Automated test : UI vs API

When you start to have too many UI tests and it takes hours to complete them, maybe something is wrong in term of automated test strategy.

The first reaction when it takes ages, you start to run nightly test, revise them one by one and find eventually duplicates and finally try parallelize them when it is possible. And if this is still not enough?

Maybe you should define the most critical (20-30%) and modify the other using API calls instead of using UI.

When you test modern application with multiple layer, the UI layer should have no logic and no data. That means if you execute on API or UI level you should have the same results. But on API level the test are more quicker. In fact what you should test on the UI side is if the UI is behaving properly using the API and the security is aligned with the API, then the rest all should be done on API side.

Moving test from UI to API is not obvious because you may also modify all the UI test to cover the alignment between UI and API.

This is why you should think from the beginning to adopt the right test automation strategy with a good balance between API and UI!

For project using open source tool, RestAssured is a good choice to automatized API test. In this Github repository, you could find an example on how to use API test.

Automated Test 3 ways of doing.

In 2017 we could say there are 3 ways of doing Automated test. The typology of the these 3 ways are based on their cost maintainability.

  1. The most easy one but the most complex to maintain is the Record type. All the commercial tools and even Selenium with Selenium IDE, propose this way of doing. But when the SUT (system under test) is changing the automated test could be drastically compromised and need a full rework of the test script.
  2. The second way and more developer oriented is the Scripted type. The maintainability is less complex and could be handle some design change if the script and the UI is properly identified.
  3. The third one is based on Workflow and Data Driven typethere are still scripts but there are more easy to maintain. If a step in the workflow changes, you just need to change those steps. Almost all the commercial tools are proposing these kind of way of doing. With Selenium if you use the BDD with Cucumber will work in that way.

Now in term of implementation the most expensive in maintenance is the most easy to implement.

So when it the time to choose which way you want to implement tests you need forecast how complex,  how many tests you want and how many time they are run.


CBTA – Prerequisites

  • An Add-On to be installed on SAP Solution Manager 7.1 SP7 (onwards)
  • An agent which is part of ST-PI
  • A  frontend component to be installed on the user’s desktop.

Useful documentation on Solman 7.2 and 7.1 and Test Automation

Solman 7.2 and Test Suite


  • Compatibility Matrix (Mainly for SolMan 7.1 but as long you use the CBTA version for latest version of SolMan 7.1 Sp14, it will work with SolMan 7.2)
  • Configuration (Also available in the How-to Guide below)
  • Prerequisites (Also available in the How-to Guide below)
  • Solman 7.2 Test automation overview on CBTA
  • Component Based Test Automation (CBTA) How-to Guide : This how to is a bit outdated and is targeting SolMan 7.1 SP12 and CBTA 3.0 SP4.
    • Create automated test from Business Blueprint (Transaction : SOLAR02) : from page 35
    • Create automated test from Test Repository (Transaction : SOLMAN_WORKCENTER) : from page 82
    • Create automated test in debug mode : from page 107 (it will execute the test step by step with outputs of the object status)
    • Creating custom library/function : from page 153

OpenStack on a single machine

If you want to go more further in Containers, and you want to have a dedicated machine for it. You will need a few more “thing” above Docker. You will need an Hypervisor and a container orchestrator.
For an Hypervisor you could take LXD and an OpenStack as orchestrator.

At the first glance this could take a lot of configurations steps. But guess what, there is a “Next, Next, Next” installation!

Here is the link and it is super easy and check the tutorial video:


Service Virtualization – Open source and commercial tools


This POC (Proof of concept) tend to analyse different service virtualization available on the market.


  • POC = Proof of concept
  • Service Virtualization = “A test double often provided as a Software-as-a-Service (SaaS), is always called remotely, and is never working in-process directly with methods or functions. A virtual service is often created by recording traffic using one of the service virtualization platforms instead of building the interaction pattern from scratch based on interface or API documentation.” (ref1)
  • Stub = “A minimal implementation of an interface that normally returns hardcoded data that is tightly coupled to the test suite.” (ref1)
  • Mock = “A programmable interface observer, that verifies outputs against expectations defined by the test.” (ref1)

Main feature to consider in Service Virtualization

  • Multi protocol (not only HTTP).
  • Record record real transaction to re-play them for testing purpose.
  • Learn and cache responses.

List of Service Virtualization (Open source)

None yet if we consider SV main features.

List of Mocks (Open source)

Name Download link Comments
Citrus Framework Works with a large spectrum of protocols (http, jms, ftp, mail, etc…)
Hoverfly Learn more
Mockito  Java only
Wilma Dockerized solution

List of Service Virtualization (Commercial)

Name Download link Comments
HP SV Learn more
Traffic Parrot

List of Mocks (Commercial)

Name Download link Comments


Samy Kacem

HP Service Virtualization

Install HPE _SV Designer 3.82 CE

  • Windows 2012R2
  • SQL 2012 Express SP1


We will use the API from ft-demo-website project => API. We will explore HPE SV and We will virtualize 2 methods from the API:

  • Read list of books
  • Add a book

Read list of books

Original : http://../api/listBooks.php
This GET Rest Api provides a list of books

  • We need to configure HPE Sv to learn the service:
    • Go to Add a new virtual service…
    • Select I don’t have a service description & click Next
    • Select REST & click Next
    • Select agent HTTP Gateway because we want to simulate the endpoint.
    • Add in Real Service > Endpoints the original rest api as specified above & click Next
    • Start to learn by selecting Learn and create a call using the Virtual Service not the Real service
    • System is learning thanks to this unique call. You will see that you have been redirected to the real service. When it the system is learning it captures the call, redirects it, plays as proxy and records the answer.
    • Switch to Simulate now & do again the call using again Virtual Service url. As you can see, you are not redirected and it is providing the same content. Now the service is kind of cached. We can say this service is stubbed as per this definition.

Add books

Original : http//../api/insertBook.php?title=toto&author=tutu&edition=Sun%20Edition
This GET Rest Api adds a book with the content defined in the param. You create a new Virtual service as the step above. Then put the 2 services in Learning mode.
Before starting, reset the db using http//../api/createTable.php (this will not simulated), and start to alternate the call (list, add, list, add etc…) then switch to Simulate and redo the same pattern. You will see the system is redoing the same outputs.
Now lets trick the system. Call only listBooks.php, you will notice he will loop as per the pattern even we don’t use insertBook.php! We you could face some issue for complex Business Process. How you could solve it? By using script or data driven set.

Beyond a simple stub

After recording you can more data scenarios. You have 2 options that could be used even together:

Script (C# or Javascript) to add business logic.
Data driven set managed with an Excel file and you define all the possible request and response required for your test.
Data test strategy

If you are stuck with some poor data or it is hard to use test data. It could be a good tool to put in place and have a new data test set. Based on your test scenarios, you create all the set of data you want to have in requests and answers and you will have fairly quick your data test set. This should be consider when you are doing your test strategy.


It works perfect if you are doing automated test but if you are doing manual explanatory tests it will be wired.
If you need to put more business logic it is still possible. For each service virtualized, you can add scripts with business logic. But how far you want to replicate the business logic? This could be complicated if you need a perfect replication, that means you will have 2 technologies to maintain with the same business logic. You should avoid creating scripts and answer as much as possible with Data driven set. This will be more easy to maintain.

Samy Kacem


Let’s do a quick overview on this open source tool for Service Mocking / Service Virtualization.


  • Using v0.11.0
  • For test purposes we are using ft-demo-website
  • Docker (if you want to use it, it is optional)


There 3 ways of using Hoverfly:

  • Using mainly for stub purpose, we set a json file that configure the appropriate answer we want. Mainly there are static answers. You can see the demo.
  • Using as simple mock with middleware in python.
  • Have a more complexe mock for complex workflows using Hoverfly Java.


For the demo we took the first way of doing, just stubbing. You can see the demo. We are mainly 2 simples configurations:

  • We record a api call. We have created a bash script that generate the json file you will need to reuse for replay purpose.
  • We replay the api call. We have created a separate bash script that start locally a replay api call. It will use the json of the first step.

That’s it!


Now when you have your json for replay you can create a docker for mocking. Just one single command line:

docker run -d -p 8888:8888 -p 8500:8500 spectolabs/hoverfly:latest -import ${HOVERFLY_JSON_URL}


Samy Kacem

Data test strategy and Service Virtualization

Actually I’m digging in different SV (=Service Virtualization) Solution like HP SV, OSV from Tricentis and CA SV. In this context, it allows you to cut from production the real data and play with test data.

What commercial SV can do for you?

  1. Record transactions : You record all the transactions you need for your test scenarios, then the system is playing back all the responses as required.
  2. Scripting the service with a business logic if you need some complex response.
  3. Data driven set : Allows you to fill an Excel spreadsheet or any data container with all the data you require for your test scenarios. Your data set need to be “stateful business process” oriented, that means during the data lifecycle, the object could change state and value.

This last feature will be a good candidate if you are stuck with a legacy or new applications in order to put in place a test data set. The majority of the commercial tools could work with different protocol (http/https, MQ, Filesystem, Ftp/s, etc…). And if the price is a show stopper and you require only http/https, some open source tools are also good candidates. Here you can find a good list of open source and commercial tool.

Samy Kacem