Global solar energy company based out of SF bay area, CA with 3.4 billion in market capitalization was reaching the cost limits of enterprise style licensing and performance limits of HA architecture for time series database. They were also anticipating business growth from 1000s of sites to 1,000,000s of sites as well as customer wanted to be able to predict weather conditions movements and energy flow to the grid using BigData analytics.
ValueLinked researched, prototyped, and developed a Cassandra framework and cluster for persisting 1,000,000 samples per minute. 1 billion/day. ValueLinked examined products, licensing models, vendors, and determined that the open source, additional tuning, tools, plus support would provide a highly scalable solution at a fixed margin over the hardware operational cost. Then used Hadoop and MapReduce for performing analytics and predictive reporting.