This BeyeNETWORK article features Ron Powellís interview with Paul Moran, account executive at SAP. Paul and Ron talk about accelerating the data warehouse with in-memory technology.Paul, could you tell us what changes you have you seen in the past 20 years with regard to BI and data warehousing?Paul Moran:
A lot Ė as you have as well. I think, over time, we have seen such a change in how we deliver BI. Itís not always easy, but at the same time, the output is extraordinary. We have evolved from spreadsheets or analytic tools that are disparate and siloed across the organization. Now weíre seeing intelligent systems that are highly automated and highly connected Ė real-time information systems that can be trusted and relied on by business leaders to make a difference.When I look at innovation and what we see in BI and DW, in-memory has really started to impact both business intelligence and data warehousing. Could you talk about the impact of in-memory?Paul Moran:
The question of innovation here is unbelievable. In my mind, analytics has followed Mooreís law. Every 18 months, weíve seen leaps and bounds in progression to what is possible today with analytics and harnessing of information. There are so many things that are critical that we see happening. One is speed. Moreover, real time is a key part of this. Access to information as itís flowing through corporations and delivering that information in real time, I think itís a key driver and major innovator. However, itís just one small piece of the innovation that Iíve seen happening with my customers over the last couple of years.
Indeed query performance has improved markedly with in-memory solutions, but one of the most significant things is agility. We have seen rigid data warehousing architecture models designed and built that are very difficult to change as companies introduce new products or reorganize their internal employees. The agility that exists with aggregation and indexing done on the fly using in-memory technologies is so liberating to our organizations today.
Then I really think combining that with the data volumes and the capabilities like predictive analytics
that we have, we see our customers deploying data volumes on in-memory technology at levels that were just unheard of before Ė doing things with tens and hundreds of terabytes of data, highly agile, extremely performant with subsecond query response times using predictive analytics to spot trends, look at deviations from the mean, and alert people to take action where itís really vital to drive business performance. These things, Ron, just werenít possible two or three years ago, and there are applications that we see coming to life today that are just phenomenally promising. Well, youíve highlighted four of the biggest areas Ė speed, agility, data volumes and the whole predictive aspect that in-memory facilitates. From a customer perspective, could you give us an example of a customer that is using in-memory and all these capabilities?Paul Moran:
Absolutely. Iíll call the company a large Silicon Valley eCommerce-based company. What we have seen them envision and bring to life over the last year is an application that monitors this eCommerce economy that they operate. The data volumes are well over 50 terabytes of raw data. We are crunching hundreds of thousands of metrics in less than an hour every single day to measure and monitor all aspects of the business from category performance by product and by region around the globe, activity volumes and trends. Weíre applying multiple predictive models against those metrics to look for signals and deviations from the mean. In fact, on their own a small subset of these metrics would never have been noticed by the hundreds of analysts that are chartered to provide intelligence on whatís happening in the business. But when you start to cluster them and combine them and look at the trend over time, and when you can access that data in real time with real performance, then answer another question and click and click, all of a sudden the data tells a story about the business that smart data scientists and analysts can actually use to change business performance. Iíve seen it firsthand; and, again, itís the type of thing that we dreamed of two or three years ago and now see being deployed in production today. Itís quite exciting.Well, Paul, it really is merging the processes today from an operational perspective with the data warehouse so itís almost bringing it to the real-time world. Thank you, Paul, for providing such great insight into how SAP is accelerating the use of the data warehouse and BI with in-memory for customers, especially in eCommerce.
SOURCE: In-Memory Technology Accelerates Data Warehousing: A Q&A with Paul Moran of SAP
Recent articles by Ron Powell