Software Performance Engineering

SPE is designed to ensure predictable performance in systems, by specifying and analyzing quantitative behavior at every tier. Our Architects and Engineers, carefully consider performance requirements, the actual design, and the environment in which the system will be running. Analysis is based on modeling techniques and design tools.

SPE ensures that software systems adhere to their performance requirements, aiding in an organization’s overall business success. ITg’s SPE expertise, offers organizations the following benefits:

• Saves significant money towards the costs of software projects by reducing inflated equipment budgets, staffing overruns, and missed project deadlines
• Reduces the probability of users rejecting new and subsequent versions of software systems
• Eliminates the need for disruptive SWAT teams to combat poorly performing software systems
• Provides speed and reliability that encourage customers to utilize self servicing systems, e.g., Human Resource portals, online ordering, restaurant reservations, and online banking
• Enables the delivery of large amounts of complex information, in real time, which is required for data mining and cross selling opportunities, group collaboration, rich media, etc.
• Allows hand held mobile devices to run full featured software applications and display the latest web content, with the speed and reliability of a desktop computer

Software Performance Architecture

Clients deserve and demand a high speed, world-class technological experience. ITg utilizes mathematical modeling to iterate through multiple design options, to ensure the delivery of software that will satisfy our client’s performance requirements. These models:

• Map out the sequence of key business processes
• Assign arrival rates and resource utilization of key business processes
• Calculate if application service times are acceptable
• Provide baseline models that are refined to fit the construction and deployment of the software system

A significant return on investment is realized, as a result of employing these models during the design phase; architecture rework costs less per improvement than recoding an application or making unnecessary investments in infrastructure resources.

Software Performance Development

We train and coach developers on the use of mathematical models, to ensure software routines adhere to performance-resource allocations and service-time requirements.
Measuring resource utilization during code construction enables software developers to write code that is both functional, and high performing. The performance data that is collected against the actual source code, improves the initial performance models that were created during the architectural stage. Development performance models continue to build confidence in the ability to accurately predict the performance scalability of the final software system. The investment in, and use of, performance models at this stage, reduces the need for extensive optimization after code integration, which in turn assures a more efficient and cost-effective software system.

Capacity Planning

ITg properly sizes servers, storage, and network resources, by leveraging a combination of mathematical modeling, load testing and analysis. These models are employed during the system design, development and testing phases, as well as throughout the ongoing production process. ITg utilizes this systematic approach to predict and maintain the most cost-effective and optimally performing infrastructure.

• The performance models created during design, improve the accuracy of hardware sizing estimates by reducing over-and-under provisioning of resources, budget bloat, and disruptions.
• Continued use of the performance models during development and testing, further refine the resource allocations prior to production deployment, improving the success of the software system implementation.
• Employing the models once the software system is in production, provide an engineering method to monitor changes in workload and system behavior. The models forecast how such changes will impact the hardware, storage, and network-provisioning strategy that is used for a software system. In addition, these models provide a feedback loop into the initial models that were created during design, and provide insights into the changes that are necessary to create future models in the design phase.

Software Performance Testing

The purpose of software performance testing is to find out how the software system works under varying stress loads and activity. Using a combination of multiple tests, we are able to determine the parts of the program that reduce its effectiveness. Load and stress testing, enables integrated software systems to pass through a final phase of performance validation, prior to deployment. Through this process, ITg offers its clients the ability to approximate production workload, against a sample of production hardware, to execute the final pre-production coding and system optimization.

ITg offers a load and stress testing solution that provides an easy-to-use and cost-effective way to load test the performance of web applications and websites. Unlike most testing solutions on the market, it does not require any software or hardware purchases and provides results faster and less expensively than traditional performance testing.