Performance Testing of Key Student Applications Migration to Azure Cloud by UWE Bristol

Challenges:

1. Identifying High-Risk Applications: UWE needed to prioritise applications for performance testing based on their importance and susceptibility to performance degradation in the cloud environment.
2. On-Prem vs. Cloud Performance Comparison: A comprehensive on-premises performance baseline had to be established for critical applications, so the university could determine if the Azure Cloud environment introduced any performance variances.
3. Meeting End-User Expectations: Ensuring that the performance of key student-facing and administrative applications met or exceeded user expectations after migration was essential, particularly for high-traffic periods like registration or assessment submission.

Performance Testing Approach:

1. Risk Assessment and Prioritisation: A detailed assessment was carried out to identify high-risk applications requiring performance testing. Applications with high transaction volumes, such as GO (URL Shortener), Course Search and Module Search were flagged as critical.
2. On-Prem Baseline Testing: Before migrating to the cloud, a series of performance tests were conducted on the on-premises environment. These tests established a performance baseline for response times, system throughput, and resource utilisation (CPU, memory, and network I/O) under normal, peak, and stress loads. This data served as the benchmark for comparing performance in the Azure Cloud.
3. Cloud-Based Performance Testing:

   - Load Simulation on Azure: JMeter was used to simulate real-world usage by generating concurrent user requests for each key application in the Azure Cloud environment. Test scenarios included critical student workflows such as GO (URL Shortener), Course Search and Module Search.

   - Peak and Stress Testing: The team conducted peak and stress tests to evaluate how well the cloud infrastructure could handle heavy loads. UWE’s applications were stress-tested beyond expected peak loads to ensure scalability and resilience in Azure.

4. Cloud Resource Monitoring: Using Azure’s native monitoring tools, performance metrics such as response times, CPU, memory usage, and auto-scaling behaviours were collected during the tests. This helped identify potential bottlenecks or areas requiring optimisation in the new cloud environment.

Results:

- Performance Consistency: The comparison between the on-premises and Azure environments showed that the key applications met or exceeded performance expectations in the cloud. Response times and throughput were either comparable or improved in the Azure environment, even during high-traffic periods.

- Scalability: The performance testing validated Azure’s ability to scale effectively during peak loads, ensuring that the applications could handle increased user demand during critical academic events.

- Stress Test Findings: The cloud environment proved resilient during stress testing, confirming that UWE’s applications could handle higher-than-expected loads without performance degradation.

Outcome:

UWE Bristol successfully migrated its key applications to the Azure Cloud, supported by robust performance testing. The university gained valuable insights into the performance differences between their on-premises environment and the Azure Cloud, with the testing ensuring that end-user expectations were met. The established performance baseline and cloud testing framework have set the foundation for UWE to scale its services confidently in the cloud, while maintaining an optimal user experience for students and staff.

This case study highlights UWE's commitment to leveraging cloud technologies while safeguarding application performance, enabling the university to continue delivering high-quality services to its academic community.