Combining Big Data and predictive modelling

In its mission to provide data center owners and operators with actionable intelligence, Romonet has brought together its powerful and patented data center cost and performance predictive modeling engine with a noSQL database architecture and Big Data analytics; allowing faster and more insightful decision-making information for users.

Romonet has also updated Portal’s user interface, providing a targeted and easy to understand view for site managers and staff along with users that have responsibility across a portfolio of many sites.

“Traditional monitoring and trend analysis tools are reactive; giving teams insight either at the point of an alarm or over a historical view of events that have already happened,” said Zahl Limbuwala, CEO of Romonet. “By feeding a predictive model with the actual monitored data from a site, Romonet Portal gives operational teams visibility and warning of future operational issues. For instance, it can actually provide a warning when device performance starts to degrade or exhibit small but important early warning signs of failure. In fact, combining these Big Data techniques with Predictive Modeling allows Romonet Portal to spot and raise the alarm on issues such as filters in air handlers that need changing; or blanking plates being removed, causing increased air leakage.”

While developing this new combined capability Romonet has addressed another hidden issue within the data center, that of monitored data quality. Liam Newcombe, Romonet’s CTO, said: “It was a surprise to us initially that on average 35% of the data coming out of the Building Management System and Environmental Monitoring System platforms out there is junk data. We now clean, verify and store all the metered data in our noSQL database; as most monitoring systems are not capable of storing this type of time-series dataset over many years while still being responsive to queries and lookups.”

Romonet’s Big Data architecture means predictions can be made faster, and with more frequency, than ever before. Performance issues can therefore be identified and solved before they have a chance to disrupt availability of service to customers.

Romonet Portal’s Big Data capabilities come from using a scalable non-relational database, as used by Internet giants such as Facebook and Google. In conjunction with Romonet’s proprietary and award-winning predictive modeling engine, this allows users to forecast, plan and measure business performance with greater insight, accuracy and agility than previously possible. The new Big Data architecture means analysis that would previously have taken hours or days can be performed in minutes; the database is capable of sorting through millions of data points in minimal time, and removes the requirements of archiving, even with massive datasets; all the customer’s data is there to access, all of the time.