On this page the order is described, in which the topics and problems are addressed and solved.
The order might change during progress.Currently I try to create
Concepts (On hold)
Creating own concepts and methods that are independent from Christians ideas.
According to the 'Top-Down Approach", I will start started creating my concept on a very abstract level.
And from there move to the less abstract levels while providing more detailed solution approaches.
The first step is was to create a very general approach that covers the whole perspective of the Tracing Invocation Sequences goal.
The second step will be divide that approach and verify possibilities for realization.
Working Order(Outdated)
...
Inner-Thread tracing
...
Inter-Thread tracing
...
Inter-JVM tracing
...
Unique ID generation (Probably should be on first place since further approaches will base on IDs.)
...
Method sensor storage
...
Continues Sending
...
Data structure to handle invocation data
...
It turns out that it unlikely to find an approach that is independent of the communication protocols used by the monitored system.
To get practical results, one protocol is picked to be traced.
HTTP
HTTP is chosen to be the first one to be traced. Due the fact that additional context information can be easily transmitted in the request header this protocol is suited best for testing initial tracing approaches and finding common problems that will likely apply to other communication protocols.
The following steps will be taken.
- Implementation of a basic web service (DONE)
- Researching and implementing the ID transmission (In progress)
- Creating test cases to assure transparency (optional since HTTP header does not block the execution)
- Researching and implementing a way to inject the ID transmission via instrumentation