-
Introduction
- What is a Darknet, effects, features for its users
- Arising problems from its characteristics
- name the analysing methods (and 1 sentence explanation)
- analytic
- simulation
- emulation
- testbed (and/or measurement in real world networks)
- Advantages of event based large scale simulation and in generell analysability of darknets
-
Darknets
- Detailed explanation of characteristics
- (short) what is a P2P overlay; how they differ
- only connected to people one trusts
- no (concrete) topology information leakage
- Implications of those characteristics
- Survey of existing darknets
- (historical perspective ?)
- mainly grouped/sorted by used test/analysing methods
- implemented/pure theoretical work
- (turtle/wase,gnugnet,oneswarm,freenet,mcon,x-vine, (eventuell prefix embedding von Andreas Höfer)
- Recap and explain used metrices and analysis methods
- Detailed explanation of characteristics
-
My Model
- Why we chose the simulation based approach
- Node based with fixed neighbor set
- Churn model extension with bootstrapping & offline detection
- Selection of OMNet++
- Implementaion of the Model
- two simple example routing models
- randomwalk
- with n-degree fanout
- loopdetection
- flooding
-
Simulation and Analysis
- Simulation environment
- Used metrices
- (average/max) path length
- sent message count
- faild routings / requests OR droped packages
- Q: Scalability of the model/framework
- Used RAM; RAM RAM and moar RAM
- Q: Impact of fanout degre at randomwalk on found pathlength
- comparison to flooding which finds shortest path
- Q: Probability of path failure on return for churn model
-
Conclusion and Future Work
- Everything is epic, but.. ;)
- Implement tested darknets and compare to their tests
- Implement untested/new nets and improve routing parameters
- or even decide if algorythm is pratical usefull or not