How to Up-Level from Teradata to Azure SQL DW Without Losing Sleep – The Dato...03.09.19
In a recent blog, Julia White, Corporate VP for Azure at Microsoft, made a stunning observation: “[The] SQL Server 2008 [..] migration opportunity alone creates over $50B in opportunity for our partners”. This number includes a wide array of tasks and services. From moving the actual data to rebuilding dependent apps. The magnitude of this number is even more significant when you realize she refers to a move within the SQL Server family: From SQL 2008 to SQL DB – which is technically a superset of the former. That should be a relatively easy transition, compared to, say, a data warehouse migration such as Teradata to SQL DW.
Kudos to White for being the first executive from a vendor company to go on record with this kind of estimate. It takes guts—and objective insight—to do so. We can assume this information has been thoroughly researched by White’s team. Their conclusion is probably as accurate as it gets, and certainly not exaggerated. Contrast this with billboards along Highway 101 in Silicon Valley: Most database vendors, from startups to established heavy-weights, dramatically downplay the effort needed to adopt their technology.
Let’s unpack this number and see what it means for your impending Teradata migration:
In our experience, less than 10% of the migration bill come from actually moving the contents of a database. Paradoxically, the real cost is in replatforming enterprise applications. We coined the term Database Migration Paradox for this phenomenon.
One area that balloons your tab are what we call dependent upgrades. For example, your ETL provider may very well offer a version of their software that is able to work with a cloud data warehouse. However, the version you have does not because it predates the cloud by years. Hence, you first must upgrade this entire component before making the move. Upgrading an ETL systems alone is typically a multi-million-dollar project. Worse yet, that type of modernization in the end leaves you with exactly the functionality you had before, with no business impact. This is modernization for the sake of keeping the lights on but does not provide any value to the business.
So, understand first what dependencies your data warehouse has. What software packages require upgrades. Which upgrades—if any—add true value, and which are inevitable but will not generate new revenue. Without this information your TCO calculation will be dangerously incomplete.
Mixing migration with modernization is perhaps the most frequently made mistake. A common misperception has it that once you “have the hood open” and make changes to the data warehouse, one might as well combine it with other modernization projects. These are often projects that the company has put off for a long time. Sadly, nothing derails migration projects quicker. By all accounts it’s the #1 Killer of not only database migrations but of entire cloud strategies. Many an enterprise had to reset their journey to the cloud halfway into the project because of this type of scope creep.
Once you understand the dependencies, it quickly becomes clear that burdening your migration with modernization will lead to an explosion in cost and delay or even kill the project altogether.
What I find stunning in our work with many Fortune 500 enterprises is that even seasoned practitioners drastically underestimate this effect. Mixing migration and modernization is a recipe for disaster and almost certain to fail. Therefore, move first, modernize later—always.
To the modern enterprise, data is everything. Databases are the foundation of every enterprise IT stack. Migrating from one stack to another may be a critical project but, more important, it is your opportunity to advance the enterprise—and your career. Moving from on-premises to the cloud is not a cost-savings measure, but a strategic move. Leading enterprises have long realized the cloud is critical to their competitive posture in the marketplace. The journey to the cloud is really the journey of the enterprise’s data to the cloud.
And analysts agree: over the next decade virtually the entire database market will be replatformed to the cloud. It’s not a question of “if” but “when”.
IT leaders must now prove their mettle and this migration event is also their opportunity to stand out and lead. Being the change agent who moves the data—and all the applications depending on it—to the cloud is going to be the most critical position in IT over the next decade.
Clearly, the stakes are high. Enterprises who fail to move their data to the cloud rapidly will be left in the dust. IT leaders need a better way to approach this transformation of the enterprise. In an agile cloud-based world, replatforming between data stacks will be the new normal, not a once-in-a-lifetime event.
At Datometry, we pioneered critical enterprise infrastructure that virtualizes data warehouse technology. Our platform lets existing applications, written for on-premises data warehouses, run natively on a modern cloud data warehouse.
Database Virtualization is the key to resolving the two chief obstacles we detailed above. Firstly, Datometry lets you replatform without the impossible prerequisite of disentangling the dependencies among your applications. So, instead of wasting time on readying the existing environment for a move, you will actually move. And, secondly, we have you migrate without the specter of having to modernize during that same critical window of time. That is, you will not mix migration with modernization. Rather, you can modernize at your own leisure, after the migration is complete.
Database Virtualization is an extremely powerful instrument to achieve agility, rapidly reap the benefits of the cloud, and ensure economic viability of the enterprise. Think “today’s applications on tomorrow’s technology”.
Circling back to the beginning and Microsoft’s Julia White. Replatforming SQL Server 2008 may be a $50bn effort but moving the entire database real estate to the cloud is going to be a heck of a lot more expensive. Extrapolating from White’s estimate, the on-premises data warehouse market alone will require an effort substantially north of $100bn to replatform. This creates distinct opportunities for enterprises that leverage modern technology and move fast. Remember: don’t rewrite, virtualize.