Predictive Analytics For Resource Planning And Improving Post-Sales Services

BACKGROUND

A global building systems company needed to improve their preventive maintenance support and optimize the uptime of their HVAC systems. A data driven approach for their service and support business was needed to meet evolving customer expectations, and for cost efficiencies. Data is collected from multiple sensors on their HVAC systems in hundreds of buildings on a continuous basis. To improve availability and optimize energy consumption, all the data collected from the HVAC installations across the globe is centralized for monitoring and analysis purposes.

BUSINESS CHALLENGE

The services group needed to improve their understanding of the HVAC units performance, efficiency, and reliability, as well as real-time warehouse inventory management. Addressing the problem before it impacts the customer required deeper insights on leading indicators that might foreshadow a system malfunction. Deciding on the proper response and intervention when technical issues arise at the global monitoring center has proven to be difficult. Sending an experienced technician on-site is costly and should only be considered when an expert is required and solving the issue remotely is not possible. Knowing when to use their limited support resources allows them to focus on the productivity and efficiency of their services group. With a better understanding of problems upfront they can send their experts where they are needed the most, resulting in customer experience improvements at a much lower cost.

The Xcalar Solution: Xcalar Virtual Data Warehouse

The services group has grown in response to the need and opportunity to use data for efficient operations and increasing  revenues. Realizing that their existing legacy data warehouse would not be able to handle the increase in analytics workloads, the services group decided on Xcalar Virtual Data Warehouse to meet their evolving business requirements:
  • Xcalar runs on the cloud (Amazon EC2); a data center was not an option.
  • Handles time series analytics on growing datasets from multiple sources.
  • Xcalar’s Scale-out architecture provides the capacity to expand as needed by simply adding new nodes.
  • The ability to run across three availability zones offers better redundancy.
  • Xcalar’s True Data in PlaceTM reduced the size of raw storage space by over 10 times the requirement of their legacy system.
Xcalar’s field team helped this customer navigate the roadblocks of deploying on the Amazon cloud across availability zones. Extensive customer testing of the Xcalar solution across Amazon availability zones show  superior performance and resiliency, with minimal latency impact.

Time series data is captured and collected at regular intervals from the intelligent buildings,ranging from one minute to 15-minute intervals. The XML data from the sensors is written to S3 storage and then directly accessed by Xcalar. The optimization engineers, who analyze the data, are testing different settings that will improve performance, extend the life of equipment, or reduce energy costs. These optimization settings are used as references by the services group for setting alarm and notification parameters that are monitored by the operations center.

 The Xcalar Virtual Data Warehouse now captures and analyzes data from 30 sites covering hundreds of buildings, growing to over 700 sites covering thousands of buildings worldwide by 2018.