Last week my colleague, Clay Ryder, and I presented a webinar, titled No! The cloud is not someone else’s data center, in which we examined how companies can reduce the complexity of a cloud migration and accelerate the benefits of digital transformation. It’s an important topic, so as a follow-up to the session, I’ve summarized five key things you need to understand to be successful in the cloud. If you missed the session, you can listen to the full discussion at the link above.
1. Understand the true value of the cloud
Let’s start with the title—it’s been said that the cloud is really just someone else’s data center. While a given technology may be the same regardless of where it’s physically located, this misses the point. What the oversimplification overlooks is that the cloud offers a strategic opportunity to enhance an organization’s efficiency and expense structure and, more importantly, to position itself for competitive advantage in the market. As George Westerman, Research Scientist with the MIT Sloan Initiative on the Digital Economy, said, “When digital transformation is done right, it’s like a caterpillar turning into a butterfly, but when done wrong, all you have is a really fast caterpillar.” To be transformative, you have to reimagine the business to find new opportunity for growth and better business outcomes. What this means is you need to take a workload-centric approach in your cloud strategies, because it’s the workloads that deliver business value. The cloud is the enabling infrastructure.
2. Understand your workloads
Of course, to be workload-centric, you need to understand your workloads. You don’t want to simply migrate existing workloads from your data center to the public cloud. The move itself doesn’t necessarily deliver benefits. And you don’t need to transform everything—you can focus on key parts of your functionality. For example, many companies migrate their front-end processes, keeping mainframes on-premises. Here at Virtana, we perform this analysis ourselves to develop our migration strategies. Here are two examples that resulted in different decisions for “lift and shift” migrations:
- Use case 1: Training—We deliver a lot of virtual training to customers around the globe, with unpredictable student attendance and conflicting instructors’ schedules, and we can’t afford any latency, so it made sense to use the global and elastic capabilities of the cloud to move those workloads as close to the customer as possible, no matter where in the world they are.
- Use case 2: Data analysis—We have analysts who collect and analyze data for performance assessments and forensic analyses. Because this work is very IO-intensive, the analyst can’t afford to turn off the systems, and the data sets can range from 1 to 5 terabytes, these workloads are better kept on-premises on assets that are already depreciated.
The bottom line is that you must take both the behavior and the business value of your workloads into account before making any migration decisions.
3. Understand what’s in your data center
It goes without saying that you need to have a handle on the ultimate cloud destination for your workloads. But a successful migration also requires you to understand where your workloads are coming from—what the point of origin looks like. The reality is that most organizations don’t always know what they have in their data center as often their system of record is out of date, so an application discovery exercise is essential. This process should identify dependencies, as well as various workload characteristics, and it should incorporate seasonality. All of this up-front work is critical to ensure that you’re making the right decisions about the sequence in which workloads need to be moved, and that you’re setting the right expectations about SLAs in the cloud.
4. Understand the behavior of your targeted applications and workloads in the cloud
Once you understand what’s in your data center, you’re in a better position to model what will happen when your targeted workloads are operating in the cloud. It starts with a cloud mapping exercise to evaluate candidate configurations and their associated costs. You want to do this with high-fidelity data to ensure the selection is rightsized for your workloads. Then you want to simulate those workloads in the targeted cloud service provider to validate, not just your configuration selections, but even your strategic approach. For example, based on the simulation, you may realize that lift-and-shift may not actually be appropriate for certain workloads. Uncovering that reality at this point—rather than after the workloads have been migrated—can save you significant time, money, and headaches.
5. Understand how to become a “migration factory”
Cloud migration isn’t a one-time event. Businesses evolve, which means they need to keep transforming. And the workloads that are already in the cloud will evolve and drift over time. Brute force may get you through one migration event, but that’s not scalable and it’s not sustainable. You need a single, comprehensive platform to create your migration factory, so you can keep track of what has moved, what will be moved, and everything going on in between across all your on-premises and hybrid/multi-cloud (public and private) environments.
#KnowBeforeYouGo
“Know before you go” with Virtana Platform, a single AI-powered observability platform to migrate, control cost, optimize performance, monitor, and drive uptime for your infrastructure across data centers and private and public clouds. Request a demo.
Ricardo Negrete
