn a CMP inward and provide the same friendly user interface and rapid resource provisioning within their own data centers. They can then leverage the CMP to procure public cloud services and build a true hybrid cloud.

However, old challenges remain around training. 

The hurdle of training technical staff to manage each individual cloud service does not disappear simply because a CMP can automate and streamline much of the work. This is especially true for the automation solutions included in these toolsets. The time needed to create, customize, and maintain templates, runbooks, and scripts is often underestimated. 

Interoperability between multiple cloud services is also a lagging area. What a CMP supports for one provider may not be supported for another. Additionally, migrating workloads between public cloud environments is still a complex task.

Finally, managing accurate pricing and usage data can be a major challenge as importing and managing data from multiple sources can be difficult to automate accurately. It is these overlooked tasks that can make it difficult to realize the time savings promised by a CMP.

Without question, a CMP can allow an organization to move more quickly and provide a more consumer-like end-user experience. However, it is a complex toolset that demands skills missing in most IT organizations. Even when those skills are present, the effort required to implement a CMP is easy to underestimate.