In technology discussions, multicloud (or multicloud) is one of those frequently used terms, like big data, digital transformation and AI, which means different things to different companies — and often different things to the people inside those companies.
In its simplest sense, multicloud is about using more than one cloud platform or environment to run your business.
But however you define it, getting multicloud right is critical for building a connected business and transforming your digital enterprise.
In this Q&A, we spoke with Anshul Chadda, senior director of product management with Virtustream, and Ben Schreiner, of Boomi, about the key considerations, challenges and strategies for ensuring your multicloud success.
What Are Some of the Issues Companies Face When Considering a Multicloud Architecture?
Anshul Chadda: I think the biggest issue is usually data security. Obviously, you need to get that right. Otherwise, you can put your company at risk.
Along with security concerns, you’re also moving from a homogeneous to a heterogeneous environment, which means you need to allow applications and data to communicate across that environment, and this need for communication can cause issues with network performance and latency between clouds.
When first exploring multicloud, organizations often don’t have the personnel and the skills to manage multiple cloud platforms, so they turn to a provider like Virtustream to help them “keep the ship running.” We’ll handle the maintenance, and the organization can focus on app creation and innovation.
Multicloud architecture is another important consideration. What are the standard services you need to develop to have a consistent experience across multiple cloud end points? This set of services will depend on customer business needs and is different for each customer.
How Can Companies Ensure Their Multicloud Environment Supports Their Mission-Critical Systems?
Anshul Chadda: Often what we see in multicloud deployments are customers who are running mission-critical applications in a cloud, and it’s not working for them. Two years ago, many companies bought into “lift and shift” for their traditional mission-critical applications and moved to the public cloud. In most cases, this didn’t work for them. Reasons why it didn’t work? The required availability and performance wasn’t there, and there were no guarantees.
In addition, the architecture of public cloud is so different that they couldn’t just lift and shift their mission-critical applications without facing complications and set-backs. Companies need to make sure that their mission-critical application characteristics (availability, performance, architecture) are supported by the target platform, and that they align the definition of mission-critical with the inherent challenges of operating in a cloud environment.
How Should a Company Get Started on Its Multicloud Journey?
Ben Schreiner: I think you start with an understanding of what you’re getting into. As Anshul pointed out, this means defining mission-critical and what workloads you’re putting where. Make sure you understand the “why.” Why are you putting this workload on this cloud or changing where a workload runs? Inventory the whats and the whys.
And then as you assess different providers, make sure you understand what you’re buying. What are you getting, and what are you giving up? Understand service-level agreements, uptime, maintenance windows and what happens if something goes wrong.
Like any IT infrastructure, cloud platforms can go down. It just depends on the steps taken to keep that infrastructure running. So, consider what risks you’re offloading to a cloud provider, the key SLAs and the platform’s track record, and how that changes your support and management model for the customers of your business.
Anshul Chadda: This is consistent with what we see at Virtustream. It’s all about planning. People are excited about multicloud. They realize the benefits that multicloud can offer, but many have no idea how to get there. And the reality is that, while multicloud has many benefits, it also has complexities and complications.
IT decision makers are often pushed from the top down to explore what multicloud is. And instead of framing a deliberate and enterprise-centric strategy around what they need, they make ad hoc decisions. They don’t have a long-term plan that will get the most out of their applications while making sure everything is secure, compliant and cooperates across multiple clouds.
Organizations that have a plan in place and a future state in mind tend to be a lot more successful. They know what they’re getting into; they’re not jumping in blind.
What Role Does Integration Play in Moving to Multicloud?
Ben Schreiner: Multicloud presents a different integration problem than on-premise. When everything is on-premise, you can have an on-premise enterprise service bus (ESB) that connects all the apps behind the firewall. But with software as a service (SaaS) multicloud deployments apps can reside anywhere, but you still have the same data sharing and exchange needs.
Having a platform that allows for integration, API management, and the overall governance and movement of the information is what companies are looking for to help manage those complexities. Multicloud isn’t a less complex architecture than on-premise; if anything, it’s more complex.
So the ability to easily put information in or get information from various sources in your architecture and make it available to new services and applications is the value Boomi brings.
What Aspects of Multicloud Deployments Do Organizations Most Commonly Underestimate?
Ben Schreiner: Most companies underestimate the number of dependencies in their existing systems. Very few enterprise architectures account for all the downstream applications or business processes that use the information in a mission-critical system. When you move or change the application, you run the risk of breaking downstream processes, because information isn’t coming from where it used to. So, underestimating the amount of process disruption is a significant risk.
Anshul Chadda: I agree that’s a huge point when you’re moving to multicloud. I also think people tend to underestimate “day-two” operations. What people focus on when they’re moving to multicloud is the move itself, but they need to plan for what they’re going to do when they get there.
How do you get the benefits, such as automating processes, that were the “why” behind the move?
Another day-two concern is making sure all business processes are online. That’s also an issue if you’re doing an upgrade as part of the migration, which just increases the complexity and the day-two issues you need to pay attention to.
What Additional Advice Do You Have for Organizations That Want to Build Multicloud Environments?
Ben Schreiner: I would look at four things:
- Know what you’re trying to accomplish — the benefit, the why. It could be more flexibility, upgrading systems, or getting out of the hardware and maintenance business.
- Be clear on your definitions. For example, what does mission critical really mean for your organization? This will reinforce #1.
- Know the steps you’re going to take to achieve the benefit
- Understand the risks and that includes the risks of disruption as well as the risk of not moving to the cloud
Anshul Chadda: I agree that establishing the “why” of your multicloud move is the clear first step. You don’t want to move to multicloud because everyone else is doing it. Transformation with a purpose is what we call it.
Often organizations need help identifying and refining the purpose, so they work with a partner, cloud provider or system integrator. Then after they determine the “why,” they can start on the “how do we get there.” Sometimes organizations do this in reverse — the how before the why — and that can be problematic. For my advice I would say:
- Know what you need. Do you need extremely high availability and performance? Do you need that for all of your applications or just some of them? And make sure you work with partners and providers that align with your needs.
- Pay close attention to cost. If you’re “evacuating” your data center to move to a cloud OPEX model, understand what that will cost.
- Have a plan for day-two operations. The move went well. Now what? How do you get the most performance out of your application? How do you incent your developers to innovate more?
- What’s your plan for the future? Where do you want to be in three years? How big will your cloud footprint be then?