Add in-depth support for a loop of continous monitoring/exploitation

[mojtaba-eshghie]

The exploit runner based on different test environment is ready.

Do we want to generate the report based on the exploits or the models? Neither; we would like to generate a contract-based report; and for each contract there could be several models running at the same time monitoring different aspects of the execution. Therefore, the generated report should be something in the lines of: SampleContract

• Monitored against
  • Model X
  • Transaction 1 -> Valid (in green color)
  • Transaction 2 -> Violation (in red color)
  • Model Y
  • Transaction 1 -> Violation (in red color)
  • Transaction 2 -> Violation (in red color)
  • Model Z
  • Transaction 2 -> Valid (in green color)

In short, the contract transaction indices are global (per contract) and we mark which transaction which model captures.

We need to build the system:

• Start from each contract, for each contract (we're basically starting the monitor in the following steps)
  • For each test of this contract (If there are no tests, just run a and c)
    1. For each model,
       • based on the model id, make a new simulation for the model using the engine API
       • Based on each model id, execute conventions for that model
       • Execute general conventions (for instance execute the constructor activity, etc.)
       • Execute model-specific conventions
       • Initiate a consumer for that model
    2. Execute the whole test using runner (currently, each test/exploit we include also has an oracle included in itself so that it checks its own success; However, in a normal usage the oracle is the model running, and the tests/exploits could be merely sending transactions. Here, we are testing the monitoring system, therefore, the oracle should be in the exploits to see if the exploit is successful, can the model detect it?)
    3. Store the results from the monitor to generate the report later based on the following

[mojtaba-eshghie]

Detailed report:

SampleContract

Monitored against: Model1

• TX 1 (clickable link to see the exact transaction trace) -> Valid
• TX 2 (clickable link to see the exact transaction trace) -> Violation

  Monitored against: Model2

• TX 2 (clickable link to see the exact transaction trace) -> Violation

Summary report:

Contract	Exploit/Test	Exploit/Test Details	Captured Violations
SampleContract	Contract1Temporal	Exploited Successfully	-

[mojtaba-eshghie]

We would like to run parallel exploits/experiments to increase the performance; however, we need to care about the fact that we need to keep track of each test that runs and monitor cannot distinguish between multiple exploit versions. Therefore, parallel sessions of the environment+monitor+exploit_runner are required together to keep track of the things correctly.