This page describes all problems, tasks and solution approaches regarding the data structure to handle invocation data.

Main topics

1. What would be the best data structure for handing this kind of data?
2. How can the data structure be modified to optimize the data handling?
3. How to handle the data in a efficient way?

4. How to convert the data from the agent to the storage data structure(relational?) or tree structure?

5. How to reconstruct the invocation tree structure from the provided data?(How to recognize the transaction context and put the transaction together?)
6. Allows the data structure to place invocation information independent from the created or missing relations? (e.g. Having only one invocation, but not knowing what was invoked before or after.)

Current solution (state)

One approach, that were already discussed in the master thesis, were to change the data structure from hierarchy based to relation based.
Potential candidates for this is the Adjacency list and Pre-order Traversing.TODO: Ask advisor about this.

Currently data is handled in a relational data structure.

New approaches /

...

ideas

Possible solutions for topic 5:

• This depends on the method used to track the invocation sequences.
  
  (1. An ID can be transmitted parallel to the transaction execution.)
  (2. The tracing could be done without tagging, but trough characteristics (attributes) recognition.
  The transaction would be put together form the communications between JVMs and the processing inside these JVMs.)

Possible solutions for topic 6:

• Organize the information in a matrix(columns= threads, rows = invocations and someway to represent communication between threads)
• Or as tree-structure
  

Questions / Possible Problems

Coming soon...-