By using website you agree to our use of cookies as described in our cookie policy. Learn More

TDWI Upside - Where Data Means Business

Strategies for Successful Cloud Migration: Lessons from TDWI’s David Stodder

David Stodder, TDWI’s senior director of research for business intelligence, discusses trends and topics in cloud migration and management.

In a TDWI “Speaking of Data” podcast earlier this year, TDWI’s David Stodder and host Andrew Miller explored the challenges and considerations of cloud migration and management. Stodder is senior research director for business intelligence at TDWI. [Editor’s note: Speaker quotations have been edited for length and clarity.]

For Further Reading:

The Key to a Successful Cloud Migration Is What Comes After

Why Understanding Data Gravity is the Key to Data Migration Success

Plan Carefully When Migrating to a Cloud Data Warehouse

Stodder began by explaining that although the basic idea of a cloud migration is simply moving data analytics to the cloud, in actuality the process involves a great deal of variation.

“There’s the question of the degree to which a company wants to put their data and analytics in the cloud. There’s the question of what kind of cloud, whether to opt to let the provider handle everything, and so on.”

When asked about key trends and drivers, Stodder was quick to point out that cloud migrations don’t happen in isolation. “Organizations are also trying to modernize their data to support better analytics, to allow more users, and to address problems in their legacy systems. The cloud is at the intersection of many of those issues.” He went on to point out that TDWI research has consistently shown the same five reasons as the main drivers for cloud migrations:

  • Expanding data access with lower upfront investment
  • Optimizing data pipelines
  • Improving data quality
  • Increasing the speed and scale of data processing
  • Integrating disparate data silos.

“Some organizations are even still digitizing paper-based processes for use in analytics,” Stodder added.

When it comes to challenges to migrations, Stodder’s key takeaway was that with all the variations in cloud implementations, organizations should make sure they choose the approach that aligns with their intended goals.

“An important consideration is how, exactly, you want to move your data to the cloud. Do you just move it wholesale -- aka, lift and shift? The problem with that [approach] is you can wind up carrying over the problems you had with your legacy systems. You may then decide to do a more phased migration, evaluating each phase as you go. However, in that case, you can find yourself navigating a tangle of dependencies among different applications and users.” Stodder acknowledged that users will inevitably be disrupted in any cloud migration, but a properly done phased migration usually produces better results in that regard.

“Another option is to start small,” Stodder explained. “Leave the things that are working on premises and build any new projects in the cloud. There is the potential missed opportunity to modernize the legacy systems, but small steps can help minimize disruptions.” The worst option, he explained, is to come at a cloud migration piecemeal without any overall strategy.

This brought up the subject of hybrid architectures -- combinations of on-premises and cloud data management.

“There’s a lot of concern about this, particularly about governance,” Stodder said. “Governance requirements are becoming more restrictive these days, especially in highly regulated industries such as finance and healthcare. Organizations in such industries really need to keep track of sensitive data.” He noted that companies are taking one of several approaches to dealing with sensitive data.

“Some are choosing to keep that data on premises, perhaps in a data warehouse specifically for the purpose. Others who are more confident in the security of cloud data warehouses are more comfortable moving sensitive data there.” As a result, it becomes a challenge to inventory all that data, especially when it’s spread across a variety of on-premises and cloud systems. “That’s one of the things driving interest in data fabrics.”

Data fabrics are an architecture that creates a layer -- usually driven by a data catalog or other metadata system -- that encompasses the complete data life cycle, wherever it may reside. This layer allows companies to govern and inventory their data, and then use that to improve data access and stability. This becomes especially important when performing a phased migration, where users must live with a distributed architecture for at least some amount of time.

“Organizations need to understand that, try as they might, they’re just not going to be able to consolidate all their data. There will be mergers and acquisitions, the business will reshape itself, and so on. Distributed architectures will always be a part of the landscape.”

Stodder concluded with some simple advice for those beginning a cloud migration journey.

“Think of everything,” he said, reiterating that cloud migrations don’t happen in isolation. “Think of all the things you’re trying to do with your migration. You’re trying to improve user experience. You’re trying to build better analytics. You also don’t want to lose the expertise you built up developing your on-premises systems.

“Another thing to think about is cost management. In our research, we see a lot of sticker shock when systems are moved to the cloud, so it’s important to give some forethought to monitoring and managing your cloud costs.”

[Editor’s note: You can listen to the entire conversation on demand.]

TDWI Membership

Accelerate Your Projects,
and Your Career

TDWI Members have access to exclusive research reports, publications, communities and training.

Individual, Student, and Team memberships available.