As the environments for enterprise use of multiple clouds expand, the impetus for considering more than one cloud provider is growing. IT professionals need to consider the many aspects of expansion to multiple providers.

The cloud environment continues to mature in a way that encourages enterprises using cloud service providers (CSPs)—or managing their own private cloud—to look beyond their current arrangements in the future. As IBM explains in a simple definition that’s hard to improve on, multicloud management refers to the relatively new field of "running enterprise applications on Platform as a Service (PaaS) or Infrastructure as a Service (IaaS) from multiple CSPs, such as Amazon Web Services (AWS), Google Cloud Platform, IBM Cloud, and Microsoft Azure."

Although multiclouds are like hybrid clouds in that they are often a mixture of public and private clouds, they require greater management of resources to keep data organized, security policies in sync, and applications integrated, as well as to provide uniform governance across a suite of cloud services that have different operating environments.

The advantages of multicloud management can be as varied as the business needs of the enterprises using them, but the most common are reduced risk of supplier lock-in (the difficulties of switching CSPs and the ability to avoid a single point of failure), access to better technologies as the major CSPs develop their offerings, and infrastructure maintenance cost benefits. A specific enterprise will also likely find benefits stemming from its own vision of its business mission as well or may find that certain CSPs better fulfill regulatory standards or security requirements that pertain to the enterprise’s lines of business.

Getting Started with a Switch

Regardless of motivations, a migratory path to multicloud starts with defining solid business objectives that are achievable by making such an environmental change. Likely, an interdepartmental task force will be the best exploratory vehicle. Explore what aspects of multicloud operations will get the enterprise to a better place. Next, survey CSPs to see what technologies and management tools they provide that will move the enterprise toward its goals. Become acquainted with best practices for controlling the system and providing the best security. Understand what the advantages and costs will be of using services from each CSP under consideration. The CSPs themselves will be eager to show their best advantages, but businesses should consider outside consulting help that isn't necessarily beholden to a particular vendor.

One aspect to pay particular attention to is the tools that each major CSP offers because one of them may prove to be an especially good fit for a particular enterprise business environment. For example, AWS CloudFormation is an Infrastructure as Code (IaC) tool that provides a language with which to manage AWS and some third-party resources. Microsoft’s Azure Resource Manager is a template-based management layer that helps control the flow of requests through a multicloud system from application program interfaces (APIs), software development kits (SDKs), and other tools. Understanding how these and other tools from various CSPs work can help match capabilities to business needs.

It's also going to be worthwhile to check out third-party multicloud management tools, of which there are too many to summarize here. However, key capabilities in a cloud management platform should include support for the major CSPs, support for private clouds if necessary, a unified view of all accounts being operated (called a "single pane of glass," or SPOG) via the system, and similar views for managing security, compliance, governance, and costs. One form of this is called a "traffic light governance dashboard," which can summarize all accounts in a system and indicate which ones may currently need special attention.

Architecture Types

Another important aspect of deciding on a multicloud structure is the different types of architectures CSPs present. A variety of approaches is available.

Cloudification is the migration-related process of tailoring existing apps to take advantage of cloud-environment characteristics.

Multicloud relocation, in contrast, is an agile strategy that reallocates apps, data, and workloads among various cloud providers on a dynamic basis.

Multicloud refactoring is an ongoing process of altering or redesigning apps to run seamlessly in any applicable CSP environment.

Multicloud rebinding is like both of these, except that it relies on automation tools to change app-execution CSP environments on the fly based on performance, costs, or changes in regulatory requirements.

Multicloud rebinding with cloud brokerage is the practice of using a third-party service to build app-execution environments. Multi-application modernization is similar to cloudification but applies to reengineering legacy apps that run natively elsewhere to be cloud-friendly.

There are also two categories of deployment patterns to be aware of, each with its own characteristics. The four distributed patterns are tiered hybrid, which spreads tiers of individual apps (e.g., databases, backends) among different clouds; partitioned multicloud, which locates all of an app's functional components across multiple clouds; cloud analytics, which confines analysis tasks to one cloud while preserving the location and processes of the main app; and edge hybrid, which handles processing of edge computing device results with data handlers close to them to reduce latency times.

There are three redundant deployment pattern categories. Hybrid focuses on maintaining performance by balancing cloud and on-premises assets. Business continuity hybrid optimizes deploying critical apps across multiple clouds as a means of maintaining processing continuity. Cloud bursting envisions variable demand for processing at different times and facilitates app expansion into cloud resources normally used only during peak periods.

Categories of Cloud Management Tools

Whether as part of a homegrown initiative to move to multicloud operations or as part of an application restructuring for cloud, there are 10 major types of cloud management tools that help enterprises keep track of multicloud operations. All of them have functional value that should be considered, particularly if multicloud operations can, or eventually will, cover multiple concurrent processes. Whether or not an enterprise buys separate tools to handle all these functions, the list of categories is a good overview of the types of tasks that must be kept under close scrutiny and control to prevent a gradual slide of a multicloud environment into inefficiency or outright chaos.

IaC tools facilitate cloud resource provisioning and automation of some management functions.

Integration and API management tools control those functions closely.

Multicloud management platforms offer unified interfaces for dealing with different cloud environments and simplifying control of cloud interactions.

Cost management and optimization tools track cloud expenses across multiple platforms and can automate some aspects of choosing least-cost alternatives when there are multiple options for carrying out a certain task.

Monitoring and performance management tools help maintain uptime and optimal performance across a cloud network.

Container management and orchestration tools control containerized apps in a cloud network.

Cloud security and compliance tools provide a means of ensuring a multicloud environment follows security controls and, when necessary, conforming to regulatory requirements.

Cloud data management tools help control the storage, retrieval, and use of data in the network. Cloud migration and mobility tools assist in transitioning apps and data from one cloud environment to others.

Finally, there are multicloud management platforms, which often present SPOG dashboards that unify many cloud management tasks within a single interface, and cloud service brokers, who act as consultants to help businesses select and establish a system of cloud services.

Best Practices for Established Multicloud Environments

At some point, this seemingly endless list of tasks and concerns will come together into a multicloud arrangement. However, the related organizational changes and responsibilities are simply ready to begin a new phase.

Perhaps most fundamentally, the organization itself will have to undergo a cultural change. A more complex operating environment means new skills for employees to learn and longer checklists of aspects someone needs to keep an eye on. Multicloud environments are still relatively new; therefore, expertise in their operations is at a premium. Unless your enterprise has deep pockets, it’s not going to be able to easily hire experienced multicloud-environment administrators. Existing employees are going to have to take up the slack, and that means having employees willing to change how they do things and being capable of being trained to fill bigger shoes than they might otherwise be used to. Security procedures will have to be more stringent than ever, and some employee positions may not be able to access all the data and functions they have habitually had available. Openness to change, expectation of training (and retraining as procedures and technologies morph over time), and more hoops to jump through to accomplish certain tasks are to be expected and planned for. Employees must understand that things will be different; functioning as a mere cog in a corporate machine that never changes will not be possible except for brief periods. Those eager for challenges and advancement will adapt to this situation, but that won’t be everyone, and some fallout must be expected. Investment in employee training and skill development will be vital.

Aside from that, there will be a new list of best practices that need to be kept in mind:

• Backup strategy, business continuity, and disaster recovery plans will have to be altered and become more elaborate.
• Identify and track all cloud assets and recheck them frequently.
• This inventory will have to be rechecked frequently.
• The multicloud strategy will be temporary and must be revisited often.
• Enable centralized event monitoring and logging.
• Prioritize attention to security and compliance.
• Automate workload management as much as possible.
• Encourage app interoperability across the environment.
• Build a strong governance model.
• Regularly review and optimize operational costs.
• Stay updated on all related technological developments.

Clearly, there may need to be a larger group of people than the number currently employed to keep up with these constant multicloud housekeeping tasks. An increase in overhead will be unavoidable but may soon be the price of staying in business for some enterprises.

Reaping the Benefits

While the task and concerns lists may seem daunting and we have glossed over all the potential benefits, moving to a multicloud environment, particularly for larger enterprises, is becoming a necessity to keep up with the competition rather than a luxury. Multicloud is going to be part of the business environment until something like AI assistance or quantum computing becomes more readily available. If the alternative is eventual obsolescence, finding a way to be a multicloud-using enterprise should be viewed as an adventure in taking full advantage of the technological world we find ourselves in.