Mike Waas
Mike Waas is the founder and CEO of Datometry, a SaaS database virtualization platform enabling existing applications to run natively on modern cloud data management systems without being rewritten. He has held senior engineering positions at Microsoft, Amazon, EMC, and Pivotal and is the architect of Greenplum’s ORCA query optimizer. He has 40+ scientific peer-reviewed publications and holds 20+ patents.

If there is one operation that just about every IT leader dreads, it must be database migrations. So much so that CIOs have routinely pursued an avoidance strategy. Instead of grabbing the bull by the horns, database migrations are typically passed on to the next CIO.

However, the cloud is changing the database market rapidly now. While database vendors rarely seem to agree on anything, everybody seems to be eager to take a bite out of the $40 billion market of conventional databases. The days of databases as a product seem numbered. Databases as a service is the future.

Because of this tectonic shift, migrations are suddenly top of mind. Yet, the number of successful completions clearly negates the hype. If everybody wants to move, and every business needs the new technology, and plenty of “solutions” to help move are available, then why is it that success is so hard to come by?

Read More:   Update Drill Now a Top-Level Apache Project

Here, we take a look at why database migrations routinely stump even the most seasoned of IT warriors.

‘New Revenue Generators’ Not ‘Resume Builders’

Too many IT leaders still view themselves primarily as the “bringers of new technology, ”and they do so often for the technology’s sake. Vendors have long catered to the decision maker’s ego and positioned their tech as the ultimate resume builder. How else could one explain the Hadoop-craze of the past decade before it fizzled out?

IT does have the mandate to shift from being a cost center to contributing to revenue. Yet, old habits are hard to break. One critical element is the fast and seamless adoption of technology. Adopting technology based on sound and comprehensive economic analysis as fast as possible may constitute a significant competitive advantage.

Vendors have long catered to the decision maker’s ego and positioned their tech as the ultimate resume builder.

On the flip side, this requires staying clear of vendor lock-in. The lower the degree of lock-in, the faster the enterprise can move. Any technology that demands massive rewriting of applications and reinventing of business processes may not be the right fit. Stay clear of any vendor who tries to sweep the effort required to adopt their technology under the rug.

Don’t be the IT leader who leaves their successor a highly fragmented landscape of database products with insurmountable vendor lock-in. Evaluate your choices carefully.

‘Clean Slate’ Approaches that Turn into ‘Scope Creep’

For us engineers, nothing seems more tempting than a “clean slate” approach. Rewriting, rebuilding, and reinventing is enormously gratifying. Nothing gets rid of the problems of the past as effectively as starting all over again. For applications that are clearly suffering a quality problem, a rewrite might indeed be in order.

However, moving from one database to another is typically driven by other factors. Going from on-premises to the cloud is usually not motivated by quality problems of working applications but by business factors. As a matter of fact, IT organizations will have to re-platform perfectly healthy applications.

What looked like a manageable project at first engulfs soon the entire business.

The approach of a clean slate very quickly turns into scope creep. As the initiative progresses, more and more applications need to be rewritten. What looked like a manageable project at first engulfs soon the entire business. Analysts and business users get enlisted to rewrite their SQL because there just aren’t enough consultants to go around to help with it and everybody wonders if this is truly the best use of their time.

Read More:   Update The Perfect Pair: Kubernetes and Distributed SQL

As the old adage goes: if it ain’t broken, why fix it? Just because you’re looking for a more economic database platform doesn’t mean you have to rebuild everything that’s working now — and will work exactly the same way when you’re done rewriting.

Overpriced Migrations

A $30 million spend over five years is a typical migration project duration for data warehouse projects. Gartner puts the success rate of these projects at below 15% . Teradata, the heavy-weight incumbent in this space, recently highlighted how it won back a customer because that company’s migration to another platform turned out to be too complicated.

If these projects are so riddled with problems, how come the industry has not gotten a better handle on it? For starters, we engineers have the reputation to underestimate the effort any software development project takes. Yes, rewrite-based migrations are software development projects, and should be treated as such.

Engineers have the reputation to underestimate the effort any software development project takes.

More treacherous, however, is the fact that migrations are classic 80-20 problems. The initial progress suggests the project will be a swift success. Then, come the 20% that will eat up 80% of the resources and drive up the cost relentlessly. Several years into the project, the funds dry up, the business becomes impatient, and all too often, the project collapses under its own weight.

Don’t get tempted by this most epitomic 80-20 problem: the first 80% of the challenge is a cakewalk. The last 20% is where these projects go to die — and you’re left holding the bag.

Bulletproof Your Career

Moving data to the cloud is not an option but an inevitable requirement for every IT department. But it doesn’t stop there. Solely swapping your on-prem database for a managed one does not satisfy the mandate for IT to contribute to revenue generation. Instead, IT will have to figure out how to transition from the data silos of today — cloud-based or not — to deeply integrated cloud databases.

Read More:   Establishing Best Practices for Adopting Kubernetes – InApps 2022

Data warehouses, the backbone of any new IT landscape in the cloud, are perhaps the hardest to move. IT leaders can prove their mettle if they execute a flawless migration. The stakes are high but so are the rewards.

Database system virtualization bridges the chasm between the old and the new in a powerful way: Existing applications can quickly move over from the current database to the new data warehouse. No major changes needed. At the same time, the business can build new applications on the new platform already. Simply put, database system virtualization gives you the best of both worlds.

Seasoned IT leaders understand that deploying resources and funds toward new development pays significantly higher dividends than rewriting applications for the rewriting’s sake. And in today’s world, the speed of technology adoption matters more than ever.

Feature image via Pixabay.