Create and run automated unit tests, system tests and integration tests for SoftwareAG's webMethods™ IntegrationServer. Thanks to its declarative approach, IwTest let's you focus on data rather than on mechanics.
There are two costs associated with test cases:
The cost to create them
The cost to maintain them
and IwTest addresses both. Its main advantages are:
IwTest's unique feature is the ability to record service invocations server-side and use the recorded inputs and outputs to generate test cases, including stubs. No need to manually save inputs and outputs.
Creating a test suite is as simple as:
Recording not only works on the IntegrationServer you're developing on. You can also record services in an environment with more realistic data, e.g. in your User Acceptance Test environment and download the recorded data for test suite generation.
Note: Exceptions are recorded and treated as expected error situations. IwTest will include them automatically when generating a test suite.
You can incorporate IwTest easily in your CI/CD-pipeline. Each test suite is a regular flow service, so deploying the tests is not any different
than deploying a package. Use for example
wget to run your test suites. These filters are supported in order to
select the test suites to run:
Run all available:
Run only unit tests:
Run only the test suites in the packages whose names start with Acme:
Download the results in JUnit format:
Use IwTest to create unit tests in no time. Let IwTest record the service invocations while you manually execute your flow services
Use IwTest to create system tests with equally little effort by additionally recording data for the services you would like to stub.
Use IwTest to create end-2-end tests by adding callbacks and additional stubs for asynchronously executed services
Are you facing a webMethods upgrade and you don't have a set of reliable regression tests? Use IwTest to create them quickly. Let IwTest record service invocations in your QA environment and build a regression test suite over night! You're saving incredible amounts of time, effort and tedious work.
IwTest minimizes the time necessary to maintain your test cases with it unique ability to repair test cases.
Say you have a service with 25 test cases and suppose you need to change the service. Now you have 25 failing test cases. Rather than update each test case manually, you can use the repair facility and have IwTest update them for you. It will even retain all the regular expressions you might have defined in the expected results.
When one of the services that you want to stub is executed multiples times when you invoke the top-level service, things get complicated. Of course you don't want it to return the same result every time, and it's cumbersome to write an actual stub service as a replacement for the real service.
This scenario is tackled by IwTest by using dynamic stubs. IwTest will record the different input/output combinations of the sub-service. During test case execution IwTest will replace the calls to the external service with a dynamic stub. The actual input is matched against the recorded inputs and on a match the corresponding output (or exception!) is returned.
IwTest supports pub/sub scenario's via so-called callbacks. In your test case you tell IwTest to capture the pipeline that's present at the beginning or end of a service invocation and run assertions against that data. This service may be located on another IntegrationServer!
IwTest doesn't care how the (asynchronous) execution of that service was triggered. It may have been a Broker message, JMS message, MQTT event, Web Service call etc.
By the way, you're not limited by just one callback: you can define as many in a test case as you like.
Of course on the remote IntegrationServer you can again define stubs so you can prevent that the asynchronously executed service sends data to an external system.
Watch this video to get an impression of how IwTest works:
Execute your test cases from your Continuous Integration tool (e.g. Jenkins, or Bamboo) with a single command. With another single command download the results in JUnit-format.
Jenkins has an excellent JUnit plugin that keeps history of the various test runs, highlights the differences between them, produces nice graphs and notifies developers when a build failed or has become unstable.
You can download and evaluate IwTest for free. The first 90 days it's fully functional, only the number of test cases that you can execute is limited to 50. After the evaluation period the recording and test suite generation function is disabled.
You can either purchase a permanent license or take a subscription. The permanent license doesn't expire, but is only valid for a major version. With a subscription on the other hand you're entitled to all upgrades.
|# Test cases||50||Unlimited||Unlimited|
|Bug fixes||90 Days|
|Minor upgrades||90 Days|
|Major upgrades||90 Days|
You can request a quote by sending an email to
IwTest consists of two regular IS packages: IwTest & IwTestAgent. A package with samples is also included.
|1.6.3||May 8 2022||IwTest v1.6.3||16b84ab6ada3cb0f101af6ff999f6612|
The latest version of the major release are always downloadable from this website. If you purchased a permanent license and would like to upgrade to the newest new major version, then ask for a quote by sending an email to
You can send bug reports and feature requests to
The following features are planned for upcoming releases:
WmTestSuite lets you leverage inputs and outputs you've gathered in Designer. In IwTest however you can record on the server and include sub-services in order to generate stubs for them. Furthermore, IwTest supports asynchronous and distributed scenario's, which WmTestSuite does not. Besides, it lacks a repair function.
CATE provides similar and more advanced functionality, but divided over two different products (CATE Developer Edition (DE) and CATE Enterprise Edition (EE)) and against a much higher license fee. Besides, CATE lacks a repair function.
IwTest does not support (JMS) Messaging directly. However, IwTest does support asynchronous (pub/sub) and distributed scenario's.
In your test case you can wait for a service to execute asynchronously, either on the same IS, or on a remote IS, and validate its inputs, outputs or even exception.
If there is no subscribing service (yet), then you could also create a separate trigger in your test package and subscribe to the message that you want to capture.
Note that you can extend your test case with multiple asynchronous events.
Test cases come at a cost. First there is the cost to create them (with IwTest, that cost is relatively small, but still) and later there is the cost to update them as you apply fixes or enhance the functionality.
As a first approach, focus on your transformation services (of course you've separated your transformation logic from the tranport logic, right?).
The best time is when initial development is done and your code is in functional acceptance testing, but before it's released to production.
Later you can shift your focus to end-to-end scenarios and - time and budget permitting - to utility services.
If you have your test data available, then with IwTest's record & generate facility, it's literally a matter of minutes. Also if you want to stub services.
If on the other hand you need to prepare your test data and vary it in order to test different mapping variations or to evoke different responses from the target system, then it will take as long as it takes you to prepare your data.
Yes, but IwTest cannot link asynchronous events to one test case. So the data will be recorded, but in pub/sub scenario's you'll have to manually extend a test case it with a so-called 'callback' to add an asynchronous event.
No, IwTest doesn't produce excessive amounts of data unless you tell it to. There is a parameter that controls the minimum interval between two service invocation recordings. The default value is 2 seconds.
Yes! Just specify the services IwTest should collect data for.
Warning! IntegrationWise does not recommend to use IwTest in a production environment. Do this only in a QA environment.
Obviously there is a performance penalty when recording data that flows through you integration platform. And although IwTest has been tested extensively, IntegrationWise cannot assume responsibility for any negative effects the use of IwTest might have.
IwTest supports the same data types the IDataXMLCoder class supports, so that's everything you can do a 'Save Pipeline to File' for. So IwTest does not support an xml 'node'. If the service you want test takes a node as an input, simply create a wrapper flow service that takes an xml-string as an input, call 'pub.xml:xmlStringToXMLNode' and then invoke your flow service with the 'node'.
Most probably your service produces different results each time it's executed. There might be for example a timestamp in the outputs, or an auto-generated id. This is something IwTest cannot detect.
The solution is to use a regular expression in the expected results. This also works for fields that are numbers or dates. For example replace:
Because CI tools can take care of this. Jenkins for example has a marvelous JUnit plugin. It compares the test results of the current run with those of the previous runs and presents the differences nicely. In addition it draws nice graphs of history of the test runs. There is nothing IwTest could add here.
No, use tools like JMeter to do that. Nevertheless, IwTest can be of great help here. Before you run your performance tests you can tell IwTest to create stubs for those services that interact with an external system. By isolating your integration platform like this, you can much better judge how well it performs. Usually the external systems are the performance bottle necks, not the integration platform.
No, IwTest does not require wM DevOps Edition.
IwTest is a product of IntegrationWise B.V., a Netherlands based IT consultancy that specialises in integration projects using SoftwareAG's webMethods™ Digital Business Platform.