Performance Testing

Software systems often exhibit performance characteristics that differ wildly from the designers’ intentions. One commonly used method of determining these characteristics is performance testing, where individual code segments, software units, or whole systems are tested using automated testing tools, to measure performance behaviour.

Need

• Minimise risk of new applications or releases
• Minimise costs associated with fixing a poorly-performing service
• Delivery of new services to time and budget requirements
• Identify performance problems before the user is impacted
• Identify resources needed to run the service upon release and in future planning window

Challenges

• Increasingly complex application architectures
• Reduced time available for testing due to commercial time-to-market pressure
• Adoption of Agile Development techniques
• Obtaining meaningful results from unrepresentative test-beds
• Availability of quality performance-focussed testing staff
• Cost of permanently manned performance testing team or contractors

Our Service

Our staff are skilled at understanding how systems perform, rather than being generic testers, thereby delivering considerable added value. Capacitas offers a comprehensive performance testing service, comprising a suite of activities:

• Performance Risk Assessment
• Paper-based risk assessment of each release
• Identifies and determines risk of each change
• Proposes appropriate testing to cost-effectively mitigate risks


• Performance Baseline & Regression Testing
• Measures transaction service time and resources used for key transactions
• Used as a baseline for comparing against future and past releases
• Resource and transactional service time data can be used for modelling


• Load Testing
• Measures the scalability characteristics of system as transactional load is increased
• Identifies resource bottlenecks and transactional wait time
• Useful for validating queuing performance and capacity models


• Stress Testing
• Measures the scalability characteristics of system as mixed transactional load is varied
• Identifies resource bottlenecks
• Useful for validating queuing performance and capacity models


• Soak Testing
• Measures key resources of system over time under load
• Useful for identifying various resource leaks, e.g. memory, handles, etc.


• Results Analysis & Interpretation
• Scaling results to reflect how the live system will perform
• Determine risks to service based on detailed results analysis
• Identify and propose performance tuning that can be conducted
• Propose mitigation for any performance problems identified
• Provide seasoned analysis to enable effective go/no-go decisions to be made

Capacitas can provide performance testing staff as and when needed at short notice. Our staff can use existing performance testing toolsets, recommend tool purchases or rent performance testing tools on a per-use basis.

Features & Benefits

Delivering Results:

• Mitigates the majority of performance and capacity risks of an application or service
• Promotes successful rapid delivery of new services to the business
• Confidence in results due to performance-specific expertise
• Reduce risk of service availability issues

Flexible Resourcing:

• Multiple staff can be used as-and-when needed, enabling the customer to ramp up and down performance testing quickly and cost-effectively
• Staff only used when needed – no on-going charges compared to permanent staff or contractors
• Testing can be conducted on-site or off-site (e.g. over VPNs) if appropriate

Value-for-Money:

• Our team combines performance testers and experienced performance test managers, thereby creating an average daily rate comparable with contract staff
• Complete turn-key solution possible to performance testing using our services

Pragmatic Approach:

• Conduct a performance risk assessment to identify what does and does not need performance testing
• Able to define performance testing approach on a release-by-release basis

Performance-Focused Expertise:

• Ability to understand performance testing results and their implications
• Ability to model and extrapolate results from unrepresentative test-beds

Example Customers

• easyJet
• HP

Next Steps

• Review the Resources section on this page
• To email for further information click here
• Call our offices on +44 (0) 20 7841 9950 to arrange a meeting