Company Profile

AES is a SaaS company providing hosted Internet payment and collections applications to the financial services sector in the US and Euro Zone.

Challenges

The application tested was Apollo’s Intelligent Debt Solutions ™, which has been selected for the prestigious Top 100 Collection Technology Products award by Collection Advisor Magazine for the 3rd consecutive year in 2007. “Decisioning Intelligence™, Apollo’s data-driven decisioning process that significantly improves debt recovery while reducing operational costs, continues to be one of the industry’s most revolutionary and effective technologies,” said Apollo Founder and President Christopher Imrey. “As the first web-based, real-time, rules-based, self-settlement system to incorporate bureau data, analytics and decisioning intelligence that match a debtor’s current ability to pay, the Apollo IDS system has attracted several marquee clients from the card issuing, debt-buying and recovery side of the business over the past year and we are now collecting over $1 billion online annually.”

Apollo’s application is highly dynamic and configurable, consisting of a distributed community of specialized software agents mapped to twelve logical tiers hosted on a variable collection of servers in the data center. Identifying and eliminating processing bottlenecks in the various deployment configurations is a critical aspect of our success. We provide our software as a service, so we live with the results of our work, day in and day out. It has to be reliable under a load.

Solution

When we were initially evaluating load-testing tools, we were still a pre-customer start-up with the usual budget and resource constraints. We needed a tool that we could learn and deploy within a week, not a month. At the same time, we needed to be able to perform complex testing scenarios with different populations of virtual users performing varying concurrent execution paths to ensure that no resource contention or deadlock situations existed. Neotys came through with a great tool and very effective technical support, even though we are an ocean apart.

The testing went well in the sense that we were able to ramp up quickly and get actionable test results in time to resolve issues.

Test Results and Added Value

As a direct result of NeoLoad stress testing, we found a documented issue in SQL Server that was undetectable under normal usage, but that crashed the server under a heavy load - typical unit and regression-testing procedures could not reveal the issue. Thanks to NeoLoad stress testing, we shipped clean and regularly clock 1000 TPS with no SQL Server stability problems at all.