How Combining IoT, In-Memory Processing, and Business Intelligence Makes Intel Faster

Fig3.png

When I looked through the recently released Intel IT’s 2015-2016 Annual Performance Review (APR), I was struck by the way Intel IT combined the Internet of Things, In-Memory Processing, and Business Intelligence to enable Intel to go faster.  There are many other ways that business intelligence techniques add value, and in discussing last year’s performance review, we focused on those approaches. In this post, I look at how business intelligence techniques implemented over the past year have increased velocity at Intel.

Intel IT has deployed Internet of Things (IoT) technologies into Intel’s wafer fabrication plants.  Just getting IoT originated data is not necessarily helpful without the ability to analyze what you need among the 5 billion sensor data points collected by Intel factories daily.   Combining analysis with sensor generated data enables Intel process engineers and managers to know when and where manufacturing equipment might be malfunctioning. They then can schedule preventive maintenance before a machine defect occurs and results in manufacturing delays.  Intel IT implemented an improved analysis system, enabling engineers to perform analysis in 30 seconds that used to take 4 hours or more.

Outside of the manufacturing area, improving analysis speed generates tremendous business value. Moving to an In-Memory processing system created value by speeding up a number of Intel’s internal business processes. We have improved the average response time for our top ten transactions by 47% and reducing key business application runtimes by 24%.   These are but a few of the accelerations we saw with this technology change.

IAH.jpg

In the final example of Business Intelligence changes in 2015 speeding up Intel, an internal Intel Data lake called the Integrated Analytics Hub generated even more advances. Creating a single place for marketing and sales data to be stored and providing self-service business intelligence tools have reduced the time to complete analysis and develop customer insights from months to days as often as low as 24 hours.

Intel IT's Business intelligence teams will continue to make improvements in 2016, and I look forward to seeing what the APR will cover next year.  For more information about these subjects, here are some IT@Intel blogs and white papers that cover the subjects just discussed: