ITX8080: erinevus redaktsioonide vahel

Allikas: Lambda
Mine navigeerimisribale Mine otsikasti
← Vanem muudatus
Anttix (arutelu | kaastöö)
218 muudatust
Anttix (arutelu | kaastöö)
218 muudatust
 
(ei näidata sama kasutaja 64 vahepealset redaktsiooni)
2. rida: 2. rida:
== Schedule ==
== Schedule ==


Lecture TH 12:00 - 13:30
Lecture TH 17:45 - 21:15 (odd weeks)


Practice FR 14:00 - 17:30
Practice TH 17:45 - 21:15 (even weeks)


== Communication ==
== Communication ==


Discussinons will be held at: itx8080 atatat googlegroups.com
Discussinons will be held at course forum in Moodle.


Send e-mail to hannesa at-no-spam-please gmail.com for an invitation.
https://moodle.e-ope.ee/course/view.php?id=4319
 
All scenario proposals must be posted to their own respective forums.
 
Please register your account today at http://moodle.e-ope.ee


The lecturers can be reached via the course coordinator:  
The lecturers can be reached via the course coordinator:  
16. rida: 20. rida:
antti.andreimann only-people-please eesti.ee
antti.andreimann only-people-please eesti.ee


== Points awarded so far ==
= Grading =


Arnis : 3 points - ab testing solution, best countermeasure proposed for scenario 4, best snort rules proposed for scenario 6
Best of the bunch will score points.


Igor : 3 points - wordpress virtualization solution, lab 3 report, IDS testing solution
Points will be awareded for the:


Kaarel : 2 points - lab 1 report and conclusions, lab 3 solution
* Best simulation proposal
* Best result obtained in the lab environment
* Best lab report


Märt : 3 points - wordpress testing procedure, best wordpress optimization (wp-cache + php-apc), best lab 2 report.
In total it is possible to score three points for every two weeks.


Roland: 2 points - best Lab 4 report, IDS testing solution
Maximum score is 3, nobody will be awareded more.


Hannes: 1 point - best lab 7 report
Minimum score to pass is 1.


Siim: 1 point - most popular scenario 7 solution
For those who can not score a single point, an
examination proposals will received from those
of you who scored 3 and voted.


= Scenarios =
== Points scored so far ==


== Scenario 1 ==
Margus - 3pts (lecture 1 vote, lab 2 testing method, lab 3 result)
There is a Wordpress based blog.


The client is worried: How many visitors can they handle before they die.
Andres - 1pts (best lab 1 report)


Establish a baseline: How many page displays per second AT LEAST we can handle.
Michal - 1pts (best lab 1 result)
Disruption of the site operation is NOT acceptable.


Minimal budget!
Alexandre - 1pts (best lab 2 result)


Be as detailed and technical as You possibly can.
Roman - 1pts (best lab 2 report)


Post to the discussion group the following information:
Allan - 3pts (best lab 3 proposal with Predrag, lucky SOB scenario 4 proposal, best scenario 5 proposal)


# Detailed description of the simulation of the site. eg. if You're going to virtualize it then how? If You are running it as a process, then how You separate it from the real machine?
Predrag - 2pts (best lab 3 proposal with Allan, best scenario 5 proposal)
# Detailed description of the attack / benchmark tools picked. Where You are going to run them? What tools? How many copies? How you are going to collect the results?!!!


=== Proposals ===
Mauno P - 3pts (best lab 3 report, best lab 4 report, best scenario 7 proposal)
Benchmarking:
* AB - Discussed in lecture
* http://tsung.erlang-projects.org - Jérémy
* wget + AB - Arnis
* http://jakarta.apache.org/jmeter/ - Roland, Jérémy
* http://grinder.sourceforge.net/ - Märt
* http://httpd.apache.org/test/flood/ - Hannes
* Normal monitoring software on live server - Urmas


Simulation of the system:
Robert P - 3 pts (best scenario 4 proposal, best lab 4 result, best lab 5 result)
* VirtualBox - Roland, Jérémy
* VMWare ESX - Igor
* VMWare Player - Marina
* Amazon EC2 - Märt
* KVM - Siim
* Original hardware - Hannes


=== Lab 1 ===
Anis - 1 pts (best lab 4 result)
We will need to create two different virtual machines, one is going to be the wordpress server and the 2nd one is the client which is used to attack the server. It's probably easiest to install only one image and then clone it.


Hard disk image cloning is best done with
Mart - 1 pts (best lab 5 report)


Linux:
Ronny - 1 pts (best lab 7 report)
VBoxManage clonehd FirstUbuntu.vdi SecondUbuntu.vdi


Windows:
Mikheil - 1 pts (best lab 6 report)
"C:\Program Files\Oracle\VirtualBox\VBoxManage.exe" clonehd FirstUbuntu.vdi SecondUbuntu.vdi


After cloning the image create another virtual machine and attach the 2nd disk to that machine. As the result there will be two different virtual machines with two different virtual hard disks.
Jan - 1pts (best lab 8 result)


To be able to connect between the two virtual machines, You need to add an
Franck - 1pts (best lab 8 report)
additional network adapter to both hosts. It can be either "Host Only" or
"Internal" type.


To enable the new adapters in Ubuntu edit the file: /etc/network/interfaces
= Scenarios =


== Scenario 2 ==
== Scenario 1 ==
Design a method for evaluating any speedup solutions for the wordpress.
A repeatable process to evaluate the claims.
Whatever optimizations.
The process must be detailed enough so that somebody else can get the same results when applying that.
The "other person" is expected to have IT knowledge sufficient to install and run a Linux desktop.
The method must specify the level performance
at which the site is considered DEAD.
Budget requirements: Low - Dedicated machine + dedicated tester.


=== Lab 2 ===
There is a Wordpress based blog.
* Get Virtual machine disk image and instructions from: http://www.tud.ttu.ee/~anttix/lab2/ <br /> '''NB! DO NOT DOWNLOAD IMAGE FROM THE LINK NOTED IN INSTRUCTIONS PDF, IT WILL BE SLOW'''
* Fix bug in image (see below)
* Measure the effect of these wordpress plugins:
** [http://wordpress.org/extend/plugins/wp-super-cache/ WP Super Cache]
** [http://wordpress.org/extend/plugins/wp-cache/ WP-Cache]
* Devise Your own optimizations for the site


Bug in image: the ip address of the VM is saved in WP database as 192.168.56.3
The client is worried: How many visitors can they handle before they die.
It is advised tu run the Target VM first and run the "sudo dhclient" command first in Target VM.
Then check active IP of target by command "ifconfig".  If it is not 192.168.56.3, then situation can be fixed by running:
In Target VM: "sudo ifconfig eth2 192.168.56.3"
In Client VM: "sudo ifconfig eth2 192.168.56.4"


==== A note to Windows users ====
Establish a baseline: How many page displays per second AT LEAST we can handle.
Use this command line to access VBoxManage program
Disruption of the site operation is NOT acceptable.


"C:\Program Files\Oracle\VirtualBox\VBoxManage.exe"
Minimal budget!


==== Wordpress Admin access ====
Assume upstream bandwidth can cope...
Username: admin
Password: QIaTemvskSeM


Be as detailed and technical as You possibly can.


==== Note to Ubuntu users ====
=== Proposals ===
VirtualBox with WiFi Does NOT work properly http://ubuntuforums.org/showthread.php?t=782936


== Scenario 3 ==
* Install a similar system on test hardware
Your client is worried about some stuff posted on a blog. They ask You to take care of it.
** Test hardware is smaller than
They have a throwaway "script kiddie" in a third world country, who will mount the attack
* Extrapolate from current system logs
so You don't need to worry about hiding the attackers identity.
** Must have some load ...
** It may NOT be linear (it usually isn't!)
* Generate some load (without breaking it) and extrapolate from logs
* Find existing performance graphs for wordpress ... (highly unlikely)


Devise a way to attack a wordpress (default installation) based site to render it
=== Simulation environment ===
unusable (page view times over 60 seconds).


Attack resources: one PC, a script kiddie, internet connection of 2Mbit/s.
* Real hardware
* Virtualization platform


How are we going to evaluate the solutions?
Pick what is closer to the real system.
* Easy to use instructions
* The most efficient use of available resources


Defacing will count, IF it works for the latest version of wordpress.
=== Benchmarking tools ===


=== Lab 3 ===
* httpPerf
Possible attacks:
* pagetop
* htop
* AB - apache bench
* curl / wget + custom scripts
* JMeter
* Grinder
* wpscan - wordpress specific


* Brute force overload (AB or other benchmark), bang the slowest page possible.
=== How to separate benchmark tools from the server? ===
* Connection flood (slowloris)
* SYN Flood (eg. with hping)
* Banging expensive operations (eg. adding big comments)
* etc.


== Scenario 4 ==
* Use to VMs, Use CPU allocation in virtualization environment to allocate different CPUs for different VMs
Develop countermeasures for the attacks simulated in Lab 3.
* Put VM disk images on different disks (probably an overkill)?
* Monitor process table and make sure that test tools consume significantly less than web server processes. Lower the priority of the test tools.
* Start monitoring load and page views. Generate some load on real system. How?


Limitations:
How do we define the system is dead?
* No additional hardware
* No reactive measures


=== Lab 4 ===
Page view takes longer than 10sec.


Evaluate the countermeasures against attacks.
=== Determining if referenced content has an effect on results ===
The countermeasures must not have negative effects
* Make a list of all references the main page has
on "normal usage pattern".
* Measure the performance of all the URLs
* Disregard the ones that load at least two magnitudes faster.


Normal usage pattern is defined as:
=== How to measure the results? ===
* Bursts of 10 connections per IP address in 1 second
* Page load times not over 20 seconds


== Scenario 5 ==
* Measure response times and number of parallel requests using the benchmarking tool itself
Create a solution and instructions for testing IDS
* Deduct the numbers from server logs
systems against the wordpress site and attacks
* Put the system under load, test with a normal browser if the load time is reasonable.
tried in lab 3.


A repeatable process to evaluate vendor claims.
=== What to document? ===
Whatever passive IDS system delivered as a VM or a dedicated box.


Creating the IDS system itself is out of scope.
* Exact configuration of the benchmarking tool (command line, scripts, config files etc. for EACH test performed)
* The version of the benchmarking tool used
* Virtualization platform
* Hardware specification
* URL of the page tested
* Wordpress configuration
* Wordpress content !!!
* Wordpress version
* Web server configuration
* Web server version (apache, php, sql, etc)
* The measurement results (page load times under each different test)
* Web browser used to measure response time (type, version, operating system)
* Operating system configuration (version, software installed, customizations etc).
* The AUTHOR! and the DATE!


The process must be detailed enough so that somebody else can get the same results when applying that.
Write evertyhing down at ONCE.
The "other person" is expected to have IT knowledge sufficient to install and run a Linux desktop.
if using VM-s save the image.


Budget requirements: Modest - 2 Dedicated machines + dedicated tester + networking equipment to connect these together.
=== Lab1 ===


The process must test at least the following:
* Virtualization platform: VirtualBox
* Port scan
** 512M memory, 1 CPU for each VM
* SYN flood
* Wordpress 3.2.1 with default settings (no extra plugins, no performance tweaks, default skin) and default home page
* slowloris/pyloris
* Ubuntu Server 10.04 LTS 32bit
* Ab overwhelming attack
* Web server: Apache2, SQL: MySQL (default versions bundled with OS)
* Point will go to the person that can produce the most dense graph showing the number of concurrent connections against the response time. Individual tests used to plot the data must be repeated at least 3 times.


=== Lab 5 ===
Instructions for installing Snort IDS and its web GUI called 'acidbase': https://help.ubuntu.com/community/SnortIDS


== Scenario 6 ==
You will need to create two different virtual machines, one is going to be the wordpress server and the 2nd one is the client which is used to attack the server.


Develop/download/find/whatever a SNORT configuration (rulesets, preprocessors, whatever)
To be able to connect between the two virtual machines, You need to add an additional network adapter to both hosts. It can be either "Host Only" or "Internal" type.
that performs better than the default configuration in lab 5 tests.
By better we mean:
* Less false positives
* Less false negatives
* The objectives are contradictory so the rule of thumb is one false negative per 10 false positives eg. solution with 10 false positives and 2 false negatives is better than the solution with 100 false positives and 1 false negatives, but the solution with 10 false positives and 1 false negative is better than the solution with 1 false positive and 2 false negatives.
* Attack is defined by a single invocation of the test script.


== Scenario 7 ==
To enable the new adapters in Ubuntu edit the file: /etc/network/interfaces
 
There is a new remote vulnerability in Windows.
You became aware of a new virus that targets this vulnerability at 7 PM.
The antivirus vendor has released an updated virus database, but it's probably not installed
on all of the computers yet. Your company has 500 employees, with half of them running on laptops. You have to decide in 12 hours, if You need to collect all the employee's laptops in the morning.
 
Collecting the laptops has a high cost in lost employee productivity: There will be a queue of laptops waiting to be inspected. You can assume that all of the laptops can be inspected in half a day.
 
Limits: You can not mitigate the risk by changing the network configuration, You do not know how many laptops are still in the building. You can not take the servers offline.
 
Resources: 5 workers, some networking equipment, real servers (you can not disrupt those), "samples" of existing systems: a few spare workstations, a few spare laptops,
 
What do You know about the virus?
* It disables antivirus software
* It disables automatic software updates
* It can be cleaned up only manually in about 1 manhour
 
You need to develop an action plan:
* What You test in those scarce 12 hours?
* How?
* How do You weigh the results to make the decision?
* The plan itself must be short and easy to understand so You can write it on a whiteboard.
* Justifications for the specific actions go to a separate file/document.
 
== Scenario 8 ==
 
Find a solution to simulate a network of 300 PC-s in a single corporate environment, accessing a software update service at once.
 
Determine the load on primary internet connection and the router.
Assume that updates server will be able to keep up.
 
Tools to consider
* [http://www.ssfnet.org/homePage.html SSFNet]
* [http://www.omnetpp.org/ OMNeT++]
 
Additional resources
* [http://www.ssfnet.org/Exchange/gallery/index.html SSFNet module gallery]
* [http://code.google.com/p/omnet-httptools/ omnet-httptools]

Viimane redaktsioon: 21. detsember 2011, kell 01:55

General Information

Schedule

Lecture TH 17:45 - 21:15 (odd weeks)

Practice TH 17:45 - 21:15 (even weeks)

Communication

Discussinons will be held at course forum in Moodle.

https://moodle.e-ope.ee/course/view.php?id=4319

All scenario proposals must be posted to their own respective forums.

Please register your account today at http://moodle.e-ope.ee

The lecturers can be reached via the course coordinator:

antti.andreimann only-people-please eesti.ee

Grading

Best of the bunch will score points.

Points will be awareded for the:

  • Best simulation proposal
  • Best result obtained in the lab environment
  • Best lab report

In total it is possible to score three points for every two weeks.

Maximum score is 3, nobody will be awareded more.

Minimum score to pass is 1.

For those who can not score a single point, an examination proposals will received from those of you who scored 3 and voted.

Points scored so far

Margus - 3pts (lecture 1 vote, lab 2 testing method, lab 3 result)

Andres - 1pts (best lab 1 report)

Michal - 1pts (best lab 1 result)

Alexandre - 1pts (best lab 2 result)

Roman - 1pts (best lab 2 report)

Allan - 3pts (best lab 3 proposal with Predrag, lucky SOB scenario 4 proposal, best scenario 5 proposal)

Predrag - 2pts (best lab 3 proposal with Allan, best scenario 5 proposal)

Mauno P - 3pts (best lab 3 report, best lab 4 report, best scenario 7 proposal)

Robert P - 3 pts (best scenario 4 proposal, best lab 4 result, best lab 5 result)

Anis - 1 pts (best lab 4 result)

Mart - 1 pts (best lab 5 report)

Ronny - 1 pts (best lab 7 report)

Mikheil - 1 pts (best lab 6 report)

Jan - 1pts (best lab 8 result)

Franck - 1pts (best lab 8 report)

Scenarios

Scenario 1

There is a Wordpress based blog.

The client is worried: How many visitors can they handle before they die.

Establish a baseline: How many page displays per second AT LEAST we can handle. Disruption of the site operation is NOT acceptable.

Minimal budget!

Assume upstream bandwidth can cope...

Be as detailed and technical as You possibly can.

Proposals

  • Install a similar system on test hardware
    • Test hardware is smaller than
  • Extrapolate from current system logs
    • Must have some load ...
    • It may NOT be linear (it usually isn't!)
  • Generate some load (without breaking it) and extrapolate from logs
  • Find existing performance graphs for wordpress ... (highly unlikely)

Simulation environment

  • Real hardware
  • Virtualization platform

Pick what is closer to the real system.

Benchmarking tools

  • httpPerf
  • pagetop
  • htop
  • AB - apache bench
  • curl / wget + custom scripts
  • JMeter
  • Grinder
  • wpscan - wordpress specific

How to separate benchmark tools from the server?

  • Use to VMs, Use CPU allocation in virtualization environment to allocate different CPUs for different VMs
  • Put VM disk images on different disks (probably an overkill)?
  • Monitor process table and make sure that test tools consume significantly less than web server processes. Lower the priority of the test tools.
  • Start monitoring load and page views. Generate some load on real system. How?

How do we define the system is dead?

Page view takes longer than 10sec.

Determining if referenced content has an effect on results

  • Make a list of all references the main page has
  • Measure the performance of all the URLs
  • Disregard the ones that load at least two magnitudes faster.

How to measure the results?

  • Measure response times and number of parallel requests using the benchmarking tool itself
  • Deduct the numbers from server logs
  • Put the system under load, test with a normal browser if the load time is reasonable.

What to document?

  • Exact configuration of the benchmarking tool (command line, scripts, config files etc. for EACH test performed)
  • The version of the benchmarking tool used
  • Virtualization platform
  • Hardware specification
  • URL of the page tested
  • Wordpress configuration
  • Wordpress content !!!
  • Wordpress version
  • Web server configuration
  • Web server version (apache, php, sql, etc)
  • The measurement results (page load times under each different test)
  • Web browser used to measure response time (type, version, operating system)
  • Operating system configuration (version, software installed, customizations etc).
  • The AUTHOR! and the DATE!

Write evertyhing down at ONCE. if using VM-s save the image.

Lab1

  • Virtualization platform: VirtualBox
    • 512M memory, 1 CPU for each VM
  • Wordpress 3.2.1 with default settings (no extra plugins, no performance tweaks, default skin) and default home page
  • Ubuntu Server 10.04 LTS 32bit
  • Web server: Apache2, SQL: MySQL (default versions bundled with OS)
  • Point will go to the person that can produce the most dense graph showing the number of concurrent connections against the response time. Individual tests used to plot the data must be repeated at least 3 times.


You will need to create two different virtual machines, one is going to be the wordpress server and the 2nd one is the client which is used to attack the server.

To be able to connect between the two virtual machines, You need to add an additional network adapter to both hosts. It can be either "Host Only" or "Internal" type.

To enable the new adapters in Ubuntu edit the file: /etc/network/interfaces

Pärit leheküljelt "http://www.lambda.ee/w/index.php?title=ITX8080&oldid=8433"