Look Before You Leap: Manual Attempts to Mass Migrate to the Cloud Often Fail
During the past year, more and more enterprises have begun migrating to the cloud. They’ve moved beyond the experimental phase and are running full production deployments. After years of buzz, the cloud has truly arrived. But getting there is not as easy as many organizations first think.
The Challenge of Mass Migrations
That’s because most enterprise migrations involve hundreds to thousands of workloads being moved. And moving this many workloads manually poses many problems. These challenges begin in the planning phase. First, you need to determine which workloads will be moved to the cloud and what changes they’ll need to optimally run there. This means you’ll need to figure out the unique compute, storage and networking requirements for each application. You must also identify all dependencies. Altogether, this phase requires an exhaustive portfolio assessment that is often rushed by organizations who are in a hurry to migrate.
Once you’ve decided what to migrate, you still have to actually move the workload, which is where it really gets tough. The primary challenge making the application run in the new target environment be that AWS, Google, Azure or a datacenter. In order to match your storage, compute and networking requirements into the cloud, you have to do an awful lot of tinkering. Many people don’t realize that the cloud is fundamentally quite different than their on-premises infrastructure. To recreate things in AWS or Azure takes a tremendous amount of trial and error.
The result of all this is downtime before you can cut over to the migrated workload in cloud. This is a huge problem. This process often drags on much longer than initially anticipated. I knew of one organization that was planning to move 1,400 applications to AWS. They got the ball rolling by migrating just 20 servers. They figured they could move the 20 servers over the weekend when employees were out of the office. Instead, it took them three weeks with long periods of downtime as they struggled to write script, configure necessary resources and convert those to the cloud.
As hard as it is to migrate to the cloud, it’s even trickier to repatriate back to your datacenter. Occasionally, customers realize that a cloud workload is more expensive than they had originally anticipated, so they decide to move it back on-premises. While cloud providers have some basic tools to help you move to their platform, they don’t have much to help you leave. Thus, the migration process is cumbersome in either direction.
Getting to the Cloud
Does this mean you should abandon any aspirations to embrace AWS or Azure? Absolutely not. The cloud has proven cost and convenience advantages that all enterprises should seriously consider. You simply need to rethink how you’re going to get to the cloud. Manual migrations are herculean undertakings that most organizations don’t properly plan and execute. To make the move smooth, you’ll want to leverage an automated migration platform that removes the need for manual scripting and resource configuration. Scripting and configuration are the two biggest bottlenecks in any manual migration. With an automated platform, you can skip the scripting and simplify the configuration to rapidly move workloads to the cloud with little downtime. Thus, you can actually complete a migration over the weekend rather than a period of weeks.
About the Author: James Kessinger is an experienced marketing executive who has built a career working at leading Fortune 500 technology companies in a variety of sales and marketing roles across hardware, software, virtualization, SaaS and Cloud. He has spent time in corporate, product marketing and field based leadership roles leading both marketing and sales organizations and currently is the CMO for CloudVelox.
Edited by Mandi Nowitz