A smart meter processing company was having trouble with its solution. Although based on Big Data, it was not quite fitting the business purpose. The solution was rigid: it could give fast answers to questions that were initially thought of, but for others, it was either excruciatingly slow or could not give the answer at all.
Their customer, a power utility, was in its turn unhappy. The processing company has initially conducted a wide-ranging study proving the importance of further work on smart meters. However, now they found out that they could not quite capitalize on the data: their data analysts required much more flexibility in data handling in order to come up with enticing offers for their clients. The company was losing to the competition and saw its attrition grow.
Elephant Scale was called in to re-evaluate the architecture. Based on their extensive experience, Elephant Scale consultants suggested a balanced transition into the Big Data technologies world. First, all the data was to be collected in Hadoop, assuring business continuity and long-range flexible data storage. Next, as soon as possible this data was made available to the data analysts for experimentation and feedback. This led to streamlining the solution in the critical junctures, allowing a practical balance of flexibility and performance. Having the authors of the recently published book “HBase Design Patterns” on the team was also helpful. There is always time for such tools as Hue, Hive, Tez, HBase, and even Solr, but the trick is to know how and when to introduce them for an efficient overall solution.
Now, with the help of the solution developed by Elephant Scale, the power utility sends daily targeted emails to their customers and optimizes the profit margins by offering the best fitting power plans to consumers. These are only the first two projects out of about a dozen that the power utility envisions. In the coming two years they expect to integrate new Big Data technologies into their infrastructure, providing good business value and multiple competitive advantages.