Cloud Migration: 9 Ways to Ease the Agony
The pandemic spurred all companies to move to the cloud. But that doesn’t mean the job is done. Here are some tips to make the move less torturous.
![green robot lying on a cloud green robot lying on a cloud](https://eu-images.contentstack.com/v3/assets/blt69509c9116440be8/blt6bbd1dcc037f28ee/64cac0ab9fb8445c9859dbf9/00-greenrobot-_Zoonar_GmbH-alamy.jpg?width=700&auto=webp&quality=80&disable=upscale)
Zoonar GmbH via Alamy Stock
There’s a reason more ado is made about the benefits of operating in the cloud than on moving to the cloud. It’s rarely easy to do and prone to be botched. But what are you going to do? Competing in the current marketplace, let alone surviving a worldwide inflation storm, requires full digitalization and a seriously strategic cloud play.
Considering there’s little choice in the matter, the focus must be on how to make -- or finish -- your company’s move to the cloud as smooth and as painless as possible. After all, few companies can tolerate project delays and broken processes for very long in this economy.
Start by taking note of what’s driving you to move any given element of your business to the cloud. That bit of clarity will sort out a lot of potential issues from the outset.
“It's OK to have multiple drivers, but the primary goal must be clear. A big chunk of the pain related to cloud migration comes from lack of alignment on what's driving the migration,” says Daniel Viveiros, chief technology officer at CI&T, a digital transformation company.
Viveiros offered an example: When decommissioning a data center to cut costs or to improve business continuity, “your main strategy may be to re-host the existing application (lift and shift).” But if growing the top line via app modernization and new cloud-native architectures is the main driver, then “you'll need to re-architect or even replace your legacy systems for SaaS products.”
Click through for more tips on making cloud migration less of a misery.
Use automation to do some of the work, but also to check your team’s work for any gaps or oversights.
“Companies can decrease migration risk by automating as much of the discovery and dependency mapping as possible, eliminating human error and overcoming the hard truth that there are gaps in knowledge of what exactly a complete view of the IT estate looks like and where there is risk,” says Keith Neilson, technical evangelist at CloudSphere.
“Because no two IT estates are the same, providing configurable inputs and service completion graphs alongside discovery automation minimizes risk even further by allowing businesses to identify significant risk areas that must be considered for a specific organization,” Neilson adds.
If you prematurely pull out of discovery mode during the migration, you’re likely to make a mess of things. And cleaning up afterwards isn’t going to be fun.
“Another often overlooked issue stems from only using discovery and assessment tools during the initial discovery step and not throughout the entire migration project,” warns CloudSphere’s Neilson.
He advises that you stay vigilant by using discovery and assessment tools throughout the entire migration process. That’s because IT estates evolve, update, scale, and change all the time, he says. Your team must constantly be discovering what is or has changed. That will go a long way in preventing mistakes and oversights during the process.
But there’s another important reason to do that, too.
“Assessments and discovery steps, especially those performed early on, age fast and, ironically, can add risk, despite being designed to reduce it, if the data being relied upon is out of date and not current,” Neilson adds.
You need flexibility if you’re going to successfully stuff that crazy mix of things that runs the business into the cloud. If you’re not careful, you’ll move all those old data center problems to the cloud.
“Having a flexible cloud-native architecture in place is crucial, as being locked into a certain cloud platform or being unable to change the migration plan could hold your company back,” says Rishi Kulkarni, senior director and cloud native offer lead, Capgemini Americas.
Avoid thinking that you can take a one-size-fits-all approach or that cloud migration is just a technology or platform transformation.
“The chance of one platform sufficiently meeting all needs is slim. Certain workloads and applications operate more efficiently on certain platforms, but if you’re boxed into a mono-cloud strategy, options are limited,” warns Kulkarni.
For these reasons and more, “enterprises should avoid committing to a single cloud provider,” says Don Alusha, telco digitization senior analyst at global technology intelligence firm ABI Research.
You can take a shortcut or two along the way to build a hybrid approach to cloud migration and keep the company in its best operating form en route.
“A full shift to cloud may be the long-term goal, but containerization and orchestration provide a viable hybrid option to gain similar levels of elasticity and performance without moving all workloads to cloud in one go. This allows organizations to establish flexible architecture to unlock potential of multi- and hybrid-cloud deployments,” says Capgemini Americas’ Kulkarni.
But keep an eye on your target operating model in your cloud migration plan, too. In other words, take shortcuts and baby steps where needed, but make sure they move in the direction of your overall migration plan.
Resist the urge to lift everything up and put in the cloud as is, aka the ole ‘lift and shift’ strategy that bestows immortality on previous mistakes, errors, and problems.
“The obvious analogy is moving from one house to another: You accumulate a lot of junk over the years, and there’s no point in bringing it with you to the new place,” says Sandeep Uttamchandani, chief product officer for Unravel Data.
Instead, reevaluate everything. Keep what’s running and working and either dump or replace the rest. Then migrate only what your business truly needs to the cloud.
“The first question to be answered when you want to migrate data applications or pipelines is, ‘Migrate what’? In most organizations, this is a highly manual, time-consuming exercise that depends on tribal knowledge and crowd-sourcing information. It’s basically going to every team running data jobs and asking what they’ve got going,” Uttamchandani explains.
Why bother doing this? Because the usual practice of ‘lift and shift’ is being replaced by a more methodical and strategic modernization approach. It’s a smart move brought on by a lot of hard lessons learned in years and tears on cloud implementations.
“Examples of past mistakes include end-state value misses, such as higher-than-expected run rates, costs stemming from less than optimized workloads, operating model inefficiencies, and new skill hires in a competitive labor market,” says Alicia Johnson, consulting principal, technology transformation, at EY.
Common mistakes often stem from an either/or scenario such as focusing too much on each step or focusing too much on the bigger picture. To succeed, you need to be able to see both clearly.
William Collins, principal cloud architect at Alkira suggests you do that following these basic directives:
Know your applications and data and put them into two categories: those that are easy to lift and shift versus those that require some re-architecture.
Take measures to avoid disruption to existing infrastructure.
Create factory approach when moving workloads to clouds. Start with the lift-and-shift apps and migrate those first.
Assemble a roadmap, but don't panic when setbacks occur.
But balancing little and big things is about more than aligning baby steps with a big destination. Take edge computing in creating big market advantages, for example.
“Focus on enabling near real-time data access for core applications to avoid significant cost overruns or a poor user experience. Cloud-smart organizations do this and do it well. They are moving to a distributed core/edge network topology to enable direct cloud connectivity to internal data sources and external applications,” says Brandon Carroll, executive director of transformation, DevOps & cloud services at TEKsystems.
Organizing now can save you weeks of mopping up later. In other words, plan the sequence in moving each job to the cloud carefully.
“Workloads are all intertwined, with highly complex interdependencies. You can’t simply move any job to the cloud randomly, because it may break things,” warns Unravel Data’s Uttamchandani.
“That job or workload may depend on data tables that are potentially sitting back on-prem. The sequencing exercise is all about how to carve out very contained units of data and processing that you can then move to the cloud, one pipeline at a time,” Uttamchandani adds.
But once you have the sequences straight and jobs finished smoothly, don’t start thinking you’re done. There’s one more, very important step ahead.
“Once you have workloads migrated, you need to optimize them for this cloud environment. Nothing works out of the box,” Uttamchandani says.
Cloud-hosted tools and standardized containers make migrating to the cloud a lot easier, but that doesn’t mean there isn’t some work still be done. Integration is high on that list and that means viewing each job move as a part of a whole ecosystem and not a standalone project.
Further, when developing an integration strategy, keep in mind the underlying biases and proceed accordingly.
“Keep in mind that cloud-based data governance technologies by and large have a bias towards Intel/Windows/Linux ecosystems” says John Thielens, CTO at Cleo, an integration platform provider. “Workloads that aren’t good candidates for containerization or deployment in a cloud infrastructure shouldn’t be ignored, however. Even if they aren’t ‘in the cloud’ they can evolve to be managed in a ‘cloudy’ way.”
If you were aiming for cloud first, you’re too late. Word on the street has it that it up and died.
“Cloud first is dead. Many early adopters used cloud-first and simply moved technical debt, outdated business processes and bad code to infrastructure outside their four walls,” says TEKsystems’ Carroll. “Take a cloud-smart approach. Use cloud technologies that deliver deployment and delivery platforms for next-gen services that drive business value -- not just IT cost savings.”
Whether you’re going first or smart, remember that the cloud isn’t your end goal, making a profit and increasing market share is. The cloud is just a way to make things happen so that your company can compete and win. In other words, a quick and perfect cloud play is not the ultimate payoff you seek.
“Enterprises must have the willingness to stake their success on the ability to make commitments today that may not pay off until years into the future. In other words, their strategy should begin with a high tolerance for risk and what they do not know over cloud and associated challenges rather than with what they do, for successful strategies must be built upon a degree of unpredictability, not despite it,” adds ABI Research’s Alusha.
If you were aiming for cloud first, you’re too late. Word on the street has it that it up and died.
“Cloud first is dead. Many early adopters used cloud-first and simply moved technical debt, outdated business processes and bad code to infrastructure outside their four walls,” says TEKsystems’ Carroll. “Take a cloud-smart approach. Use cloud technologies that deliver deployment and delivery platforms for next-gen services that drive business value -- not just IT cost savings.”
Whether you’re going first or smart, remember that the cloud isn’t your end goal, making a profit and increasing market share is. The cloud is just a way to make things happen so that your company can compete and win. In other words, a quick and perfect cloud play is not the ultimate payoff you seek.
“Enterprises must have the willingness to stake their success on the ability to make commitments today that may not pay off until years into the future. In other words, their strategy should begin with a high tolerance for risk and what they do not know over cloud and associated challenges rather than with what they do, for successful strategies must be built upon a degree of unpredictability, not despite it,” adds ABI Research’s Alusha.
-
About the Author(s)
You May Also Like