Moving critical data assets to the public cloud is one of the most challenging problems IT leaders are tasked to tackle in 2021. There’s a general misconception that data virtualization might somehow help in this migration. In this article, Mike Waas, CEO, Datometry, examines database virtualization as the alternative.
One of the most challenging problems IT leaders face in 2021 is moving data to the public cloud and finding a new home for the data warehouse. Enterprises must migrate from existing systems and get their data warehouse to the public cloud to innovate and compete.
Unfortunately, there’s a general misconception that data virtualization might somehow help with this migration. It probably speaks more to a desperate need and the complexity of the problem than for an actual solution. The conventional approach of database migration is such a terrible ROI that everything else might just be a better solution?
First, consider the underlying concept of data virtualization. It really is the idea of integrating all databases into one single uber database. This new database system will then query its constituents and combine the different data sets. It is really a data integration system more than a database. To use it, one has to interact with it in a new SQL-like query language.
Let’s take a look at some of the misconceptions in this space and the top four reasons why data virtualization is unlikely to solve an enterprise’s problem of moving to the cloud.
Learn More: Why System Integrators Must Modernize to Enable Data-Driven Business Models — Or Be Left Behind
1. Esperanto Is a Great Language — That Hardly Anybody Speaks
Remember Esperanto? It’s the synthetic language that was supposed to solve the problem of translating back and forth between languages worldwide. Esperanto speakers would be able to converse with ease anywhere they went. No need to learn English or Mandarin, or any of the hundreds of other languages — if only everybody spoke Esperanto. So much for the theory. More than a century after its introduction, Esperanto is effectively spoken only by a small group of linguistic enthusiasts.
What’s this got to do with data virtualization? Data virtualization is effectively a modern-day take on the same issue. By creating a universal database language, its inventors were convinced the need for individual database languages would soon disappear. And with it, also the need for database migrations.
Today’s labor market offers plenty of talent for any given SQL dialect. Database vendors spend a fortune on educating their customers on their technology. They have no incentive to train customers on a generally weaker language (more on this below). Not to mention a language that can easily be used with their competitor’s system.
As a result, today, there is no critical mass in terms of skilled employees in the market for data virtualization languages. Like Esperanto, the idea of a universal language is appealing from a conceptual point of view. However, its implementation remains elusive.
Learn More: Open Data: A Catalyst for Driving Innovation in Finance and Healthcare
2. Standardization Eliminates Competitive Advantage
Any data virtualization system performs only as well as the most limiting system that it supports. This has far-reaching consequences. To live up to its promise, data virtualization systems ensure that users are limited to functionality that is common to all existing platforms.
Since database systems differ widely, this limitation may stifle innovation severely. Unless they are implemented in the data virtualization system, enterprises won’t be able to use powerful features of a database such as Stored Procedures or advanced analytics functions. Until they are available in all competing database systems, they won’t be able to use these features.Â
Again, this situation is not new. The standardization efforts of SQL itself have a long and checkered history of trying to rein in vendors. However, it seems every vendor has strayed from the path over the years, sometimes to work around specific architectural problems of their system. But more often, to provide their customers with a competitive edge in the form of new features.
This creates a conflict for data virtualization vendors. Enterprise customers will happily color outside the lines and use proprietary features of just about any database if it gives them a clear competitive advantage. Data virtualization may then be thought of as a limitation rather than a solution.
3. To Solve Your Migration Problem — You Must First Migrate?
And then, there is the problem that to use data virtualization systems, the entire enterprise needs to commit and implement this transition. At first, this may seem akin to giving up any bad habit, like bad dietary choices. It is clear that implementing a better regimen will have health benefits in the future.
However, there is a big difference between giving up a bad habit and transitioning to a data virtualization system. The latter does not just force companies to make a change going forward. It forces them to start over. To leverage a data virtualization system, the enterprise must first migrate to the data virtualization system.
To make matters worse, migration to a data virtualization system may actually be harder than migrating to another database. Given their limited nature, adopting this technology may require simplifying or even downgrading existing functionality. This is not only a formidable technical challenge but may be unacceptable to the business.
Data virtualization comes with a rich value proposition but also with a high barrier to entry. Enterprises need to understand the trade-offs. They also need to take into account their own track record for implementing business processes years in advance. This type of investment may not be practical.
Learn More: Getting the Data and AI Implementation Right for Your Organization
4. High-Risk and Unclear Return
This brings us to the last point. As we just saw, in terms of effort, there is little difference between migrating to a new destination database and migrating to a data virtualization system. In both cases, every single application needs to be rewritten or adjusted.Â
To many enterprise data warehouse (EDW) owners, this may ring familiar. Most IT leaders responsible for a data warehouse appliance have at some point in the past tried to move off it. To them, having to migrate to a data virtualization system just to migrate to another database system may sound like double jeopardy.
Even if the migration to the data virtualization system succeeds, there is no guarantee that existing applications will work as hoped. Instead, additional time and budget need to be set aside to allow for troubleshooting and engineering work to finish up the project.
Most IT leaders are under enormous time pressure to decommission the legacy database. Complicating the move with an additional migration ratchets up the risk considerably.
Database Virtualization — Not Data Virtualization — Is an Alternative
Data virtualization is an interesting concept, even more so for enterprises that have the luxury to start from scratch and who don’t view database technology as a competitive advantage. However, it does not solve the re-platforming problem from on-prem to cloud.
Database virtualization, as opposed to data virtualization, offers a very different solution. Database virtualization takes a page from the core concept of virtualization. It connects applications and databases with a hypervisor-like platform. The hypervisor emulates the functionality of the legacy database in real-time.
Database virtualization adds the capabilities of the legacy system to the new destination database. That is, with database virtualization, a modern cloud database can “speak†legacy languages. The new stack supports existing applications then without costly, time-consuming, and risk-laden migrations.
In contrast to data virtualization, database virtualization solves the “last mile†problem by connecting applications as-is to the new platform. No rewrites are required. This makes database virtualization a powerful alternative to manual migrations.
Did you find this article helpful? Tell us what you think on LinkedInOpens a new window , TwitterOpens a new window , or FacebookOpens a new window . We’d be thrilled to hear from you.