3 phases on the journey to multicloud

2021 was the year that government agencies fully embraced multicloud. As the pandemic shifted more work online and remote work became the norm, the move to a multicloud environment accelerated.

Now, however, many agencies are realizing that not all apps and workloads are suited to the cloud. Skyrocketing costs due to data egress, poor performance from lack of in-house skills to manage workloads in public clouds and security complications related to compliance demands have made many cloud workloads problematic.

That's why 2022 will be “the year of multicloud strategy” when federal IT leaders take a step back and prioritize creating a comprehensive plan for deploying multicloud environments.

There are three critical phases to thoughtful deployment that will allow agencies to reap all the benefits that multicloud has to offer. These three stages include assessing the current IT system, determining where workloads belong and mapping out plans and achieving business goals set out in those plans.

Phase 1: Assess the current IT ecosystem

Throughout this process, agencies may determine that some apps must be repatriated, or moved back from the cloud to on-premises. For best decision-making, this is the time to examine data egress (when applications send data back and forth from clouds or downloads and files are moved to external storage).

This is especially important for agencies that have experienced unforeseen cloud costs due to unanticipated egress costs and misunderstanding how chatty their on premises apps would be with apps in the public cloud.

While  cloud service providers can help agencies calculate expected costs, they can’t determine how chatty all an agency’s apps are, making it nearly impossible to predict true costs of moving certain workloads to the public cloud. The key to success for this phase will be less about applications and more about where workloads and corresponding data reside.

Phase 2: Determine where workloads belong and map out a plan

After thoroughly vetting the current IT environment and app workloads, it’s time to map out getting from an “as-is” state to the “to be” state. This process will encompass gauging what workloads need to be containerized and ported, which refactored and which rewritten entirely.

When it comes to determining whether applications should go to a private cloud, public cloud, colocation facility, edge environment or data center, agencies should consider:

1. Immediate needs and any triggering events for prioritizing applications for cloud migration.
2. Time and cost effectiveness, whether that's rewriting a cloud application, re-platforming or containerizing an app, rehosting an app (moving it as is or repurchasing it as a software-as-a-service offering), retaining the application as is or retiring the application.
3. The requirements, benefits and limitations of new infrastructure and consumption models.

One alternative to moving an entire workload is to move only the parts that need to be rewritten, turning those parts into microservices. Because rewrites can be especially difficult, this will make it easier to break the process down, to glean lessons learned, and to still have cloud-native instances that are easier to manage.

Phase 3: Achieve goals set out in the plan

Now it’s time to put that multicloud plan into motion to achieve agency goals. This phase should also be iterative, never stopping after implementation, and aim to reduce time to value, minimize risk and manage costs more effectively.

And while it’s been said time and again, it bears repeating that navigating multicloud is not simply a matter of technology. Successful transitions involve people, processes and technology. Agencies will have to prepare for a cultural shift, process changes and be equipped with the necessary technologies and training to enable successful multicloud deployment.

To effectively manage costs, agencies will need automated continuous monitoring that focuses on instances. Too often, organizations have been surprised by shadow IT where employees knowingly or unknowingly use cloud services that contribute to exhausting the cloud budget. Actively managing instances and services in the multi-cloud environment is vital to monitoring costs.

Additionally, it’s important to keep in mind that not all instances are created equal – some run better than others. That’s why it’s critical to continuously check in and understand which workloads perform best for large transaction databases and which perform best for large capacity but small transactions. Focusing on managing instances will help IT staff continually optimize the agency’s multicloud environment.

By incorporating a more thoughtful approach to multicloud, federal agencies stand to glean more of its benefits in the coming year, including increased agility, flexibility, efficiency, performance, security and cost management.

Stephen Smith is director of cloud technologies at Iron Bow Technologies.