Business Drivers for Cloud vs. On-Premises Analytics

When it comes to your company’s data and analytics tools, how much does location matter? For years, business leaders have been hearing this debate about cloud versus on-premises data and analytics from their IT and security teams. I understand – at AnswerRocket, we went through the same debate.

In late 2013, we began developing AnswerRocket’s search-powered data analytics and visualization software. At that time, everybody was predicting the inevitable and fast move to the cloud for data storage. CIO Magazine said so. Forbes said so.

Great, we thought. Let’s build our solution for the cloud.

Except that the more we talked to executives of global companies, we realized this wasn’t happening as predicted. These companies had security concerns with cloud storage. Or they had expensive infrastructure they needed to leverage. Or their data was growing too large too quickly to move it to the cloud fast enough.

So we needed an on-premises/self-hosted strategy. Not instead of cloud – but in addition to.

We also needed really fast user response times, since we were aiming for a Google-like user experience. Our first thought was to look for on-premises columnar data stores like HP Vertica and ParAccel. But this would add significant cost and complexity to the whole product offering.

Fortunately, about this time, we became aware of the innovation happening with open source software like Hadoop. Hadoop was traditionally being used to process large batch jobs. But Apache Spark, Apache Drill, Impala, and others built upon this platform to make data more accessible in a fast and interactive way.

With these solutions, we were able to build AnswerRocket to access data to fit with a company’s existing data strategy and infrastructure – with a path to adapt as needed. This way, companies had flexibility with their data storage. If they needed the extra security and network bandwidth for on-premises storage, open source data frameworks gave them the ability to query multiple data sources.

Now, the move to the cloud for analytics does seem increasingly inevitable. The cloud offers lower cost, greater agility, and speed to market.

But how we get to the cloud varies by company. In a recent survey by IDG Services of IT and business decision makers about analytics storage over the next three years:

  • 32% expect to adopt a cloud-only approach
  • 39% plan to implement a hybrid approach

So what does this mean for you? If you have a vested interest in data analytics and business intelligence for your company:

  1. Look to the cloud first. The benefits of lower costs, greater agility, and speed to market provide real benefit and perhaps competitive advantage.
  2. Seek solutions to leverage your existing infrastructure. If your company has already invested heavily in big data infrastructure, look for solutions that let you quickly tap into your data. Ideally, from either the cloud or on-premises.
  3. Consider open source data frameworks. If your company is just getting started with big data, spend some time reviewing open source data frameworks. They may let you manage costs while providing the agility you need going forward.

The technology will continue to evolve in the coming years. To best harness your data, you need a solution that solves your current business challenges but also offers you the greatest flexibility in the long run.

Use this essential guide to selecting a natural language analytics tool.

Walter CC BY

Scroll to Top