The main responsibility of CIM (Common Information Model) providers is to return useful information about each system-managed resource (device, software component, or any other object on a system). Therefore, in order to ensure that adequate testing is performed (that is, to ensure high coverage), superior testing of CIM providers uses many possible concurrent resources. In addition, in order to ensure that sufficient test variations are covered, it is better still to provide these tests on a variety of different systems. However, the more systems deployed to perform functional and benchmark testing, the heavier the workload will be.

Script-Driven Testing Tool for CIM Providers helps testers to easily build the necessary scripts for batch testing of CIM providers. This batch testing allows for sequential testing of various managed resources without intervention, thereby saving a significant amount of time. The tool also helps testers automatically collect information from each managed resource for each tested system, enabling a useful comparison of results.

This tool logistically handles system log-on, thus obviating a need to be present in order to initiate each test system start; and the tool finishes by generating an HTML test report summarizing the test results for all CIM providers.

Script-Driven Testing Tool for CIM Providers runs on UNIX®-like systems and is dependent on the user's obtaining an open-source component called SBLIM-testsuite on SourceForge.net. The user inputs the desired test scripts for each CIM provider. The basic grammar and script syntax is defined by SBLIM-testsuite. The Script-Driven Testing Tool extends the grammar a little bit to allow the test scripts to be run on remote systems.

How does it work?

SBLIM-testsuite (one component of the project SBLIM, Standards-Based Linux® Instrumentation for Manageability) on Linux/UNIX allows testers to perform only a single CIM provider test on the property level. SBLIM-testsuite requires testers to write scripts (shell, Perl, or any executable script) in order to get the expected output of CIM providers.

In contrast, Script-Driven Testing Tool for CIM Providers enables the appropriate scripts that meet specific test criteria to be run remotely on many different systems without the deployment of these scripts to the systems, all because of network capabilities of the Secure Shell Handler (SSH) protocol. Through an SSH session, commands can run on remote systems as long as the SSH daemon, SSHD, is running on the remote system. As a result, all required test scripts need be installed only on a single system, greatly reducing the complexity of such testing.

On the systems that install all test scripts, all test scripts (include executable scripts and the scripts defined by SBLIM-testsuite) are stored in paths named by the name space of these providers, so that the tool can enumerate all the test scripts and run them sequentially.

Because the output of the open-source SBLIM-testsuite isn't easy to browse or manage, especially when there are many CIM providers to be tested, the tool will parse the output logs and convert them into HTML format. Meanwhile, a summary page will be generated.

Because the tool is dependent on some Linux/UNIX utilities, it can be run only on Linux or UNIX-like systems.

Use of Script-Driven Testing Tool for CIM Providers typically includes the following steps:

1. The tester makes sure that SSHD is running on each tested system.
2. The tester writes test scripts for each provider, following the specification of SBLIM-testsuite. The only difference is to add a special macro in the executable script so that the script can be run by the Script-Driven Testing Tool on a remote system.
3. The tester specifies the host name, user name, and password of the systems to be tested and starts up the tool. (If the tester merely wants to test some of the CIM providers, the provider class names must be specified when starting up the tool.)
4. The tool will go through the input path of all test scripts, get the name space from the path name, and call SBLIM-testsuite for each provider.
5. The tool collects all results from SBLIM-testsuite and parses them, converting them from text files into HTML files and generating a test summary.

Guo Cong Bin, a software engineer, has been working on the CIM and director development for System p® at IBM Systems and Technology""s Beijing development lab since January 2006. He focuses on CIM provider development.