ddos flooding attack detection throug h a step by step
play

DDoS flooding attack detection throug h a step-by-step investigation - PowerPoint PPT Presentation

DDoS flooding attack detection throug h a step-by-step investigation


  1. DDoS flooding attack detection throug h a step-by-step investigation 0101001111110010011011010101010100101101110011111001010011101101011111010110111110011101111111100010001001110010000111100011100011101110111011110110100111010010101010011111010100010111001111001111100101001110110101111101011011111001110111 Jaehyun Jun Realtime Image Processing & Telecommunication Lab 16 December 2011

  2. Contents 010100111111001001101101010101010010110111001111100101001110110101111101011011111001110111111110001000100111001000011110001110001110111011101111011010011101001010101001111101010001011100111100111110010100111011010111110101101111100111011111111000 • Introduction • Existing DDoS attack detection method – DDoS attack methods – DDoS attack detection method • DDoS flooding attack detection through a step-by-step investi gation • Experiment • Conclusion 16/12/11 Realtime Image Pr Processing ng & Te Telecommuni unication n Lab. 2 2

  3. Introduction (1/2) 010100111111001001101101010101010010110111001111100101001110110101111101011011111001110111111110001000100111001000011110001110001110111011101111011010011101001010101001111101010001011100111100111110010100111011010111110101101111100111011111111000 • People have been provided many convenient, because general ized computer network infra building and increasing the numb er of internet users could make easy access to computer data • The crime occurs economical damage by flooding traffic to net work or illegal access to computer system which could stop th e right service • Still the DDoS attack has been tried and the damage is contin uously raised – Target for and attack has been diversified for example, game business site, stock site, internet portal site, (Yahoo, Amazon, etc) 16/12/11 Realtime Image Pr Processing ng & Te Telecommuni unication n Lab. 3 3

  4. Introduction (2/2) 010100111111001001101101010101010010110111001111100101001110110101111101011011111001110111111110001000100111001000011110001110001110111011101111011010011101001010101001111101010001011100111100111110010100111011010111110101101111100111011111111000 • Distributed Denial of Service (DDoS) – The structure of DDoS attack – Role of DDoS attack nodes NAME ROLE Attacker is leading all attack operate instrume nt by remote control and delivers command di Attacker rectly Master is receiving the command from attack er and orders attack zombies managed by it Master They are controlled by master. Attack progra m operates the command came from each ma Zombie ster, and finally performs their attack to the vi ctim As final victim, simultaneously they are attack Victim ed from several hosts 16/12/11 Realtime Image Pr Processing ng & Te Telecommuni unication n Lab. 4 4

  5. Existing DDoS attack detection method (1/4) 010100111111001001101101010101010010110111001111100101001110110101111101011011111001110111111110001000100111001000011110001110001110111011101111011010011101001010101001111101010001011100111100111110010100111011010111110101101111100111011111111000 • DDoS attack methods – SYN Flooding attack, UDP Flooding attack, ICMP Flooding attack 16/12/11 Realtime Image Pr Processing ng & Te Telecommuni unication n Lab. 5 5

  6. Existing DDoS attack detection method (2/4) 010100111111001001101101010101010010110111001111100101001110110101111101011011111001110111111110001000100111001000011110001110001110111011101111011010011101001010101001111101010001011100111100111110010100111011010111110101101111100111011111111000 • DDoS attack methods – SYN Flooding attack, UDP Flooding attack, ICMP Flooding attack 16/12/11 Realtime Image Pr Processing ng & Te Telecommuni unication n Lab. 6 6

  7. Existing DDoS attack detection method (3/4) 010100111111001001101101010101010010110111001111100101001110110101111101011011111001110111111110001000100111001000011110001110001110111011101111011010011101001010101001111101010001011100111100111110010100111011010111110101101111100111011111111000 • DDoS attack methods – SYN Flooding attack, UDP Flooding attack, ICMP Flooding attack 16/12/11 Realtime Image Pr Processing ng & Te Telecommuni unication n Lab. 7 7

  8. Existing DDoS attack detection method (4/4) 010100111111001001101101010101010010110111001111100101001110110101111101011011111001110111111110001000100111001000011110001110001110111011101111011010011101001010101001111101010001011100111100111110010100111011010111110101101111100111011111111000 • DDoS attack detection method 5-tuple srcIP, dstIP, srcPrt, dstPrt Packet flows clusters Threshold using feature value distribution (cluster key, entropy) Classification Dominant state Behavior classes Dominant state analysis Structural modeling 16/12/11 Realtime Image Pr Processing ng & Te Telecommuni unication n Lab. 8 8

  9. DDoS flooding attack detection through a step-by-step i nvestigation (1/4) 010100111111001001101101010101010010110111001111100101001110110101111101011011111001110111111110001000100111001000011110001110001110111011101111011010011101001010101001111101010001011100111100111110010100111011010111110101101111100111011111111000 DDoS attack First danger Second danger Third danger N Total volume detection < volume threshold T 1 Y N Dest_IP entropy < Dest_IP entropy threshold T 2 Y Src_port entropy N < Src_port entropy threshold T 3 Y N Packets / sec < Packets / sec threshold T 4 Y Normal state DDoS state 16/12/11 Realtime Image Pr Processing ng & Te Telecommuni unication n Lab. 9 9

  10. DDoS flooding attack detection through a step-by-step i nvestigation (2/4) 010100111111001001101101010101010010110111001111100101001110110101111101011011111001110111111110001000100111001000011110001110001110111011101111011010011101001010101001111101010001011100111100111110010100111011010111110101101111100111011111111000 • Volume threshold(T 1 ) method judging DDoS attack detection Traffic Compare Volume threshold (T 1 ) 16/12/11 Realtime Image Pr Processing ng & Te Telecommuni unication n Lab. 10 10

  11. DDoS flooding attack detection through a step-by-step i nvestigation (3/4) 010100111111001001101101010101010010110111001111100101001110110101111101011011111001110111111110001000100111001000011110001110001110111011101111011010011101001010101001111101010001011100111100111110010100111011010111110101101111100111011111111000 • The comparison between entropy of destination IP address (T 2 ) and entropy of source port number (T 3 ) for detecting DDoS attack First danger YES Compare NO dest_IP entropy threshold (T 2 ) Second danger YES Nomarl State Compare NO Src_port entropy threshold on dest_IP (T 3 ) Third danger 16/12/11 Realtime Image Pr Processing ng & Te Telecommuni unication n Lab. 11 11

  12. DDoS flooding attack detection through a step-by-step i nvestigation (4/4) 010100111111001001101101010101010010110111001111100101001110110101111101011011111001110111111110001000100111001000011110001110001110111011101111011010011101001010101001111101010001011100111100111110010100111011010111110101101111100111011111111000 • The comparison packet creation rate per second (T4) for dete ction DDoS attack Compare Src_port entropy threshold on dest_IP (T 3 ) Third danger Nomarl NO Compare State Packets / sec threshold (T 4 ) YES DDoS attack state 16/12/11 Realtime Image Pr Processing ng & Te Telecommuni unication n Lab. 12 12

  13. The result of experiment (1/4) 010100111111001001101101010101010010110111001111100101001110110101111101011011111001110111111110001000100111001000011110001110001110111011101111011010011101001010101001111101010001011100111100111110010100111011010111110101101111100111011111111000 • Experiment environment – by using OPNET simulation – Component : node_1~12(zombie node), node_13~16(normal node), ro uter_1~5, server – Utilized traffic : DDoS attack traffic, Normal traffic 16/12/11 Realtime Image Pr Processing ng & Te Telecommuni unication n Lab. 13 13

  14. The result of experiment (2/4) 010100111111001001101101010101010010110111001111100101001110110101111101011011111001110111111110001000100111001000011110001110001110111011101111011010011101001010101001111101010001011100111100111110010100111011010111110101101111100111011111111000 • Experiment result and analysis Creation rate each node packet Traffic amount flow in router_5 when DDoS attack 16/12/11 Realtime Image Pr Processing ng & Te Telecommuni unication n Lab. 14 14

  15. The result of experiment (3/4) 010100111111001001101101010101010010110111001111100101001110110101111101011011111001110111111110001000100111001000011110001110001110111011101111011010011101001010101001111101010001011100111100111110010100111011010111110101101111100111011111111000 • Experiment result and analysis The entropy of source port number of traffic The entropy of traffic destination IP address judged the second danger flowed in router_5 when DDoS attack happ ens 16/12/11 Realtime Image Pr Processing ng & Te Telecommuni unication n Lab. 15 15

Recommend


More recommend