Performance is one quality of software that can make or mar it. Poor performance can cause major losses. Hence, performance is an integral part of the various functions of Banks and Financial Institutions. Financial Institutions invest heavily in IT products and services for their better survival. Performance Engineering reduces TCO for IT and optimizes application performance.

Software Performance:

Software performance problems usually occur due to a fundamental misunderstanding of dealing with performance objectives. The idea is to tune for performance after getting the functionality right. Performance problems are complicated and require extensive code changes. Systematic planning and predicting the performance of the emerging software throughout the development process help in managing software performance properly.

Software Performance Engineering:

Software Performance Engineering is a systematic, quantitative, software-oriented approach to manage software performance proactively. Software Performance Engineering helps in recognizing problems early in development and corrects them.

The three measures of performance are Speed, Scalability and Stability, in which discrepancies may occur. Performance Engineering is all about solving a problem to achieve a desired and beneficial outcome.

The Cost of Performance Problems:

1. Ncreased hardware & development costs.

2. Cancelled projects.

3. Damaged customer relations.

4. Lost income.

5. Reduced competitiveness.

The Growth of Load Testing market:

Large development and implementation costs and the risk of losing of brand equity escalate the cost of application failure. According to Newport Group research, the hourly cost of downtime per million dollars of daily-generated online business revenue can range between $9,500 and $27,500,

Importance of Performance Engineering for Banks and Financial Institutions:

The services like loan, card, insurance, ERP and many more provided by banks and other financial institutions are subjected to improved Software Performance Engineering. Testing Banking Systems help to benchmark applications for users, predict the application's performance, fix problems in early stages, enhance the speed and response time of applications, prevent customer annoyance and loss, and get the most cost effective investment plan.

Testing Banking Applications require performance-engineering services such as Load and Stress test, End-to-end Transaction Testing, Architecture Benchmarking, Capacity Planning, Simulation/ Performance Prediction, Endurance Testing etc.

ReadyTestGo's Performance Engineering Services:

ReadyTestGo's Performance Engineering Services offer unbiased and unlimited testing of the software, besides offering anytime, over the net and lab testing services. Our consultants possess significant expertise in web servers, databases, networking and more. Our 20,000 sq.ft. Lab at Chennai, co-hosted at San Jose and Singapore is equipped with world-class hardware, software and testing tools for diagnosing and solving problems.

Process Overview:

Any performance engineering process should ensure repeatability, consistent delivery, complete coverage and a strong feedback mechanism to leverage knowledge. Figure 1 illustrates our approach to performance engineering:

Approach to Performance engineering:

To get proper utilization of Performance Engineering, test for common performance bottlenecks and transactions must be conducted. Creation of re-usable test scripts and tracking defects to closure are some way to make it happen.

Server Testing:

Server Testing needs to monitor Web, Application and Database Servers very well.

Performance Test Runs:

Performance Test Runs include:

Baseline and Diagnostic Test Runs, and Enhancement Test.

Monitoring:

Server Side and Client Side Monitoring are crucial for finer performance measure.

Reporting:

Testing Banking Systems is followed by the submission of a test report containing ReadyTestGo's primary client and server side observations, its analysis of the correlated client and server side data and it would attempt to identify any and all bottlenecks in the application.

The graph in the following figure depicts a typical client and server side co-relation showing that the web site under test failed to scale beyond a user load of around 55 concurrent users due to high CPU utilization. It also shows that user processes were the primary contributors to the high CPU utilization observed and that CPU utilization by the Kernel and the I/O subsystem was nominal.