Service Virtualization

 View Only
  • 1.  VS created via WSDL unable to use Learning Mode?

    Posted Sep 29, 2016 10:36 PM

    I tried creating a Virtual Service Image via WSDL and the VS model does not contain the steps for Live Invocation, and therefore unable to use the other execution mode such as Learning mode and Live Invocation.


    Will there be any concerns if I add in the live Invocation Step manually? Have anybody else here did such modifications and encounter any errors or unexpected behavior?


    I am also considering extracting the xml messages and create the VS via R/R pairing instead, since the VSM will come with those steps by default.



  • 2.  Re: VS created via WSDL unable to use Learning Mode?

    Posted Sep 30, 2016 10:04 AM

    I like the option of using R/R pairs better than using WSDL because many times most of the elements in the WSDL are defined as optional and the resulting VSI contains values that do not appear very realistic.


    Take the VSM generated from WSDL and manually add the Live Invocation step -- should be OK but you have to deal with all of the branching logic.  With R/R pairs, you get the luxury of allowing DevTest to create a proper model including the Live Invocation with values that make much more sense.


    In terms of learning, I cannot help a great deal in that area.

  • 3.  Re: VS created via WSDL unable to use Learning Mode?

    Posted Oct 02, 2016 10:57 PM

    That was my initial thought on adding the step manually, that it could be too tedious and also prone to mistakes as I am pretty new to this.


    At the moment, I am extracting the messages from the Virtual Service Image created via WSDL and use those message pairings to create a new Virtual Service image via R/R pairing.


    That should be quite safe?




  • 4.  Re: VS created via WSDL unable to use Learning Mode?

    Posted Oct 03, 2016 03:22 PM

    Generally speaking, it would be a reasonable approach to decompose the service from WSDL into R/R pairs to generate a more viable service.  Another option would be to create or reuse a VSM that has the Live Invocation step.  Overlay the VSM created by the WSDL and change it to point to the VSI (in the VS Image Selection Step) generated by the WSDL.  


    Keep in mind though that *optional* WSDL elements may or may not be passed on the incoming request when the system is connected to the service.  The danger of not having specific Use Case data scenarios is that the WSDL may generate a VSI containing all of the elements -- converted into request arguments.  Whereas, the system-under-test may only send a subset of these elements during a given call to the service.  This disparity could cause your VSI to mismatch since the number of arguments on the request is different from the number of arguments in the VSI.  In this case, you may need to add an RDM DPH to manipulate the argument list that is passed to the VSI.  We would not be able to advise you in this regard without more knowledge of the interface contract.

  • 5.  Re: VS created via WSDL unable to use Learning Mode?

    Posted Oct 03, 2016 08:55 PM

    Thanks! I am aware of that risk of mis-match request messages. I am just  working on the image as I wait for the test environment to be available for me to do a recording of the scenarios, and happen to find the differences in the VSM when using different methods of creation.


    I would have assumed that all Virtual Services support the same features, regardless the method of creation.