Maximizing automation helps prepare agencies and their workloads for cloud deployment. Those same steps will make cloud migration and moves between clouds successful.
More and more agencies are moving technology functions to the cloud. However, merely becoming cloud-based doesn’t always accomplish what they hope. That isn’t something one wants to find out after investing financial and other resources to make the switch, explains Steven Carter, Chief Cloud Architect for Red Hat’s public sector.
Instead of jumping head first into the cloud, Carter suggests agencies might first benefit from increasing automation. Automation improves the efficiency of the varied applications that operate in a complex, diverse and distributed information technology environment.
Maximizing an organization’s automation helps prepare agencies and their workloads for cloud deployment. Those same steps will make cloud migration and moves between clouds successful.
Getting more from the cloud
One of the more common mistakes agencies make when moving to the cloud, according to Adam Clater, Chief Architect at Red Hat, is something he calls lift and shift. “I’m going to pick this up and put it into the cloud and eliminate the cost of my data center and eliminate the cost of my hardware refresh, turning that workload into a 24/7 operation with cloud provider X,” Clater says.
While many agencies have taken this option, it doesn’t make the most of cloud technology. Instead, Clater advises automating the lifecycle of the workload. “How does it work going into development? How does it work going into production? How do we scale that workload horizontally? Then how do we scale it back down when the workload is reduced.”
As an example, because of tax season, the IRS needs more servers online in March and April than in May. It is valuable for the IRS to scale up its infrastructure in an automated fashion inside the cloud to address that demand and then scale back after filing season is over.
Reduces IT workload
Automation allows agencies to expand into the cloud without putting a crushing operational burden on IT teams. For example, if an agency is using Amazon Web Services, they must adapt to the Amazon platform, Carter explains. That doubles the complexity of trying to work in the cloud. Without automation, those complexities increase exponentially every time a new cloud or process is added.
Minimizes human dependence
In many organizations, there are key IT individuals who know how everything works and functions. Carter refers to these individuals as “heroes.” This creates many problems, not the least of which is your operational capabilities walking out with a disgruntled employee. Since automation provides a way to describe the infrastructure as code, these “heroes” become part of an integrated team that contributes institutional knowledge and best practices into a central repository with change management and revision control. Using automation, agencies are less affected by staffing fluctuations. When virtual servers or other infrastructure needs provisioning or updating, automated processes are a better option for consistency, resilience, a minimal error.
Protects against problems
A range of problems can occur with cloud providers. Clater cites situations like a significant outage, a price adjustment or loss of a security certification. When that happens, an agency must be able to quickly move workloads to another cloud provider. Automation then becomes the key element of getting into or out of any cloud.
“Once you’ve figured out how you’re going to automate workloads into a cloud, you’ve also figured out — in some way — how you’re going to automate your way out of those clouds,” Clater says.
Carter and Clater recommend a staged approach to cloud adoption. They suggest choosing a small, user-focused application you can use to build a proof point around both cloud and agile development approaches. In the beginning, stay away from applications that require wholesale change to existing data and application infrastructure.
Small successes will allow agencies to build subsequent application migrations on the patterns and procedures developed in the process of successfully moving these smaller workloads.