Research shows that as many as one-third of businesses that use a data warehousing solution have migrated to the cloud (or built their data warehouse in the cloud it to begin with). Cloud-based data warehousing is on the rise, and this momentum is likely to continue in the coming months. Everybody wants in on the action; there are offerings from new vendors as well as more established ones.
What’s driving the shift away from the traditional on-premises data warehouse to private, public, and hybrid cloud deployments? There are a number of factors. As the cloud matures, concerns over security or lack of enterprise capability in the public cloud are fading. Meanwhile, a deeper understanding of the benefits of cloud deployment is emerging. Most businesses that have recently decided to migrate to the cloud or to build their data warehouse there from scratch are likely responding to one or more of these three major drivers:
1. Save money now (and later)
One of the most obvious advantages of cloud deployment is the up-front cost savings. Buying or upgrading hardware and software is both costly and time-consuming. Implementing a new infrastructure isn’t just about improving the bottom line (as critical as that obviously is). Businesses can invest the money saved on other strategic initiatives, such as better understanding customers, streamlining logistical processes, and entering new markets. Moving the data warehouse to the cloud can give your business a competitive advantage over businesses that keep theirs on premises. Those up-front savings can lead not only to longer-term savings, but to increased profits and stronger market share.
And it isn’t just the time and money; cloud deployment can also free up resources that would otherwise be dedicated to managing the new environment. The impact of this benefit depends on a number of factors:
- The level of service offered by the cloud provider. Some provide a bare-bones, infrastructure-as-a-service cloud offering, while others provide a full-service, set-it-and-forget-it package.
- The extent to which the new data warehouse is a standard or customized effort. (For the more robust projects, somebody will need to do all that work.)
- How much control the organization is prepared to give up in making the move to the cloud. The natural disdain that many IT shops have for “not-invented-here” can be easily transferred to “not-managed-here.”
But even the most cautious organization will experience a significant lightening of administrative load when it comes to managing their data warehouse in the cloud. As with the monetary savings, the real benefit comes from wisely investing the resources that have been freed up.
2. What happens in the cloud…
Unlike in Las Vegas, what happens in the cloud doesn’t necessarily stay in the cloud. But there is a growing consensus that perhaps more of it should. Once upon a time, the on-premises data warehouse collected most or all of its nicely structured, transactional data from on-premises operational systems. Today, not so much. Businesses are dealing with massive volumes of data from mobile apps, Internet of Things applications, and social platforms. That data all starts out in the cloud, and some businesses are beginning to ask why it ever needs to go anywhere else.
The Big Data phenomenon has not spelled the end of the data warehouse, as some early pundits inaccurately predicted. Other industry watchers took a more nuanced approach and predicted that what we are really witnessing is the death of the EDW, the traditional enterprise data warehouse. Even that may be overstating the case, but it’s important to note that Gartner recently touted the idea of the LDW, the logical data warehouse, as a sort of next-generation alternative to the EDW, if not its heir apparent. An LDW “uses repositories, virtualization, and distributed processes in combination” and is as much at home in the cloud — or to be perfectly frank, probably much more at home in the cloud — than on-premises. You might say we are being drawn into the cloud by the very nature of the data we’re dealing with, and the new challenges and opportunities that data carries with it.
3. Turn on a dime
When you install servers on premises, you have the infrastructure you have. It may be perfect for today, or even next week, but totally inadequate for what’s coming in 18 months. Or you may have far more capacity than you currently need or will need for a while, and that’s wonderful — except for the fact that you have sunk a lot of money and effort into this environment, which might have been more profitably spent elsewhere. Or you may even have infrastructure that’s too big most of the time — except for those few minutes each month when demand spikes.
Cloud-based deployments eliminate such problems through elasticity. Resources are dynamically reassigned as workloads change. So you can grow your environment, shrink your environment, or, for example, spin off a whole new environment just for a two-week promotion. As with the on-premises deployment, you have the infrastructure you have. But unlike on-prem, you might have something completely different tomorrow. Keeping pace with a rapidly changing business and data landscape is no easy task; having an environment that is ready to change in any direction at any time might just make all the difference.
It’s cloud time
As businesses continue to move their data warehousing environments to the cloud (and build net new ones there), we will get a clearer idea of whether cost and resource savings, having a native solution, or elasticity is the primary driver for the move. But that comparison may miss the point. For most businesses, it could well be that it is the interplay between these three major drivers that really matters.
Why deploy in the cloud? Maybe the best answer is, “All of the above.”
For more cloud deployment strategies, see The Ascendance Of Cloud In The Enterprise.