Cognitive computing

Five starting points for building a cognitive business

Share this post:

There’s no way around it. As the pace of innovation and change accelerates, IT leaders will be tasked increasingly with ensuring competitive advantage. And that responsibility means being able to consistently derive insights from oceans of unstructured data to make real-time decisions.

To get to the point where your organization can act at the speed of thought and seize the opportunities of the digital, connected economy, you’re going to need cognitive tools supported by infrastructure intentionally designed for data and complex analytics.

But how do you get started? Here you’ll find my answers to five questions I’m asked frequently about designing for an IT infrastructure for cognitive workloads.

  1. Where does the move to cognitive begin?

It all begins with data. Cognitive solutions feed on data to sense, learn and adapt. You need to rapidly capture the largest volume and variety of data to develop cognitive insights faster than the competition and provide a better customer experience.

Storage that delivers high performance in a consistent manner is essential for cognitive analytics. Yet much of the installed infrastructure found today was born in the bygone era of the disk-based storage array. Flash is a great way to bridge that generational gap. A well-designed all-flash array used as a storage tier can sustain the performance level needed to feed a modern, in-memory database—using less server memory at lower cost.

  1. What role does software-defined infrastructure (SDI) play in building an IT infrastructure for cognitive workloads?

Cognitive analytics is about using all available data to get deeper insights than ever before. To truly become a cognitive business, most organizations need to have an awareness of social data. Sensors and meters are creating valuable data as well. But the traditional structured database just isn’t sufficient to process it all.

Enter SDI, which can help CIOs reduce cost while still achieving the performance needed to become a cognitive enterprise. You can use it to ingest, store and retrieve data of diverse types from many different inputs at high speed, and scale out your infrastructure without negatively affecting performance. Along with a workload management solution, SDI also enables you to run analytics at the most optimal location – in many cases, next to the data.

  1. How can I ensure performance to achieve real-time insights?

It takes a good match between compute and storage to provide both the fastest data retrieval and the most powerful processing available as needed to support cognitive analytics. Use servers, flash storage and workload management specifically designed for cognitive solutions. Look for innovative hardware accelerators and low-latency storage that can rapidly open up new business opportunities for you by cutting data analysis from hours to milliseconds.

  1. What’s likely to be the biggest obstacle in my data center?

Your biggest obstacle may well be the interface between compute and storage. Much of what prevents the rapid ingestion of data and hinders performance in today’s data center is not just the hardware pipe but also the software stack.

The software layers that accumulate in data centers over the years can slow data access in the same way that residue in old plumbing can make a spigot incapable of supplying a steady stream of water. But by bypassing layers of old software, you can be assured of the ability to streamline data into memory for the rapid and efficient analysis essential in the cognitive era. I often recommend CIOs learn more about our coherent accelerator processor interface (CAPI), which provides a very direct connection between the storage and the server processor.

  1. What does my organization need to do next?

Find out where your organization stands today and how you can get started by accessing IBM’s new cognitive assessment tool. This interactive guide helps you figure out the options that work best for your company.

For example, where will you locate your infrastructure for cognitive workloads? In the cloud? On premises? In a hybrid environment? Using the tool can give you and your colleagues a better idea of how that infrastructure can pay big dividends for your enterprise.

Click here to read the Spanish translation.

IBM Fellow, CTO and Chief Architect, Flash Systems

More stories

5 top takeaways from Think 2018

The first Think conference is in the rear-view mirror. There, visionary speakers and experts shared their solutions for some of the world’s most daunting challenges, and IBM revealed some of our big bets on infrastructure innovation in the coming years. Now that I’ve had some time to reflect on the conference, I’ve assembled my five […]

Continue reading

Taking the AI training wheels off: moving from PoC to production

In helping dozens of organizations build on-premises AI initiatives, we have seen three fundamental stages organizations go through on their journey to enterprise-scale AI. First, individual data scientists experiment on proof of concept projects which may be promising. These PoCs then often hit knowledge, data management and infrastructure performance obstacles that keep them from proceeding […]

Continue reading

Announcing PowerAI Enterprise: Bringing data science into production

Today I am tremendously excited to announce IBM Cognitive Systems’ newest applied artificial intelligence (AI) offering, PowerAI Enterprise. This new platform extends all the capability we have been packing into our distribution of deep learning and machine learning frameworks, PowerAI, by adding tools which span the entire model development workflow. With these capabilities, customers are […]

Continue reading