How Load Impact’s API testing works
The challenges to API performance
- APIs allow for multi-use, one-to-many integrations with internal as well as external systems. This places great demands on the functionality and performance of APIs.
- Websites and apps provide access for human consumption; APIs provide access for system (i.e. websites and apps) as well as human consumption
- Anticipating how resources will be manipulated as well as retrieved via a URL requires skillful web API planning
- Web API performance is affected by server response times, network latency and throughput. Each of these require a different level of optimization work if found to be the source of poor performance.
Why API testing matters:
Originally, APIs were used to enable integrations with a closed set of distributed systems. Today, API usage has evolved, they are now used by a wider audience as a new approach to deploying infrastructure and are considered best practice for any kind of app development.
This is particularly true for mobile app development. In 2013, the number one reason APIs were deployed was to support mobile app development. Going forward, mobile phones and tablets will be the biggest driving force for the development and consumption of APIs (API Evangelist).
With the number of public APIs available expected to hit 10,000 in 2014 and private APIs estimated at ten times that number, staying competitive will require serious performance optimization – ensuring that resources exposed by one’s API can be created, listed, updated and deleted appropriately under expected load (Programmableweb, 2014).
With Load Impact, you can quickly create a load test which simulates different types of REST and SOAP API users – from customers and partners to affiliates and integrators.
For advanced scenarios that include complex flows, randomized requests, authentication and dependencies, use our advanced scripting functionality – which includes built-in support for parsing and serializing JSON, and parsing XML.
When testing APIs it is also often times required to model the test in terms of throughput, requests per second. See our knowledge base article on how to load test APIs with Load Impact.
We help you with all of the following
- Guarantee that resources exposed by your API can be created, listed, updated and deleted appropriately under expected, as well as unexpected, load
- Validate server scalability and measure end user performance
- Ensure the response times experienced by your API users are acceptable and within the boundaries of Service Level Agreements (SLAs)
- Establish a long-term baseline for your API performance on which to validate the impact (good and bad) of infrastructure changes over time
- Determine your maximum capacity in order to understand which components fail first, in what order and what the nature of that failure is