Multi-cloud management tools fall short of expectations
Multi-cloud presents new management issues for enterprises. Users want better visibility and control of multiple clouds, but vendors struggle to meet these needs.
Companies continue to invest in multi-cloud strategies. This gives them access to a breadth of services, but it also creates management challenges due to platform differences.
Multi-cloud management can simplify deployments across disparate environments and lower maintenance costs. However, the tools currently available on the market remain a work in progress. Their shortcomings force users to take on a great deal of complex systems integration work -- a burden most enterprises are unable to bear.
As a result, many enterprises that deploy more than one cloud often manage those environments autonomously, instead of using a multi-cloud management tool. With this approach, resources are wasted, workloads are unable to move to the preferred platform and users pay more than they should. What they need is a unified offering to ease these troubles.
Enterprises desire more
Organizations want vendors to deliver more comprehensive multi-cloud management services, though the industry hasn't settled on the system requirements or delivery mechanisms for these tools, according to Roy Illsley, an analyst at Ovum, a market research company.
"Corporations are looking for a multi-cloud management nirvana," Illsley said. "No vendor now delivers it. At best, they have 70% or so of the functionality that users want."
Corporations want features that support cross-cloud orchestration and provide visibility into different cloud services' performance. This would give companies the ability to automate certain maintenance tasks, such as dynamically moving a workload from one cloud to another.
Corporations are looking for a multi-cloud management nirvana.
Roy IllsleyAnalyst, Ovum
Multi-cloud users also want to conduct cost and risk analysis to determine the best fit for their workloads, and they want to enact policies across clouds, such as network security, in a consistent, automated manner. For that to occur, the management tools need to recognize and respond to cloud platform system changes -- capabilities that remain a work in progress.
Third-party vendors put in enormous effort to link the various cloud-native management systems. There aren't standards that outline how different management systems exchange information, so each vendor must build connections, one at a time, between systems. This pulls management data from the various platforms into a single console for users, but because of the complexity involved, few management vendors are able to offer much more than that.
To compound the problem, these connections are made via each supplier's management APIs. Due to competitive demand, public cloud vendors concentrate on their own system's management capabilities, rather than open those systems to others. As a result, the current APIs are not as rich as autonomous management systems, and functionality is lost when different systems are connected.
However, there are signs that the major cloud vendors might change in the face of user demand. In April 2019, Google Anthos -- previously known as Google Cloud Services Platform -- was made generally available. It provides a common software tack that runs on Google Cloud and on premises, but it's also designed to manage workloads on third-party clouds like AWS and Microsoft Azure. Still, because the service is so new, it remains to be seen how deeply Anthos will integrate with those other platforms.
Users must fill in the gaps
Vendors continue to make promises about multi-cloud system capabilities, but it has become increasingly difficult for users to believe them. Once the conversation turns from the hypothetical to the real-life scenarios, there's always a catch, Illsley said.
"Unless you are Facebook or Google, you do not have enough manpower to tweak these systems enough so that they function in a complete, true multi-cloud management fashion," Illsley said.
For example, a multi-cloud management system might require a dedicated Python programmer to integrate and maintain the system, but that's not a position most companies have on staff. The orchestration functions that many users desire are only available to enterprises with large tech support staffs and technical experts that can build those capabilities themselves.
