Meeting 2

Meeting Notes

Agenda:

• Overall explanation of the readings:
  • SDN related attacks
  • Datasets are artificial not in real life
  • High volume DDoS successfully is found
  • Most common attacks are DDoS
  • Mininet is used for testing and emulation
  • feature selection, ML is 90-95% with train
  • low-volume attacks 80-85% but the their test dataset is not artificial
  • future work and gaps defined in this paper, we can focus on this points.
  • Instrusion detection system is not for the DDoS but we want to ask whether it is good area or not.
  • we take a look to the shrew attacks. here the results seem good?
  • flash events and DDoS are hard to differentiate
  • multi-controller or data center based SDN can have different problems.
  • buffer saturation , congestion of the controller - data link
  • Security of OpenFlow Switches
  • Traffic dataset