To reduce power theft and improve customer satisfaction, an electricity company set out to minimize planned IT infrastructure downtime, both planned and unplanned.
The electricity company deployed SAP HANA on IBM Power Systems, enabling super-fast queries that highlight hundreds of insights to help investigate incidences of suspected theft in a timely manner.
Enablesaccurate targeting of power-supply inspections
Slashestime to generate budget reports
Business challenge story
Faster business decisions: crucial for growth
An electricity company that manages more than ten million accounts realized that analyzing generation, distribution, consumption and billing data could help it manage its services to target fraud and optimize customer satisfaction. For example, why would one house show low or zero usage compared with another in the same street, or why would a business use a fraction of the power expected?
However, with a billing database of more than 120 TB, growing at 2 TB a month, data analysis presented a significant challenge. Naturally, the company wanted to provide the inspection teams with suspected cases of fraud suitable for investigation. To increase prediction precision, known as ‘assertiveness,’ the company wanted both to analyze large data sets and to look deep into the data, and with its huge data volumes this was proving next to impossible.
In 2017, the company migrated its core SAP Business Warehouse (SAP BW) to the SAP HANA database. The operational database shows the total energy provided by each delivery point, such as a substation or termination. The billing database holds details of how much each consumer was billed, and the total energy consumption. By cross-checking the delivery point energy with the energy consumed according to the billing system, the company could identify discrepancies, correct them and deliver a better service.
A spokesperson for the electricity company explains, “Comparing the energy at tens of thousands of delivery points with millions of consumer accounts, each with hundreds of meter readings, would place significant workload on our SAP HANA infrastructure. That could potentially hinder our fraud-detection efforts and lower customer satisfaction, and we wanted to find a solution capable of scaling up to this challenge.”
“ Using the improved accuracy of the fraud scoring solution based on SAP HANA running on IBM Power Systems, we have made considerable efficiency savings and optimized customer service. ”
— Spokesperson, Electricity Company
Watching core performance soar
The electricity company reviewed its infrastructure options for SAP HANA, and selected an IBM Power Systems E870 server, based on its scalability, attractive return on investment, and tremendous system performance. The local tax law favors capital investment, and the company purchased and deployed the solution on-premises. As workload grows, the server offers simple scale-up options, and the addition of further processors and main memory will enable greater capacity without disruption to existing software, networking and backup processes.
Using IBM PowerVM® virtualization, the server is divided into multiple virtual machines (logical partitions), each of which is optimized for its workload. The core SAP HANA database runs in a virtual machine running the SUSE Linux Enterprise Server for SAP Applications operating system, alongside two virtual machines running the IBM AIX® operating system, supporting SAP NetWeaver, and two virtual I/O servers.
This configuration allows the company to allocate memory, processors and priority to its workloads, ensuring that it can run deep analytics at the same time as completing its business-as-usual operational work. For example, the IBM Power Systems E870 server delivers significant performance improvements, slashing the time taken to produce budget reports, forecasts and other reports.
“We selected IBM Power Systems for its superb availability,” comments the spokesperson. “For example, we can reduce downtime even during planned maintenance windows.
“In addition, IBM Power Systems runs processes faster using fewer resources. For example, reports are available much faster, driving faster decision-making.”
With this combination of the SAP HANA in-memory database and IBM Power Systems, the company could turn its full attention to scouring its databases in order to deliver better service.
“ We selected IBM Power Systems for its superb availability. For example, we can reduce downtime even during planned maintenance windows. ”
— Spokesperson, Electricity Company
Creating a fairer society
Since its implementation of SAP HANA on IBM Power Systems, the company has driven up prediction assertiveness, by looking more closely at consumption data and integrating the analysis with data on power delivery points. For analysis, the company relies on SAP BW powered by SAP HANA, exploiting the advantages of the in-memory technology for speed and the IBM Power Systems E870 server for performance.
SAP HANA on IBM Power Systems provides the ability to include all meter and delivery data-points over a longer period, which means that the company is less likely to miss an unexplained dip in billing, or to overlook a mismatch between consumption and delivery data. This enables the company to deliver significantly improved customer service.
The spokesperson concludes, “Using the improved accuracy of the fraud scoring solution based on SAP HANA running on IBM Power Systems, we have made considerable efficiency savings and optimized customer service.”
This electricity company and its power generation, fuel oil, biomass, hydroelectric, solar and related subsidiaries serve more than 10 million customers, and employ more than 10,000 people.
- IBM Power Systems running Linux - SUSE
- IBM Power Systems running AIX
- SAP HANA
- SAP NetWeaver Business Warehouse
- Software - PowerVM (not HPC, nor VA Linux)
Take the Next Step
To learn more about IBM Power Systems for SAP HANA, please contact your IBM representative or IBM Business Partner, or visit: ibm.com/power/saphana
View more client stories or learn more about IBM Systems Hardware