CASE STUDY
Financial Services Giant Saves $20M with Pepperdata Real-Time Cost Optimization
CASE STUDY
Financial Services Giant Saves $20M with Pepperdata Real-Time Cost Optimization
About the Client
A leading Fortune 100 multinational investment bank and financial services corporation headquartered in the United States.
Challenge
The company needed to gain control of runaway IT resource spend while maintaining optimal workload performance levels and meeting business-critical SLAs.
Solution
Pepperdata Capacity Optimizer’s real-time cost optimization recaptured wasted resources to lower costs and maximize the capacity of the customer’s existing infrastructure to control spend.
Results
The financial services giant achieved a 30 percent reduction in annual server costs, seeing approximately $20 million in savings over 7+ years.
FinServ Provider Facing Uncontrollable Big Data Costs
A leading Fortune 100 multinational investment bank and financial services corporation headquartered in the U.S. needed an intelligent solution to address the problem of runaway data center growth and constantly increasing spend for its compute infrastructure. The company provides banking services and products for companies, governments, and institutions in over a dozen countries and is a financial linchpin of the global economy.
The company deployed a Cloudera platform running Hadoop and Apache Spark with a global footprint of approximately 17+ petabytes of data. The platform ingested large amounts of data and used the Hadoop cluster as a data lake for storage. The company then managed, processed, and analyzed this data in the cluster using artificial intelligence (AI) and machine learning (ML).
This valuable data was the foundation for many of the company’s decision-making and business development processes, and was used to generate equity and fixed-income research as well as economic, geographic, and product-specific analyses.
The company initially engaged Pepperdata in 2017 when the I&O (Infrastructure and Operations) team was considering application cost optimization options to improve the efficiency of their analytics workloads and help meet strict SLAs. Concurrently, infrastructure cost management was becoming a priority: increased data ingest rates and data compute requirements were resulting in a 100 percent year-over-year growth rate—causing the company’s server capacity to double annually.
The I&O team understood that they were on an unsustainable path and sought out Pepperdata Capacity Optimizer.
“Pepperdata Capacity Optimizer helps optimize the efficiency of our infrastructure to reduce our spending on servers; we also now know what’s needed to achieve optimal performance for each workload.”
— Vice President, Big Data Solutions Engineering, Global Financial Services Customer
How Pepperdata Cost Optimization Automatically Saved the Company 30%
The selection of Pepperdata for real-time cost optimization was a natural choice because of its ability to track more than 400 application metrics every three seconds. By automatically identifying underutilized servers and redirecting resources to those workloads in real time, Pepperdata Capacity Optimizer reduced the company’s upward spend on server infrastructure by approximately 30 percent.
Without needing to make application code changes or apply recommendations, Capacity Optimizer automatically directed the system scheduler to allocate tasks to nodes where more work could be done—increasing the efficiency of the infrastructure and ensuring resources were properly used.
For this customer, the 30 percent improvement in server utilization and capacity enabled the company to reclaim and repurpose thousands of nodes, thus extracting maximum value from their hardware investment.
Pepperdata Capacity Optimizer also provided a unified solution for resource optimization that eliminated the need for manual tuning, helping the team focus on higher-level tasks and responsibilities. With Pepperdata, the I&O team’s focus is now laser sharp on managing their growing infrastructure and creating a more efficient Hadoop/Spark cluster environment.
The company considers Pepperdata to be an indispensable and integral part of the company’s large-scale data analytics platform as well as its long-term growth and management strategy.
