Enterprise Big Data on a Shoestring

We live in a world where people are used to accessing and updating information in real time and with user experiences that are fast and fluid.  The technology now exists to, economically, enable this kind of experience for enterprise systems like Integrated Business Planning (IBP) tools even where data volumes are huge.

Google serves over 3.5 billion searches per day with millisecond response times.  When a popular news site posts an article it only takes seconds for Google to pick it up and serve it in search results.  Google’s domination in the search space is due in part to its instantaneous performance and in many ways the adoption of enterprise IBP tools is equally dependent on real time performance.  It is likely that users would be delighted to move away from their spreadsheets if the corporate tools had similarly near instantaneous response times.

The challenge is real: consider a simple CPG company with 500 products and 400 customers and on average 50,000 customer-product combinations.  For a weekly forecast horizon with four years of history and two forecast, weekly snapshots and 15 income statement levels in addition to volume, there are 250 million rows of data in the live forecast and 13 billion rows of snapshot data to update and query in near real time.

Tools have existed for a few years to help in this space like SAP Hana, however recently there has been a growth in the Big Data space that enables real time performance to be deployed very economically.

A good example of this is MemSQL, a database created by a Silicon Valley start up that is sharded (automatically split across multiple servers) and in-memory (like SAP Hana).  MemSQL is already deployed by large Tier 1 companies like Kellogg’s, Uber and Pinterest.  Blueshift, an IBP solution provider headquartered in Australia, has achieved 100-fold performance gains by leveraging MemSQL in combination with Kafka which is the technology developed and used by LinkedIn.

Whether it is making wide ranging promotional strategy changes, updating pricing or baseline information, performing adjustments in SOP review meetings, or simply conducting basic scenario modelling users expect and even require real time performance.  The good news is that there are now solutions, such as Blueshifts ONE Solution, that deliver which are both extremely economical and scalable to thousands of products and customers and beyond.

Justin Stafford, CEO and Co-founder