Performance Test Approach For SAP SaaS
Performance Test Approach For SAP SaaS
Performance test is conducted by simulating the expect workload on the production system or a system
representative of the production size. There are a few types of performance test and conduct of these
tests are varied according to their objective. The following types of performance tests are commonly
performed for SAP systems.
Load Test: System will be tested with the expected load. This load is the expected concurrent number
of users performing a specific number of transactions within the set duration and batch job execution.
Stress Test: Load tests are repeated and the simulated load on the system is increased with each
iteration until the system is not able to handle the load. This kind of test is done to determine the
upper limits of system’s capacity.
Soak Test: Load test is executed for a much longer duration to determine if the system can sustain
under a continuous heavy load.
There are two types of workloads on an SAP system and the conduct of performance test for the two
workloads are different.
The figure above illustrates the end-to-end response time which consists of the following:
Client response time: Time taken for the user’s machine to process the response and render the
page.
Network response time: Round trip time for data to be transmitted from client to backend system.
This is dependent on various network performance metrics like throughput and latency.
System response time: Time take for the system to process the request and respond.
Since the objective of performance test is to measure the performance of the system, the response time
is the system response time.
SaaS applications are designed for much larger user base and a client’s workload is much smaller
compared to the workload the system is sized for. As such, the conduct of a load test for SaaS
application does not provide benefits.
To measure the system response, the performance test tools are installed in the same LAN or data
centre as the SAP system. For SaaS application, this is not possible, and the response time measured
will include the network time. Hence, it will be a challenge to determine if the performance bottleneck
due to network or system.
SaaS performance is typically the responsibility of the service provider and the performance statistics
should be provided by them.
Instead of a performance test, Performance Benchmark Test was conducted. The purpose of the
Performance Benchmark Test is to assess the end to end response time for varying network latency. The
outcome of this test will provide a reference for the end to end response time that users will experience
for the different latency. Performance Benchmark Test will be conducted using a WAN emulator to
simulate the network latency.
Performance test
tool.pptx
Data Requirements
Test data will be required for online load test. For example:
Test user ID.
Master data.
Transactional data.
Support Personal
Support from the following teams will be required for the conduct of performance test.
Team Activity
Infra/ Basis team Create infra required for performance test.
Monitor and tune system performance.
Backup and restore systems
Functional team Advice on testing scenarios and configure system.
Data team Test data creation
Planning
The first step for testing is to define the high-level performance test design and timeline. The output of
this process is this document.
Design
The design phase specifies the test scenarios and parameters. The design should include scenario
definition, user description/activity, transaction description/volume, step-by-step transaction semantics
and specification of required master/transaction data. The outcome of this stage will be the Performance
Test Scripts.
Preparation
This third phase includes setting up the test infrastructure and scenario scripting for the test tool. The
preparation phase also includes the preparation for mock-up data such as test IDs, employee and payroll
data. It is advisable to assign a functional individual who understands the business processes, to work with
the team for the duration of data preparation.
Execution/Analysis/Tuning
The fourth phase starts with the testing of the scenarios (Execution). The data collected by various
monitors during the run is then used to analyse (Analyse) system performance and develop/implement
suggestions for system improvement/tuning (Tuning). The scenario will be iteratively re-run for the Load
test to optimize the on-premise systems.
Report
In the last phase, a report will document the completion of all stress test scenarios and cycles. It will
summarise the test results from each scenario and cycle and validate them against the performance
requirements set. The report will also specify any recommendations.
Entry Criteria
S/No. Criteria
1 Functionalities to be tested has passed SIT and are working in the performance test
landscape.
2 Test tools required for performance testing has been procured and setup.
3 Test Scenario and Performance Test Scripts signed off.
4 Test IDs and Data has been created.
5 Test scripts created and tested.
Exit Criteria
S/No. Criteria
1 All test scenarios have been executed and documented.
2 All identified performance problems have been recorded with a mutually agreed plan to
resolve the problem.
3 Performance Test Closure Report signed off.