Five Steps to Creating a Multi-Cloud Optimisation Strategy

Brandon Tikalsky, Sr. Director, Cloud & Network Architecture
May 30, 2019

Chances are you are using multiple clouds with a mix of different service types and delivery models if you’re involved in business technology today. In fact, the average organisation uses five different clouds, and that doesn’t even take into account the individual apps and services within those environments. When you stop to consider that Amazon Web Services alone has more than 90 cloud products to choose from, the picture of the average cloud ecosystem starts to look overwhelming.

One of the reasons for this complexity is simply that organisations have more choices than ever before in how they build and manage their cloud architectures. In the early stages of the cloud’s evolution, many of the key questions revolved around whether to rely more on public or private clouds. Today, the answer to that isn’t either/or, it’s how many of each.

Instead of picking one model and sticking with it, you can now pick whichever best fits the needs of a specific business demand or service that you want to run at any given moment, and often change on the fly. With this greater complexity, though, comes a critical need to weigh how to evolve your multi-cloud management strategy to optimise performance, cost, and availability.

How to Manage Your Multi-Cloud Management Strategy

In thinking about how to manage a multi-cloud architecture successfully, it’s important to get back to some of the advantages that have driven cloud adoption in the first place:

• Optimised workload distribution by transferring workloads between private and public clouds
• Redundancy through the use of multiple cloud services and regions
• Reduced dependency on a single vendor
• Best-of-breed approach for individual services

The challenge for modern businesses isn’t that they can’t achieve these benefits, it’s that cloud optimisation becomes exponentially more difficult as firms add multiple environments from many different vendors — especially when you’re trying to optimise between several different vectors like performance, security, and cost. The complexity of managing costs in a multi-cloud environment is a concern echoed in RightScale’s 2023 State of the Cloud report: 64 percent of organisations surveyed had cost savings for existing cloud use as a top initiative.

What modern organisations require is a dynamic framework to ensure that their multi-cloud management strategy continues to evolve alongside their investment in new services, vendors, and options. This framework will help to identify service-level needs as well as other investments (e.g. cloud management platforms, networking) that may be necessary to meet their business and IT demands.

Steps Toward Optimizing Multi-Cloud

Just as there is no one-size-fits-all cloud environment, the same is true for multi-cloud strategies. However, you can take steps toward developing a dynamic process for adjusting to rapidly changing IT demands and needs.

Below are several key considerations and steps to take, first for understanding where your cloud environment currently stands and then to transform your organisation into an agile cloud environment for the future.

Step 1: Map Your Cloud Environment

Mapping your existing cloud environment may sound like a basic task, but going through the exercise can yield considerable value. Creating a high-level picture of your current cloud environment and needs may reveal opportunities for future optimisation as well as gaps in your current strategy.

Think of your map as the overall conceptual design of your cloud architecture. Your map should include different cloud services, how you connect to them, how they connect to each other, and what your service needs are for each. For each cloud service your business uses, consider listing the following elements:

• Cloud vendor used
• The departments that use the service
• Any service-level agreements in place
• Related business functions (e.g., used by marketing and sales, used primarily by operations, etc.)
• There are a few other key considerations to keep in mind as you’re mapping your cloud services. For example, identifying your performance needs and comparing them to your SLAs may reveal areas for improvement or a need to switch vendors.

You will also want to consider the cloud architecture and delivery models you’re leveraging. Not all your data or every app you run needs to be in the public cloud. This exercise can help you discover which public cloud services may be better suited in the private cloud environment or vice versa.

Step 2: Map and Plan Integration Needs

Once you have a high-level view of your full cloud environment, you can start to map your integration needs much more easily. Consider the types of environments that you have and how they will share data. For example, if you have separate private cloud environments, will they require data or application integration, or both? Here are a few broad considerations to think about as you map your cloud integration strategy:

• Determine how your private cloud environments will communicate with each other
• Determine how your public clouds from different vendors will communicate and share data
• Determine if there are any existing gaps and close them
• Identify whether you need to build middleware to connect different services

Although integration is a challenge often cited alongside public software-as-a-service adoption, it can present issues in other cloud models. Ensuring seamless integration among private clouds, for example, may require IT to reconsider networking infrastructure across different sites to ensure high performance and availability. The challenge becomes more complex when using third-party vendors as there is no guarantee that a given vendor will publish APIs or make it easy for their services to communicate with others.

Although not every service in your environment needs to be integrated with each other, cloud adoption without cloud integration means your business will continue recreating work across different environments, wasting some of the benefits of migrating to the cloud in the first place.

It’s important to consider what end users will require to be productive when mapping your integration needs. Seamless integration between applications and data is vital for services essential to critical job functions. You can see where there are gaps and begin to close them by mapping your current set of integration tools.

Step 3: Establish an Automation Framework

Automation is among the universal benefits of cloud computing, but, in a multi-cloud environment, you have a much larger range of considerations to make in your automation plans. Large businesses will already have at least one tool for automating cloud deployment, but they may not necessarily have a framework for continually evaluating these tools and matching them to their scalability needs.

While it may seem counter-intuitive, the best place to start thinking about your automation framework isn’t necessarily with technology platforms. Instead, think about business needs and features. After all, if you build a platform-agnostic strategy, you can check any current tools you’re using against it. If the two aren’t aligned, it’s time to find a platform that is a better fit for you.

First consider: where is most of your time spent and what needs to be automated? For example, if you plan to invest in many different cloud services over the next several years, you may benefit from ensuring standardized rules for deploying different cloud environments are included in your automation plan. That way, you can translate those into configuration rules when you’re ready to automate new deployments.

This step also helps to build off the value of your cloud environment map and integration considerations. Once you understand which apps need to communicate, you can not only better plan how to automate between clouds, you may also identify new opportunities for performance optimisations.

For example, if there are new integrations that must be created, you may also want to consider the networking infrastructure that supports these connections. If it’s a public cloud, how do you connect to the vendor? If you’re looking at private cloud environments, are there performance tweaks you can make within your own IT environment?

Step 4: Plan Your Networking Needs

Although considering needs like processing power and data storage is important, the infrastructure that connects it all is equally so. When looking at simple cloud topology, the network infrastructure can show a modest degradation in performance compared to using non-cloud infrastructure. When you start to consider more complex multi-cloud architectures, each small negative effect on performance starts to add up and can lead to more significant problems, ranging from constantly high latency to service outages.

To set the stage for this plan, consider how your current cloud architecture is built, and where you want to be in the next five years. Will you primarily be using public cloud, private cloud, or a mixture of both? Which mission-critical business apps are sharing data and how are those connections made?

While there is no one-size-fits-all networking strategy, networking infrastructure is a key area for cloud optimisation tweaks. For example, hybrid cloud architecture is an ideal fit for cloud interconnection services that connect your internal IT environment directly to providers, so that you rely less on public internet and can gain higher performance and security through direct connections.

This can also be a good time to consider services that can enhance your overall network. For instance, colocation is often touted for its ability to reduce spend on building infrastructure, but these services can also offer performance benefits. One area where the performance benefit is gaining prominence is in edge computing, where businesses leverage colocation to move their IT infrastructure closer to end users.

This trend is already taking place in the consumer streaming video market, where performance demands are pushing TV and video providers to move IT closer to the people consuming their products. However, the same benefits could also be realized across other industries. Businesses with multiple satellite locations, for example, could use colocation to bolster performance without investing in entire new data centres for newer offices.

Step 5: Create a Strategy for Evaluating Future Cloud Adoption

Once you have a comprehensive understanding of your current environment and a broad idea for your goals, you can begin to establish your framework for evaluating future vendors and services. For example, you will have an understanding of which services are linked with the map of your environment and which will need to be linked as you migrate more services to the cloud.

Some additional future action items include developing a framework for determining security needs as you adopt more services and your strategy for translating between business and IT needs on a regular basis.

It may help to think of different “tiers” of your data, where the highest tier requires the highest protection, highest level of redundancy, and/or highest performance, for considerations like security and service-level requirements. The way you build this framework will depend heavily on the unique needs of your organization and your IT environment.

However, another key element that can help with planning for the future is identifying individuals within your organization who can translate between business and IT needs as your organization considers cloud migrations. For example, if sales and marketing want to adopt a new CRM system, they would work to identify must-have features, develop an idea of access control standards, and determine feasibility for specific feature requests. All of this will inform vendor and platform selection moving forward.

For example, a Digital Realty partner looking to upgrade its cloud network had broken the migration into different phases. Although phase one was focused within the U.S., one of the primary factors in the decision to leverage Digital Realty was our global footprint, because our partner knew it would also need to simplify and upgrade the network internationally in the future.

This process will enable IT and business departments to collaborate more effectively, which will be an increasingly important quality as IT and cloud become more integrated into overall business strategy. Ultimately, this helps you to formalize many processes that may have previously not been written down or conducted in differently throughout your organization.

Your Path to Multi-Cloud Optimisation

The bottom line is that many businesses are already running multi-cloud environments and they may even have some foundational strategies for managing them. However, we expect to see exponential growth in the number and variety of cloud services the average business uses going forward. The organizations that will succeed in managing this complexity will be those who develop a dynamic, evolving framework that successfully links core infrastructure and cloud architecture needs, across every layer of the stack, with business needs.

Without a dynamic framework and a comprehensive map of their IT environments, businesses are missing out on cost reduction and performance optimisation opportunities. In other words, you are leaving a money on the table by not taking a closer look at your cloud environment architecture.