ChallengeResponse Systems - PowerPoint PPT Presentation

1 / 24
About This Presentation
Title:

ChallengeResponse Systems

Description:

Udi Manber of Yahoo! approached researchers at Carnegie-Mellon University this problem. ... These robots sign up email accounts, then send messages containing ... – PowerPoint PPT presentation

Number of Views:87
Avg rating:3.0/5.0
Slides: 25
Provided by: rain47
Category:

less

Transcript and Presenter's Notes

Title: ChallengeResponse Systems


1
Challenge-Response Systems
  • Group 8

presents
2
What is Challenge-Response?
Challenge-response is the an abstract concept
referring to the act of limiting access to some
resource by a certain procedure. This procedure
is defined as follows
  • Person attempts to gain access to restricted
    material
  • System challenges person
  • Person responds to challenge
  • System grants access if response is satisfactory

3
Why do we need a challenge?
  • Privacy Some content is considered personal,
    and users will want either exclusive access, or
    want to give only a very small set of people
    access.
  • Security Some content is sensitive and should
    only be seen by persons with certain credentials.
  • Protection Content vital to the operation of
    businesses, countries, etc should only be handled
    by persons properly trained and equipped to do so.

4
Why do we need a database?
There are a number of things that we need to keep
track of, such as
  • The definition of privileged in the context in
    which we are implementing our CR1 system
  • What content is considered priveleged
  • Who should have access to what
  • Questions and Answers There needs to be at
    least as many questions as possible users, and at
    least one (and preferably only one) correct
    answer to each question.

1 Challenge-response
5
Why do we need a database?
We need to be able to keep track of each of these
things, separately, or all together, in a
convenient way. Convenient in this case means
  • The information should be restricted in much the
    same way that the CR system operates.
  • Changing the information should be easy for
    anyone who is authorized to do so.
  • The verification process should be as efficient
    as possible.
  • We should be unrestricted in the kind of
    information that we can keep track of.

6
Why do we need a database?
  • All of these criteria can be satisfied by storing
    the information in a properly designed and
    implemented database.
  • Today, most, if not all database management
    systems have built-in functions supporting
    authentication.
  • In fact, this functionality is now incorporated
    into DBMSs to safeguard the information contained
    within.
  • Storing information in a database allows us to
    easily attach information not critical to
    authentication (biological data, for example), to
    information required for authorization.

7
What types of CR systems are there?
Challenge-response systems can be divided into
two categories.
  • Cryptographic A CR system falls in this
    category simply by making use of some sort of
    cryptography in its implementation.
  • Non-cryptographic
  • In some ways, older than cryptographic CR.
  • Basic forms of this existed only for as long as
    it took for attackers to devise means of
    circumvention.
  • More recently, new kind of non-cryptographic CR
    has appeard in the form of counter Turing tests.

8
Non-cryptographic CR
Basic authentication is the simples form of a
challenge-response system. Explanation of the
process will be accompanied by a live
demonstration
  • User requests access to restricted content
  • System asks user to present proof of
    authorization
  • User provides authorization information
  • System verifies whether information is valid
  • If it is, user is granted access to content

9
The need for encryption
With the concept of remote authentication came
vulnerabilities which attackers can (and do)
exploit.
  • Enter ye unscrupulous person
  • User requests access to restricted content
  • System asks user to present proof of
    authorization
  • User provides authorization information
  • Attacker listens to network traffic and copies
    information
  • System verifies whether information is valid and
    grants access.
  • The attacker can then pose as the user and gain
    access to the restricted content!
  • OH NO!

10
The need for encryption
To counter this, the authorization process is
altered slightly.
  • User requests access to restricted content
  • System asks user to present proof of
    authorization
  • This time, sends random string with request.
  • User provides authorization information
  • Information is encrypted using the random string
    as a key
  • Attacker listens to network traffic and copies
    information
  • System decrypts and verifies information, and
    grants access
  • An attacker, posing as the user, is given a
    different encryption key
  • Previously recorded information is now invalid

11
Build a better lock
  • and you build a better thief.
  • Malicious person can get around the previously
    described setup by listening to traffic both ways
    and, recording the key, and finding out how
    information is encrypted.
  • More clever ways of protecting information must
    be devised.
  • This just mean dedicated attackers will find
    another way to go about their evil activity.
  • This back-and-forth is expected to continue
    indefinitely

12
Challenge-Response Proper
  • Rise of the CAPTCHA

13
The Turing Test
In 1950, Alan Turing devised a test to rate the
intelligence of a machine
  • A human participant would hold a conversation
    with two other correspondents
  • One of these would be a machine with some measure
    of artificial intelligence
  • If the human cannot tell the difference between
    the other human correspondent, and the computer,
    then the computer passes the test, and is
    considered intelligent.

14
CAPTCHA!
  • Stands for Completely Automated Public Turing
    Test to Tell Humans and Computers Apart
  • Is a variation of the Turing test, devised in
    such a way that humans will pass and machines
    will fail
  • This aspect of CR systems, is what most people
    are referring to when they mention
    challenge-response.

15
Why would we need to tell Humans and Computers
Apart?
Case Study 1 The AltaVista add-URL problem
  • If a website is to be found via search engine, it
    must first be submitted.
  • Search engines have algorithms for rating sites
    which correspond with keywords.
  • One aspect of AltaVistas rating system was (is)
    how many times a page is submitted, assuming that
    different, interested persons will go through the
    submission process.
  • Tech-savvy persons took advantage of this by
    writing scripts to repetitively add their sites
    to the search engine.
  • This skewed the search engines ratings
    significantly.

16
Why would we need to tell Humans and Computers
Apart?
Case Study 2 The Yahoo! chat room problem
  • Udi Manber of Yahoo! approached researchers at
    Carnegie-Mellon University this problem.
  • Robots2 were invading Yahoo! chat rooms and
    soliciting information from participants
  • They were also filling the chat rooms with
    advertisements and links to other sites.

2 Automated scripts
17
Why would we need to tell Humans and Computers
Apart?
Case Study 3 Spam!
  • More then 60 of email traversing the internet is
    spam3.
  • Most, if not all spam is generated and propagated
    by robots.
  • These robots sign up email accounts, then send
    messages containing advertisements and other
    unwanted content to millions of other accounts.
  • Most of the time, this is done solely to bog down
    the mailservers, i.e. for no good reason.

3 unwanted, unsolicited messages, also called
junk mail
18
How can we tell Humans and Computers Apart?
In the initial development stages, some desirable
traits of a suitable CAPTCHA were put forward.
  • The test should present a challenge which can be
    automatically generated.
  • Human users should be able to take the test
    quickly.
  • It should accept virtually all human users (even
    young or naive users) with high reliability,
    while rejecting very few
  • The test should reject virtually all machine
    users.
  • It should resist automatic attack for many years
    even as technology advances and even if the
    algorithms are known (e.g. published and/or
    released as open source)

19
How can we tell Humans and Computers Apart?
  • Even though scientist are trying to close the gap
    between man and machine, the current separation
    can be, and has been utilized.
  • For the CAPTCHA project, the relevant gap is the
    shortcoming of computer vision, or more
    precisely, character recognition.
  • The human faculty for recognition, though taken
    for granted, is astounding.
  • Computers recognition facilities are not so
    astounding.
  • This is not expected to change very much in the
    near future.

20
So how does it work?
  • Yall know what were talking about, but just in
    case
  • When trying to access content that only a human
    should be privy to, the user is presented with a
    sequence of characters, generated
    pseudo-randomly4.
  • These are subject to a subset of a number of
    visual distortions including scaling, skewing,
    warping, speckling and background noise.
  • The user is asked to type the characters as they
    appear.
  • Human users will have very little passing the
    test quickly.
  • Robots cannot pass without cheating.

4 from a large enough selection set so as to
appear random
21
Antispam
The most talked about application of CR systems
is as a means of countering junk mail, or
spam. The explanation of how this works will be
accompanied by live demonstration.
  • Legitimate user sends email to protected user
  • Enter Microshaft SpamGuard XP! Proffessional 2006
    Elite Spam-blocking Edition PRO!
  • SpamGuard checks whilelist5 for address on email
  • Address is not on list, so SpamGuard sends a
    CAPTCHA back to sender

5 a list of addresses that do not need
verification
22
Antispam
  • The user takes test and sends back correct answer
  • SpamGuard verifies answer, allows message to go
    to protected user, and puts newly verified
    address on whitelist
  • User will not have to take test again, since
    address is now on whitelist

5 a list of addreses that do not need verification
23
Antispam
  • Enter ye Ms. Roboto!
  • Robot sends email to protected user
  • SpamGuard checks whilelist5 for address on email
  • Address is not on list, so SpamGuard sends a
    CAPTCHA back to sender
  • Robot, if it even notices that there has been a
    reply, cannot make head or tail of the test
  • SpamGuard receiving no reply to the CAPTCHA,
    tosses the email
  • Protected user, ideally, does not even know about
    it

24
Pros and Cons
  • The obvious advantage is that the CAPTCHA does
    what it is supposed to do (mostly).
  • CAPTCHA has come under considerable criticism in
    online circles. Some complaints are.
  • It can be circumvented in a number of ways by
    spammers who are clever enough.
  • It is annoying to have to fill out a test just to
    send an email
  • Important mail sometimes ends up in the junk
    folder
  • Sometimes one CR system will challenge another CR
    system, and the message ends up in oblivion while
    the systems send challenges back and forth
  • Foils mailing lists which though automated, are
    actually solicited
Write a Comment
User Comments (0)
About PowerShow.com