Performance, Giving Banks More Trading Time
"Using In-Memory OLTP in SQL Server 2014, we realized a 136 percent performance gain for one of our most critical systems. That makes us confident we’ll be able to support our customers‘ increasing volumes.“
Karl Catling, Director of Adaptiv Labs, SunGard
Financial software provider SunGard wanted to improve the performance of its credit risk application in response to customer demands. The firm also wanted to speed up one of its key system processes. To address these needs, SunGard tested In-Memory OLTP n SQL Server 2014. Lab tests showed a 136 percent performance increase in a critical system process. As a result, the company can support higher customer trading volumes and give bank employees more time to trade each day. SunGard also anticipates cost savings from using existing hardware.
This case study is for informational purposes only.MICROSOFT MAKES NO WARRANTIES, EXPRESS OR IMPLIED, IN THIS SUMMARY.
Document published April 2014
Business Needs
Headquartered in Wayne, Pennsylvania, SunGard is a provider of business applications and other IT solutions for banks and financial institutions around the world.
The SunGard Adaptiv suite helps banks and financial institutions manage risk across the enterprise. One of the suite’s key products is Adaptiv Credit Risk, a financial risk management system used to manage counterparty credit exposure. Hosted in the SunGard data centers and based on Microsoft SQL Server 2012 software, Adaptiv Credit Risk gives customers the ability to perform real-time credit inquiries 24 hours a day.
SunGard must comply with customers’ strict service level agreements (SLAs) for high availability and rapid response. “We’re tightly bound by SLAs to deliver fast response times during the trading day and for any overnight processes,” says Karl Catling, Director of Adaptiv Labs,SunGard. Meeting those requirements had become more challenging because customers are experiencing rapidly increasing trading volumes. “Some are anticipating a 30–50 percent trade increase,” says Catling. “That’s something we need to be ready for, and so we wanted to improve the performance of the database to support our customers’ needs.”
The company also wanted to improve application performance in an effort to speed up the End of Day process, which prepares the system for the next day’s trading activity. “This core process needs to be completed before traders arrive at work each morning,” says Catling. “We wanted to make it faster, because many of our clients operate globally and don’t have a significant overnight window.”
Solution
Because the Adaptiv suite had used Microsoft SQL Server as its OLTP database engine since 2005, SunGard was eager to see what features the latest version of the software, SQL Server 2014 Enterprise, would offer. One of those features is In-Memory OLTP, which uses main memory optimization and no-locking/no-latching concurrency control to eliminate bottlenecks. “During our key processes, we have many servers running workloads at the same time, all trying to access the same data. We knew that the In-memory OLTP feature could help with these issues,” says Catling. The in-memory capabilities in SQL Server 2014 also include natively compiled stored procedures, which can help enhance processing performance.
Early in 2014, SunGard began a series of benchmark tests to see if SQL Server 2014 Enterprise improved the performance of Adaptiv Credit Risk. During the tests, the company focused on a critical part of the End of Day process,specifically cycling through each customer’s financial portfolio and preparing it for the next trading day. “This process generates one of the highest workloads for the database server and can cause a lot of contention,” says Catling. “It also runs a high volume of SQL batches per second. It’s one of the biggest pieces of work that needs to be done for the End of Day process.”
For the benchmark tests, SunGard used 15 HP ProLiant DL380 G7 server computers and two database servers, one running SQL Server 2014 with In-Memory OLTP and one without that feature implemented. In each test run, 34,590 customer portfolios were put through the process. “We ran the servers hard during the tests, because we wanted to see what performance would be like in the most challenging circumstances,” says Catling. Satisfied with the test results, SunGard decided to implement SQL Server 2014. The firm planto make the Adaptiv Credit Risk system available on SQL Server 2014 by mid-2014.
Benefits
Using In-Memory OLTP in SQL Server 2014, SunGard realized a more than two-fold performance increase during lab tests. Once implemented fully, the solution will give banks more trading time and reduce operating costs.
Twice the Performance Power
During testing, SunGard saw healthy performance gains on the database running In-Memory OLTP in SQL Server 2014. “Using In-Memory OLTP in SQL Server 2014, we realized a 136 percent performance gain for one of our most critical systems,” says Catling. “That makes us confident we’ll be able to support our customers’ increasing volumes. And we think we can improve the system’s performance even more in the future, by adding more in-memory tables.”
Tests also indicated that twice as many SQL batches could be processed per second using the In-Memory OLTP feature in SQL Server 2014. “We went from a little more than 1,500 batches per second to more than 3,000,” says Catling. “That kind of performance will definitely help us complete the End of Day process faster.”
More Trading Time for Banks
With the ability to support higher volumes of data and complete critical processes faster, SunGard will be able to give Adaptiv Credit Risk customers more time to trade. “Thanks to faster processing as a result of using In-Memory OLTP in SQL Server 2014, we’ll be able to push data through faster. This guarantees that processing and reporting are completed by the time bank employees are at their desks,” says Catling. “So they will have more time to trade and produce analytical reports. Where some of our customers operate in multiple locations, we don’t have the luxury of a significant overnight window. We are under real pressure to meet the opening of the Asian markets. And with the reduced contention and lack of locking we’ll get from using the in-memory tables in SQL Server 2014, we will be able to give those customers real-time access to data.”
Reduced Operating Costs
SunGard also anticipates cost savings once it deploys SQL Server 2014 as the database engine for its full Adaptiv product suite. “We expect to see some reduced CPU activity once the solution is rolled out, and, as a result, some of our servers might not need to be upgraded,” says Catling. “Because we host our applications on more than 1,000 servers, we could significantly reduce costs without updating the hardware we already have in place.”
SunGard also expects further performance gains for its applications. “The concepts and new features we implemented could be applied to other activities,” says Catling. “This would make the end-to-end process considerably shorter. We are very confident that deploying SQL Server 2014 will introduce further significant benefits to our entire product line.”
This case study is for informational purposes only.MICROSOFT MAKES NO WARRANTIES, EXPRESS OR IMPLIED, IN THIS SUMMARY.
Document published April 2014