Regaining Control of Your Cloud Destiny: A Case Study
Generic questionnaire-based cloud assessments run by central IT generally miss the mark on what the business needs out of a cloud migration. Sick of hearing “Not suitable for cloud”, one CIO took matters into her own hands.
One sultry winter’s day in 2017, Samantha sat by the POTS phone that was provided by IT and waited for a call. In anticipation of what word might rain down from IT, the CIO of a major business division gazed out her window, through the falling sleet, and into the half-empty parking lot. “Where is everyone?” she whispered to herself. “Do they know something I don’t?”
Then the phone rang.
“Your application is not suitable for the cloud.”
Her heart sank. “Will IT really try and force a 3 year $50 million hardware refresh project down my throat when the business is screaming for greater agility and lower costs to operate? Not on my watch”, she vowed.
If this fable sounds familiar to you, please know that you are not alone.
Stories such as this one are often told by veteran cloud migrators at cloud conferences the world over, from The Venetian to the MGM Grand. We learn from each one, creating organizational change strategies for side-stepping the status-quo and forming approaches that create long-lasting value for our cloud-migrating clients.
One thing each war story has, is a lesson learned that bears sharing with others in the community who will be facing the same challenges at some point.
This case study I will share focuses on a legacy suite of applications that had only recently gone live but were already approaching their hardware refresh cycle. Containing majority COTS products from Oracle, IBM, and Microsoft, there was perceived “little benefit available in cloud” from central IT, who were eager to keep the applications on AIX in the datacenter.
After weeks of painstakingly completing a 360+ question spreadsheet for central IT to evaluate its suitability for cloud, the “not cloud compatible” assessment results were already known to the project team members who were paying attention. However, having already had success in delivering cloud native applications on Azure and AWS, this group pushed back on the findings of central IT and decided they needed to do more of the heavy lifting themselves.
Working backwards from what the business needs out of their applications in the future makes the most sense when you’re about to invest money in a migration initiative. The organization’s business goals were clear:
- Lower OpEx. $7.2m pa in hosting costs for 30 daily active users is too high.
- Faster duration than a 3 year hardware refresh. Spoiler: they did it in 18 months
- Lower management overheads. They hadn’t been able to keep their applications up to date with the vendors’ releases. Changes on premise, with strict ITIL processes, were simply too hard.
- Better security. Log all the things, and let my people use the logs!
With those 4 goals identified during discovery, the newly formed migration team could now get started on performing a preliminary cloud application assessment.
Getting It Done
Leveraging the Tidal platform and Tidal Tools, the migration team quickly discovered their infrastructure, application technologies, and databases. On the components with custom software, they analyzed the source code too, quantifying the potential roadblocks they might encounter. Discovery was completed in 2 weeks, and yielded more than 50 actionable insights for the migration team - far more than what the 360+ questionnaire had produced.
To get a sense of potential cloud costs in either a Rehost (lift-and-shift), optimized, or transformative migration, the team went to Calculator to view their potential total costs in seconds. The app hooked into their discovered inventory, required no agents, and allowed them to select their cloud region and preferred currency.
Optimize for FAST
With the potential savings now quantified, the team then moved on to satisfying goal number two: Be faster than a hardware refresh project. Considering the time needed to gather hardware requirements, get quotes, order, ship, rack, cable, install operating systems etc. the cloud-savvy migration team said “We got this. If we invest some time upfront in Infrastructure as Code and employ patterns like immutable infrastructure, we will be able to measure twice and cut once.”
They were partly right, but underestimated the value of having cohesive and repeatable architecture automation.
The team was eventually able to automate the provisioning of the entire legacy application stack: 8 COTS components, across 30 individual nodes and many other cloud native resource types. In the end, because of their investment in Terraform and Ansible pipelines, the team was able to stand up four environments in just weeks. What’s more, they were able to test and validate their migration multiple times before go-live, ensuring that the business had peace of mind going into the fateful cutover date.
All in all, the 3-year estimate for hardware refresh was completed with 18-months of automation. The investment made in automating the management and deployment of the application continues to pay off for the team now running the application.
In the three years since going live in the cloud, the team has been able to upgrade the other COTS components in line with vendor support matrices, reduce downtime and improve morale of a newly empowered group of people. The increases to agility (goal #3) have been very much appreciated by the business and raised the bar for future IT projects.
The team even earned an award for their project, by demonstrating that IT investments can truly have long lasting business impacts. This is something that an on-premise hardware refresh project has yet to demonstrate.
The Bottom Line
As you evaluate your options for going to cloud, consider that everything can go to cloud. We are merely moving service delivery from one paradigm of technology delivery, to another.
Each of the 6Rs of migration offer different benefits and come with tradeoffs that you will need to consider. Be wary of the silver-bullet migration pitches, and run a rapid cloud application assessment that considers your business goals, as well as each application’s technology composition.
It is unlikely that you will get more funding from the business later, especially if it is not executed well, so be sure to invest in your migration initiative wisely.
David Colebatch is the Chief Migration Hacker at Tidal. Connect on twitter at @dcolebatch or follow on LinkedIn.