Security, Privacy and Trust - Lecture 11 - Web Information Systems (4011474FNR)

2 December 2005
Web Information Systems
Security, Privacy and Trust
Prof. Beat Signer
Department of Computer Science
Vrije Universiteit Brussel
http://www.beatsigner.com

Beat Signer - Department of Computer Science - bsigner@vub.ac.be 2December 12, 2014
Security Aspects
 Authenticity
 knowing the sender or receiver of data
- who is trying to access data on a web server
- who is offering a service
- who sent an email
- …
 Privacy
 keeping information private
- protect credit card information that is sent to a server
- protect information sent in emails
- …
 Integrity
 ensuring that information is not changed when transferred

HTTP Authentication
 Native authentication functionality offered by HTTP
 instead of directly sending a response for a given request, the
server can always respond with an authentication challenge
(401 status code)
 HTTP is extensible to support different authentication
protocols and offers the following two standard protocols
 basic access authentication
- simple Base64 encoding of the string <username>:<password>
 digest access authentication
 Protected resources can be grouped in security realms
with different sets of authorised users or groups of users

Basic Access Authentication
Client Server
GET /wise/exam.pdf HTTP/1.0
Client Server
Client Server
Client Server
ask
password
try to access
a protected
resource
HTTP/1.0 401 Authorization Required
WWW-Authenticate: Basic realm="WISE"
Authorization: Basic YmVhdDpydWxleg==
HTTP/1.0 200 OK
Content-type: application/pdf
Internet

Base64 Encoding
 Base64 encoding can be used to represent binary data
in a portable format (alphabet)
 used by MIME for content transfer encoding
 used to embed binary data in XML files (e.g. in XML-RPC)
 note that Base64 encoded data needs more space
 Takes a sequence of bytes (8-bit) and breaks it into 6-bit
chunks
 padding with 0s to make it a multiple of 24 (LCM of 6 and 8)
 complete 6-bit padding chunks are represented by the special
character '='
 Each 6-bit chunk is then represented by a character from
a 64-character alphabet

Base64 Encoding Example
 Let us encode the string
'No' to Base64
 padding to 24 bit
 lookup of 6-bit chunks in
index table
 use '=' for completely padded
6-bit chunks
val
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
char
A
B
C
D
E
F
G
H
I
J
K
L
M
N
O
P
val
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
char
Q
R
S
T
U
V
W
X
Y
Z
a
b
c
d
e
f
val
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
char
g
h
i
j
k
l
m
n
o
p
q
r
s
t
u
v
val
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
char
w
x
y
z
0
1
2
3
4
5
6
7
8
9
+
/
01001110
N o
01101111 00000000
19 38 60
T m 8 =
Base64 index table
Text
Bit Pattern
Index
Base64
padding

Proxy Authentication
 We can use the same authentication approach for
controlling access to proxy servers
 The proxy will return slightly different HTTP headers
HTTP/1.0 407 Proxy Authentication Required
Proxy-Authenticate: Basic realm="WISE"

Web Server Configuration
 Example configuration for an Apache HTTP Server
 Create a new password file (using the –c parameter)
 Put an .htaccess file with the configuration into the
directory that has to be protected
 alternatively add information to httpd.conf
#htpasswd -c /usr/local/apache/admin/passwords nelson
New password: nelson123
Re-type new password: nelson123
Adding password for user nelson
AuthType Basic
AuthName "WISE"
AuthUserFile /usr/local/apache/admin/passwords
Require user nelson

Basic Access Authentication ...
 Basic access authentication is not secure
 username and password are sent almost in "cleartext"
- Base64 value can be very easily decoded
 easy to do replay attacks
- simply reuse the username and the password
 Potential solutions
 combine the basic access authentication with an encrypted data
transfer (e.g. via TLS/SSL)
- does not necessarily prevent replay attacks
 use of alternative digest access authentication

Digest Access Authentication
 Password is no longer sent in cleartext
 only a one-way digest that is computed out of the password
(one-way hash function) is sent to the server
 Message Digest #5 (MD5) is a popular digest function
 What about digest replay attacks?
 server sends a special token (nonce) that changes frequently
 client adds the nonce to the password before computing the MD5
- any changes of the nonce result in changes of the digest which helps to
prevent replay attacks
h1 = MD5(username:realm:password)
h2 = MD5(httpMethod:requestedURI)
response = MD5(h1:nonce:h2)
Computed response based on MD5

Digest Access Authentication ...
Client Server
Client Server
Client Server
Client Server
ask
password
HTTP/1.0 401 Unauthorized
WWW-Authenticate: Digest realm="WISE",
qop="auth,auth-int" nonce="6G543RED"
Authorization: Digest username="nelson",
realm="WISE", nonce="6G543RED",
qop="auth", response="HF779RW47R7HF",
...
HTTP/1.0 200 OK
Authorization-Info: nextnonce="7HZT7F6"
...
Internet
try to access
a protected
resource

Digest Access Authentication ...
 The Authorization-Info: nextnonce="..." is used
to send the next nonce in advance
 client can send the computed hash value already with the original
request (preemptive authorization)
 The quality of protection (qop) field is used to
negotiate different protection mechanisms
 auth
- authentification
 auth-int
- authentification and message integrity protection
- add an MD5 of the body

Transport Layer Security (TLS)
 Cryptographic protocol to
ensure secure network
communication
 successor of the Secure
Socket Layer (SSL) protocol
 situated at the TCP/IP
Application Layer or OSI
Presentation Layer
 Types of authentification
 unilateral authentification
- only server authentification
 mutual authentification
- client and server authentification
TCP/IP stack
Transport
Application
Link
Internet
TLS/SSL

Cryptography
 In cryptography a cipher (coding scheme)
is used in combination with a key to create
a ciphertext out of a plaintext
 Cryptanalysis tries to get information out of the ciphertext
without having access to the secret information (key)
MEET ME
AT NOON
PHHW PH
DW QLLQ
MEET ME
AT NOONcipher
(encoder)
cipher
(decoder)ciphertext
key key
plaintext plaintext

Symmetric Key Cryptography
 A symmetric key cipher uses the same key for the
encoding and decoding of a plaintext message
 Many existing symmetric key ciphers
 DES, Triple DES, Blowfish, Rijndael/AES, ...
 The algorithms are often common knowledge and the
key is the only secret thing
 key has to be kept secret
 Brute force attack (enumeration attack) tries all keys
 The key length defines the number of potential keys
 e.g. 128 bit key considered safe today
- can change with more powerful machines

Symmetric Key Cryptography ...
 One problem of symmetric key cryptography is that we
have to secretly share the common key before we can
exchange any messages
 this has to be repeated with different keys for any two partners
willing to establish a secret communication
 how should we establish the exchange over the Internet?
- initially only an insecure channel is available
 where should we secretly store all those keys?

Public Key (Asymmetric) Cryptography
 Instead of a single key, public key cryptography uses an
asymmetric pair of keys
 publicly available key for the encoding
 secret key for the decoding
 Each party has only a single public key which is used by
everybody to encode messages to this party
 only the receiver can decode message with their private key
MEET ME
AT NOON
hJ7FHDu
KJF Z8e
fsdlgi MEET ME
AT NOONcipher
(encoder)
cipher
(decoder)ciphertext
public key B private key B
plaintext plaintext
A B

Public Key (Asymmetric) Cryptography ...
 Public key cryptography can be used to establish secure
Internet connections to any computer around the world
without having to secretly share a key beforehand
 An asymmetric public key cipher has to ensure that an
attacker cannot compute the private key based on any
information they can intercept
 public key
 ciphertext (with corresponding plaintext)
- can easily be created by any party by using the public key
 A well known public key algorithm is the RSA cipher

RSA Cipher (Rivest, Shamir and Adleman)
 Public-key cipher that can
be used for encryption as
well as signing
 published in 1978 by Rivest,
Shamir and Adleman while
they were at MIT
 The public and private keys are
generated based on two large distinct prime numbers
 the potential attacker will know about the product of the two prime
numbers but nothing about the numbers themselves
 use modular arithmetic for the encoding/decoding
 as long as the attacker is not able to do a factorisation into the
two prime numbers, RSA is assumed to be secure
Adi Shamir, Ron Rivest and Len Adleman

Public Key (Asymmetric) Cryptography ...
 A drawback of asymmetric public key cryptography is the
fact that the algorithms are much slower than symmetric
ciphers
 Hybrid solutions combine public key with symmetric key
cryptography
 the public key encryption is only used in the setup phase to
securely exchange a pair of symmetric keys
 afterwards a secure channel is established based on the
symmetric keys
 Security of public key cryptography?
 new developments (e.g. quantum computing) might break public
key cryptography

Digital Signatures
 A digital signature can be used for two purposes
 to prove the authenticity of a message
 to guarantee that a message has not been changed during the
transfer (integrity)
 Sender creates a plaintext digest, encodes it with the
private key and adds it as a signature to the message
 the receiver creates the same digest and compares it with the
decoded signature
cipher
cipher
private key A public key A
plaintext plaintextplaintext
signature
digest
digest digest
same?
A B

Digital Certificates
 Information about a
person/company that is
digitally signed by a
certificate authority (CA)
 owner's name
 validity time
 signature of the CA
 owner's public key

Digital Certificates ...
 No single standard but most certificates store their
information in the X.509 v3 certificate standard form
 Basically digital certificates can be used on the server
side as well as on the client side
 in practice client-side certificates are not often used

HTTP Secure (HTTPS)
 Secure version of HTTP
 combines HTTP with asymmetric, symmetric and certificate-
based cryptography
 HTTP sent over TLS/SSL
 HTTPS protocol is selected by the https:// URL prefix
 Browser connects to the HTTPS default port (port 443)
 Initial SSL handshake
- negotiate protocol versions
- negotiate common cipher
- authentication
- generate temporary symmetric session keys

Email Security
 Emails are generally sent as unencrypted plain text
 An email is stored on multiple intermediary servers
before reaching its target
 relatively easy to intercept
 would you also put anything you write in an email on a postcard?
 Note that the sender of an email can easily be faked
 If we want to fix these problems we have to use third-
party tools such as Pretty Good Privacy (PGP)
 privacy
- strong encryption
 authentication
- digital signatures

Email SPAM
 Abuse of an electronic messaging
system (email) to deliver unwanted messages
 A major part of all SPAM is sent by only a few hundred
spammers
 It is estimated that SPAM costs businesses more than
100 billion dollars per year
 SPAM is illegal in many countries and some spammers
have already been sentenced to jail
 "Solutions"
 SPAM filters
 micropayments for emails

Email SPAM ...
 Phishing attacks
 send emails that look like coming from an official authority
and contain a request for sensitive data (e.g. password)
 send emails with links to websites that look like official companies
(e.g. your homebank)
 Spammers often use botnets to send their SPAM

Botnets
 Computers infected by malicious software become part
of a large botnet that can be remotely controlled
 the largest botnets contain more than 1 million machines
 An attacker can buy part of such a botnet to perform
various harmful tasks including
 the distribution of SPAM
 distributed denial of service attacks (DDOS)
 Distributed denial of service attacks are a very powerful
weapon as it has for example been shown when Estonia
was attacked in May 2007
 cannot easily be detected and filtered by firewalls since the traffic
is created by many different machines

Firewalls
 Software and hardware firewalls introduce artifical
"bottlenecks" that have to be passed by all the traffic
 block specific ports
 filter and block content
 protect private intranets from incoming Internet traffic
- often only a subnetwork (demilitarised zone) is connected to the Internet
Internet
Client Server
Firewall

Privacy
 While users access information over the Internet,
there is a continuous logging of their requests
 Each server stores information about clients who
accessed specific resources
 Data mining techniques can be used to combine this
logging information and create user profiles
 can for example be used for user-targeted advertising
 Users also "deliberately" publish personal information
 e.g. on Facebook
 Published information often cannot be easily deleted
 e.g. still accessible via Internet Archive (http://www.archive.org)

Web Log
 Log entry created every time a web server is accessed
 A log entry typically contains information about
 IP address of the requesting machine
 accessed URL
 request time
 refer link (previous page accessed by the client)
- sent as part of the HTTP Request
 browser type
 errors that occured
 ...

Web Log
 Web logs can be combined with other information
 e.g. login information can be used to reveal a user's identity
 Refer link
 enables access to potentially private information
 e.g. if previous request was an HTML form request using the GET
method then all the data will be available as part of the URL
XXX.XXX.XXX.193 - - [02/Dec/2009:05:50:40 +0100] "GET /knives-shun-c-81_114-l-en.html?gclid=CLOFucf5tp4CFc5L5Qod8jQzpA HTTP/1.1" 200 65478
"http://guelph.kijiji.ca/f-Shun-Classifieds-W0QQKeywordZShunQQisSearchFormZtrue" "Mozilla/5.0 (Windows; U; Windows NT 5.1; en-GB; rv:1.9.0.15)
Gecko/2009101601 Firefox/3.0.15"
XXX.XXX.XXX.116 - - [02/Dec/2009:05:50:42 +0100] "GET /images/Jamie%20Oliver/flavourShakerSchwarz.jpg HTTP/1.1" 200 3594
"http://www.tenera.ch/kenwood-pasta-roller-at970a-for-lasagne-base-unit-p-1314-l-en.html" "Mozilla/4.0 (compatible; MSIE 7.0; Windows NT 5.1;
GTB5; .NET CLR 1.1.4322; MS-RTC LM 8; .NET CLR 2.0.50727; .NET CLR 3.0.4506.2152; .NET CLR 3.5.30729)"
XXX.XXX.XXX.139 - - [02/Dec/2009:05:52:19 +0100] "GET /stylesheet.css HTTP/1.1" 200 10185 "http://www.tenera.ch/kai-seki-magoroku-redwood-
nakirimesser-165-cm-p-1433-l-de.html" "Mozilla/4.0 (compatible; MSIE 6.0; Windows NT 5.1; SV1; Mozilla/4.0 (compatible; MSIE 6.0; Windows NT 5.1;
SV1) )"
XXX.XXX.XXX.139 - - [02/Dec/2009:05:52:19 +0100] "GET /kai-seki-magoroku-redwood-nakirimesser-165-cm-p-1433-l-de.html HTTP/1.1" 200 60636
"http://www.google.ch/search?hl=de&source=hp&q=seki+magoroku&meta=&aq=0&oq=seki+ma" "Mozilla/4.0 (compatible; MSIE 6.0; Windows NT 5.1; SV1;
Mozilla/4.0 (compatible; MSIE 6.0; Windows NT 5.1; SV1) )"
XXX.XXX.XXX.139 - - [02/Dec/2009:05:52:21 +0100] "GET /images/pixel_trans.gif HTTP/1.1" 200 43 "http://www.tenera.ch/kai-seki-magoroku-redwood-
nakirimesser-165-cm-p-1433-l-de.html" "Mozilla/4.0 (compatible; MSIE 6.0; Windows NT 5.1; SV1; Mozilla/4.0 (compatible; MSIE 6.0; Windows NT 5.1;
SV1) )"
...
web log with refer links

Web Log File Analysis
 Site owner can use
various tools to analyse
the log files
 e.g. Webalizer
 How much information do
we give away when
accessing a website?
 What is happening with the logged data?
 combined with other information to reveal IP addresses?
 combined with log files from other sites?
- user profiling
 intended use of data should be mentioned in the privacy policy

Cookies Revisited
 Persistent cookies can be used to track a
user over time
 similar to IP address but more precise
 Third-party cookies can be used to build an anonymous
user profile
 if a website contains elements that have to be accessed from
another server (e.g. banner ads), then the server can set a cookie
- the third-party server creates a unique resource URL for every page on which
the resource has been embedded
- the user can be tracked on any site that uses the same service (e.g. banner
ads) and an anonymous user profile can be created
 Cookies should not be used for authentication
 can be modified by a user to forge identity (cookie poisoning)

Web Bugs
 User tracking based on the same idea as
with third-party cookies
 Embed a small object (e.g. 1 pixel image) in a webpage
and get informed every time the webpage is accessed
 request containing the IP address is sent to the server
 The web bugs approach cannot only be used for
webpages but also for other resources such as email,
Word documents etc.
 if the user reads an email containing an embedded HTML web
bug, the server knows when the email has been read but also
gets information about the IP address of the mail client

Other Services with Privacy Issues
 Google Earth shows a lot of sensitive information
 e.g. military bases etc.
 Google Street View shows not only streets and buildings
but also citizens
 privacy of individuals might be violated since they are shown at
strange places or in weird situations
 since the blurring of faces and number plates does not always
work, some countries would like to stop the service
 Many other free services from Google as well as other
companies harvest personal information and use it, for
example, for customer-targeted advertising

Video: Google Analytics

Google Analytics
 Very nice tool for web administrators to analyse their
web traffic
 easy to "install" over the Web
 website administrators have to add a piece of JavaScript code to
their website
- similar to web bug approach shown earlier
 Google gets information about site visitors
 While a user can normally choose to use a free service
(e.g. Gmail) or not, the user has no choice when it
comes to the tracking via Google Analytics
 How save is the captured data?
 what if somebody manages to steal the data?

Exercise 11
 Security

References
 David Gourley et al., HTTP: The Definitive
Guide, O'Reilly Media, September 2002
 Google Analytics Video
 http://www.youtube.com/watch?v=rHeKRvo6OhI
 R.L. Rivest, A. Shamir and L. Adleman, A Method for
Obtaining Digital Signatures and Public-Key
Cryptosystems Authentication, Communications of the
ACM, February 1978

2 December 2005
Next Lecture
Future Trends and Summary

Security, Privacy and Trust - Lecture 11 - Web Information Systems (4011474FNR)

More Related Content

Similar to Security, Privacy and Trust - Lecture 11 - Web Information Systems (4011474FNR) (20)

More from Beat Signer (20)

Recently uploaded (20)

Security, Privacy and Trust - Lecture 11 - Web Information Systems (4011474FNR)