Logo etsi

ETSI's Bug Tracker

Notice: information submitted on the ETSI issue Tracker may be incorporated in ETSI publication(s) and therefore subject to the ETSI IPR policy.

View Revisions: Issue #7910 All Revisions ] Back to Issue ]
Summary 0007910: Allow parallel control parts/components
Revision 17-12-2020 17:20 by Gyorgy Rethy
Description Proposal:

Allow creation of components (called maybe parallel control components or PCC) from a control component. The current main control part would run on the MCC (master control component).

Example:

control {
  var ControlComponent pcc1 := ControlComponent.create;
  var ControlComponent pcc2 := ControlComponent.create;
  pcc1.start(f_start_testcase());
  pcc2.start(f_start_mirror_testcase());
  all component.done;
}

function @control f_start_testcase() {
  execute(TC());
}

function @control f_start_mirror_testcase() {
  execute(TC_mirror());
}

Semantics:

Every parallel control part has the same semantics and capabilities as the current control part. Since there are no global variables in TTCN-3, there should be no interference between the testcases, so parallelization should not be an issue.
 
The semantics of 'all component', 'any component' would need to be defined in the control context. It should refer only to the control components.
Stopping and killing of control components should kill the testcase currently running on the component (if any).

Intercomponent communication for synchronization purposes between control components could also be allowed (for instance if information needs to flow between the MCC and the PCCs).

Even system communication could be allowed, e.g. when system resources need to be allocated globally for a parallel testcase run or other upper tester communication.
Revision 17-12-2020 16:12 by Gyorgy Rethy
Description
Proposal:

Allow creation of components (called maybe parallel control components or PCC) from a control component. The current main control part would run on the MCC (master control component).

Example:

control {
  var ControlComponent pcc1 := ControlComponent.create;
  var ControlComponent pcc2 := ControlComponent.create;
  pcc1.start(f_start_testcase());
  pcc2.start(f_start_mirror_testcase());
  all component.done;
}

function @control f_start_testcase() {
  execute(TC());
}

function @control f_start_mirror_testcase() {
  execute(TC_mirror());
}

Semantics:

Every parallel control part has the same semantics and capabilities as the current control part. Since there are no global variables in TTCN-3, there should be no interference between the testcases, so parallelization should not be an issue.
 
The semantics of 'all component', 'any component' would need to be defined in the control context. It should refer only to the control components.
Stopping and killing of control components should kill the testcase currently running on the component (if any).

Intercomponent communication for synchronization purposes between control components could also be allowed (for instance if information needs to flow between the MCC and the PCCs).

Even system communication could be allowed, e.g. when system resources need to be allocated globally for a parallel testcase run or other upper tester communication.
Revision 13-02-2020 15:26 by Jacob Wieland - Spirent
Description

Proposal:

Allow creation of components (called maybe parallel control components or PCC) from a control component. The current main control part would run on the MCC (master control component).

Example:

control {
  var ControlComponent pcc1 := ControlComponent.create;
  var ControlComponent pcc2 := ControlComponent.create;
  pcc1.start(f_start_testcase());
  pcc2.start(f_start_mirror_testcase());
  all component.done;
}

function @control f_start_testcase() {
  execute(TC());
}

function @control f_start_mirror_testcase() {
  execute(TC_mirror());
}

Semantics:

Every parallel control part has the same semantics and capabilities as the current control part. Since there are no global variables in TTCN-3, there should be no interference between the testcases, so parallelization should not be an issue.
 
The semantics of 'all component', 'any component' would need to be defined in the control context. It should refer only to the control components.
Stopping and killing of control components should kill the testcase currently running on the component (if any).

Intercomponent communication for synchronization purposes between control components could also be allowed (for instance if information needs to flow between the MCC and the PCCs).

Even system communication could be allowed, e.g. when system resources need to be allocated globally for a parallel testcase run or other upper tester communication.


MantisBT 1.2.14 [^]
Copyright © 2000 - 2024 MantisBT Team
Powered by Mantis Bugtracker