mleoking的个人博客分享 http://blog.sciencenet.cn/u/mleoking

博文

主动队列管理算法 与 拒绝服务攻击 模拟实验平台

已有 5771 次阅读 2012-4-11 17:12 |系统分类:科研笔记| 管理, center, 算法, AQM

AQM&DoS Simulation Platform

Copyright (c) 2010-2012 Changwang Zhang (mleoking@gmail.com). All rights reserved.

This Active Queue Management and Denial-of-Service (AQM&DoS) Simulation Platform was established for the Robust Random Early Detection (RRED) algorithm [1]. If you use any part of this platform in your research, you have the responsibility to cite this platform as:

The experiments (or simulations) are conducted on the AQM&DoS Simulation Platform that was created for the Robust Random Early Detection (RRED) algorithm [1].

1. Changwang Zhang, Jianping Yin, Zhiping Cai, and Weifeng Chen, RRED: Robust RED Algorithm to Counter Low-rate Denial-of-Service Attacks, IEEE Communications Letters, vol. 14, pp. 489-491, 2010.

Platform Homepage: http://blog.sciencenet.cn/home.php?mod=space&uid=571128&do=blog&id=558146 

Redistribution and use in source and binary forms, with or without modification, are permitted provided that the following conditions are met:

1. Cite this platform in the redistribution using the way mentioned above.

2. The above statements are kept in the redistribution.

1. The function of the platform

The Active Queue Management and Denial-of-Service (AQM&DoS) Simulation Platform is based on the Network Simulation 2. It is able to simulate a variety of experimental scenarios related to Distributed Denial-of-Service (DDoS) attacks and Active Queue Management (AQM) algorithms.

The platform can simulate numbers of Denial-of-Service attacks:

  • Denial-of-Service (DoS) attacks
  • Distributed Deinal-of-Service (DDoS) attacks
  • Spoofing DoS or DDoS attacks
  • Low-rate Denial-of-Service (LDoS) attacks
  • Distributed Low-rate Denial-of-Service (DLDoS) attacks
  • Spoofing LDoS or DLDoS attacks

And a variety of Active Queue Management (AQM) algorithms:

  • 1 DropTail; 2 RED; 3 RED/PD; 4 Blue; 5 SFB;
  • 6 CBQ; 7 FQ; 8 SFQ; 9 DRR; 10 PI;
  • 11 Vq; 12 REM; 13 GK; 14 SRR;
  • 15 RED/Robust (RRED) 16 SFB/Robust (RSFB);

To analyse the impact of DoS attacks on normal TCP flows and AQM algorithms, this platform also provides mechanisms to automatically calculate and record the average throughput of normal TCP flows before and after DoS attacks.

The experimental network has a dumbbell topology as the network experimented in the RRED algorithm [1].



2. The installation of the platform

AQM&DoS Simulation Platform is mainly tested on ns-2.33, but it is expected to be compatible with higher versions of ns.  If you are using a different version of ns, please replace "2.33" with the version number of your ns in all the following instructions.

To experiment on the AQM&DoS Simulation Platform, you should fellow these steps:

1. Unzip the package of the AQM&DoS Simulation Platform in your Linux system (the subdirectory "result" is necesary to output the simulation result, you should keep it) and run the following command in the directory "aqm-dos-sim-plat".

                chmod +x leodos.sh

2. Install the ns-allinone-2.33 simulation software in your operation system.

NS-2.33: http://sourceforge.net/projects/nsnam/files/allinone/ns-allinone-2.33/ns-allinone-2.33.tar.gz/download

Note1: AQM&DoS Simulation Platform is tested on ns-2.33, but it is expected to be compatible with higher versions of ns.

Note2: AWK is also required to run the platform. But most users do not need to manually install it as it is already included in most Linux distributions. If it is not included in your Linux system, you can refer to the following link to install it.

AWK: http://www.gnu.org/software/gawk/

Or if you are using Debian or Ubuntu Linux, you can use the following two commands to install AWK:

                1. sudo apt-get install gawk

                2. cd /usr/bin/ && sudo ln -s gawk awk

3. Integrate RRED into your NS2 distribution.

                Please follow the instruction in "ns2-integrationintegration-of-rred.txt"

4. Modify simulation settings in "leodos.sh" to conduct your specified experiments.

                You need to modify the parameters in the "leodos.sh" to conduct a variety of simulations.

                4.1 The following line of code means to conduct a single simulation using the parameters specified in the head of "leodos.sh":

                                                dosim 0;

                4.2 The following line of code means to conduct a batch of simulations on a specified AQM algorithm x:

                                                "task_aqm_ldos x;"                           

                                x is the number of the AQM algorithm. The mapping of x to AQM algorithms is:

                                1 DropTail; 2 RED; 3 RED/PD; 4 Blue; 5 SFB 6 CBQ 7 FQ; 8 SFQ;        9 DRR; 10 PI; 11 Vq; 12 REM; 13 GK; 14 SRR 15 RED/Robust 16 SFB/Robust;

                                If you want to experiment on a specific AQM algorithm, please remove the # before its line.                           

                The original setting of the "leodos.sh" is to only conduct a single simulation.

                You might need to understand and modify the logic of the function "task_aqm_ldos" to conduct your specified batch of simulations. 
                      

5. Run the simulations using the following command in the directory "aqm-dos-sim-plat".

                ./leodos.sh

                The experimental results are located in the sub-directory "result", including:

                                1. The overall trace file "leodos.tr"

                                2. The TCP trace file "leodos_tcp.tr"

                                3. The queue monitor trace file "leodos_queue_monitor.tr"

                                4. The bottleneck queue trace file "leodos_queue.tr"

                                5. The nam trace file "leodos.nam" (To get the nam trace file, you need to change the value of "ns_of" from 2 to 3 in "leodosh.sh")

                     6. The log files "leodos.log" and "leodos_sh.log". "leodos.log" records the parameters of each simulation and its statistical results in a format shown in Section 4. If you run a batch of simulations using "task_aqm_ldos", these log files will be located in a sub-directory named "AQM_x" (x is the number of the AQM algorithm) under "result".


Optional steps:    

o1. Integrate the ip spoofing function into your NS2 distribtuion (Do this step only if you need to simulate spoofing DDoS attacks).

                Please follow the instruction in "ns2-integrationintegration-of-ip-spoofing.txt"

o2. Integrate SFB/blue into your NS2 distribution (Do this step only if you need to simulate SFB).

                Please follow the instruction (README) in "ns2-integrationns2-blue.tar.gz" - the code and instruction of SFB/blue.

o3. Integrate RSFB (Resilient Stochastic Fair Blue) into your NS2 distribution (Do this step only if you need to simulate RSFB and have finished the step o2).

                Please follow the instruction in "ns2-integrationintegration-of-rsfb.txt"

3. The parameters of the platform

Name

Description

Unit

bn_bw

Bottleneck bandwidth

Mbps

bn_dl

Bottleneck link delay

ms

bn_qs

Bottleneck queue size

packets

bn_qm

The AQM algorithm used in the bottleneck link. 
The mapping of bn_qm to AQM algorithms is:

1: DropTail
2: RED
3: RED/PD
4: Blue
5: SFB
6: CBQ
7: FQ
8: SFQ
9: DRR
10: PI
11: Vq
12: REM
13: GK
14: SRR
15: RED/Robust (RRED)
16: SFB/Robust (RSFB)

 

nt_bw

Network bandwidth (except the bottleneck link).

Mbps

nt_dl

Network delay (except the bottleneck link).

ms

hp_n

It is not used in this version of the platform.

 

ur_n

The number of normal users

 

ur_cr

The number of new connections per second. It is used to simulate http traffic.

 

ur_ps

User flows packages size

Byte

ur_st

The start time of normal user flows

second

ur_sp

The stop time of normal user flows

Second

ur_rs

Whether normal user flows randomly start in the period between ur_st and ur_sp.

0: normal user flows will all start at ur_st
1: normal user flows will randomly start in the period between ur_st and ur_sp

 

ur_pt

Normal user flows' type.

1: TCP

 

ur_app

The application of normal user traffic.

0: FTP; 1: Telnet; 2:PackMimeHTTP; 3:PackMimeHTTP_DelayBox

 

ak_n

The number of attackers

 

ak_ng

The number of attacker groups that attackers are organised into. Most of the time, you do not need to change the default value of this parameter "1".

 

ak_tg

The start time difference between attacker groups. Most of the time, you do not need to change the default value of this parameter "0".

 

ak_st

The start time of attack flows

 

ak_sp

The stop time of attack flows

 

ak_rs

The option currently only support Low-rate DoS attacks (when ak_bp < ak_ap). For most of time, you should set this to be 0.

Whether attack flows randomly start in a Low-rate DoS attack period between 0 and ak_ap.

0: attacker flows will all start at ak_st
1: attacker flows will randomly start in every attack period between 0 and ak_ap

 

ak_ps

Attacker flows' packages size

Byte

ak_ap

Attacker flows' attack period

ms

ak_bp

Attacker flows' burst period.

For a DoS or DDoS attack ak_bp equals ak_ap
For a LDoS or DLDoS attack ak_bp is smaller than ak_ap

ms

ak_pr

Attacker flows' packages rate

Mbps

ak_tp

The attack type. Most of the time, this value should be 1.

1:represents period attack
2:represents follow tcp cwnd attack

 

ak_mw

Not used in this version of the simulation platform

 

ak_cp

Not used in this version of the simulation platform

 

ak_spf_lv

Whether attackers use spoofing IP address.

0: attackers use real IP addresses
1: attackers use spoofing IP addresses

ak_spf_mn and ak_spf_mn are two integers. The spoofing address range is from ak_spf_mn to ak_spf_mn.

 

ak_spf_mn

The minimum spoofing address.

 

ak_spf_mx

The maximum spoofing address.

 

tm_fi

The simulation finishing time.

second

ns_db

Whether output the debug information.

0: do not output debug information
1: output debug information

 

ns_of

The output files of the simulation platform.

When ns_of:
>=3 output leodos.nam (used for nsnam to figure the simulation topology)
>=2 output leodos.tr leodos_tcp.tr leodos_queue_monitor.tr 
>=1 output leodos_queue.tr (The data trace of the bottleneck queue. It is the primary analysis data source for this simulation platform)

 

4. The output of the platform

An example of the platform output is:

ak_spf_mx           60000

nt_dl       2

ur_sp      240

ak_st      120

ur_cr       100

ur_n        30

li              1

ns_of      1

ak_spf_lv              0

ak_bp    200

ak_pr     0.25

bn_qm   15

ur_st       20

ak_spf_mn           100

ak_ps     50

tm_fi      240

ak_tp     1

ur_app   0

ak_rs      0

ak_ng     20

bn_bw    5

ak_n       20

ur_ps      1000

ur_pt      1

nt_bw     10

ur_rs       0

bn_qs     50

ak_ap    200

hp_n       25

ns_db     0

ak_cp     10

ak_sp     220

ak_tg      0

ak_mw  1

bn_dl      6

bn_qms RED/Robust

leodos_queue_awk: dt=0.010000 s_l=0.000000 t_st=0.000000 t_sp=240.000000 ur_n=30 ak_n=20 ur_st=20.000000 ur_sp=240.000000 ak_st=120.000000 ak_sp=220.000000 p_ct=pktcount

rate_f1_normal  600.680000      rate_f1_attack  597.220000      nth_f1  0.994240        rate_f2_attack  17.940000

 

The lines from "ak_spf_mx" to line "bn_qms" are detailed parameters of this simulation (please refer to Section 3 for the meaning of these parameters). The followed lines are statistical results:

  • "rate_f1_normal" depicts the average throughput rate (packets/s) of normal TCP traffic through the bottleneck link when there is no DoS/LDoS attack.
  • "rate_f1_attack" depicts the average throughput rate (packets/s) of normal TCP traffic through the bottleneck link when a DoS/LDoS attack is attacking (from ak_st to ak_sp).
  • "nth_f1" represents the preserved ratio of normal TCP traffic throughput under a DoS/LDoS attack, which equals to rate_f1_attack/rate_f1_normal.
  • "rate_f2_attack" depicts the average throughput rate (packets/s) of attack traffic through the bottleneck link when a DoS/LDoS attack is attacking (from ak_st to ak_sp).
References

[1]   Changwang Zhang, Jianping Yin, Zhiping Cai, and Weifeng Chen, "RRED: Robust RED Algorithm to Counter Low-Rate Denial-of-Service Attacks," IEEE Communications Letters, vol. 14, pp. 489-491, May 2010. [PDF | REF]


 ---|--|--|--|--|---



https://blog.sciencenet.cn/blog-571128-558146.html

上一篇:低速率拒绝服务(Low-rate DoS)攻击相关研究进入SCI最新文章列表
下一篇:研究发现(艾滋病)病毒与计算机网络感染类似并警示应及早治疗
收藏 IP: 128.16.8.*| 热度|

1 黄富强

该博文允许注册用户评论 请点击登录 评论 (2 个评论)

数据加载中...

Archiver|手机版|科学网 ( 京ICP备07017567号-12 )

GMT+8, 2024-3-28 21:45

Powered by ScienceNet.cn

Copyright © 2007- 中国科学报社

返回顶部