CS155: Computer and Network Security

CS155: Homework #2

Spring 2014

Due: Thursday, May 29

Problem 1: HTTPS vs. Signatures

Suppose a software company widgets.com sells a product P and wants to distribute a software update D. The company wants to ensure that its clients only install software updates published by the company. They decide to use the following approach: The company places D on its web server and designs the software P to periodically check this server for updates over HTTPS.
  1. The company decides to buy a public key certificate for its web server from a reputable CA. Explain what checks P should apply to the server's certificate to defeat a network attacker. Is your design vulnerable to ssl strip?
  2. How would you design the program P and the company's web server so that the update is secure against a network attacker, but there is no need to buy a certificate from a public CA? Your design should use an HTTPS web server as before.
  3. Later on engineers at the company proposed the following very different design: Sign D using a widgets.com private key to obtain a signature s and then distribute (s,D) in the clear to all customers. The corresponding public key is embedded in the program P. The following two questions compare this new design to the design used in part (a):
    • If we want to distribute the patch D using a content distribution network like BitTorent, which of the two designs should we use? Explain why.
    • List the most time consuming crypto operations that the widgets.com web server does in each of the two designs? It suffices to only focus on pubic-key operations such as public-key encryption/decryption and digital signature generation/verification. Which of the two designs is more efficient?

Problem 2: Stealth Port Scanning

Recall that the IP packet header contains a 16-bit identification field that is used for assembling packet fragments. IP mandates that the identification field be unique for each packet for a given (SourceIP,DestIP) pair. A common method for implementing the identification field is to maintain a single counter that is incremented by one for every packet sent. The current value of the counter is embedded in each outgoing packet. Since this counter is used for all connections to the host we say that the host implements a global identification field.
  1. Suppose a host P (whom we'll call the Patsy for reasons that will become clear later) implements a global identification field. Suppose further that P responds to ICMP ping requests. You control some other host A. How can you test if P sent a packet to anyone (other than A) within a certain one minute window? You are allowed to send your own packets to P.
  2. Your goal now is to test whether a victim host V is running a server that accepts connections to port n (that is, test if V is listening on port n). You wish to hide the identity of your machine A and therefore A cannot directly send a packet to V, unless that packet contains a spoofed source IP address. Explain how to use the patsy host P to test if V accepts connections to port n.

    Hint: Recall the following facts about TCP:
    • A host that receives a SYN packet to an open port n sends back a SYN/ACK response to the source IP.
    • A host that receives a SYN packet to a closed port n sends back a RST packet to the source IP.
    • A host that receives a SYN/ACK packet that it is not expecting sends back a RST packet to the source IP.
    • A host that receives a RST packet sends back no response.

Problem 3: The HTML canvas element

The canvas HTML element creates a 2D rectangular area and lets Javascript draw whatever it wants in that area. Canvas is used for client-side graphics such as drawing a path on a map loaded from Google maps. For the purpose of the associated same-origin policy, the origin of a canvas is the origin of the content that created it. In the map example, the origin of the Javascript that creates the canvas is Google. Canvas lets Javascript read pixels from any canvas in its origin using the GetImageData() method.
  1. Canvas lets Javascript embed images from any domain in the canvas. Suppose a user has authenticated to a site that displays private information. Describe an attack that would be possible if Javascript from one domain could embed an image from another domain in the canvas and then use GetImageData() to read pixels from that image.
  2. How would you restrict GetImageData() to prevent the attack above?
  3. A canvas element can be placed anywhere in the browser content area and can be made transparent so that the underlying content under the canvas shows through. What security problem arises if calling GetImageData() always returned the actual pixels shown on the screen at that position?
  4. How would you design GetImageData() to defend against the vulnerability from part (c)? Propose a design that does not require the browser to test if the requested pixel is over content from another origin.

Problem 4: CSRF Defenses

  1. In class we discussed Cross Site Request Forgery (CSRF) attacks against web sites that rely solely on cookies for session management. Briefly explain a CSRF attack on such a site.
  2. A common CSRF defense places a token in the DOM of every page (e.g. as a hidden form element) in addition to the cookie. An HTTP request is accepted by the server only if it contains both a valid HTTP cookie header and a valid token in the POST parameters. Why does this prevent the attack from part (a)?
  3. One approach to choosing a CSRF token is to choose one at random. Suppose a web site chooses the token as a fresh random string for every HTTP response. The server checks that this random string is present in the next HTTP request for that session. Does this prevent CSRF attacks? If so, explain why. If not, describe an attack.
  4. Another approach is to choose the token as a fixed random string chosen by the server. That is, the same random string is used as the CSRF token in all HTTP responses from the server over a given time period. Does this prevent CSRF attacks? If so, explain why. If not, describe an attack.
  5. Why is the Same-Origin Policy important for the cookie-plus-token defense?

Problem 5: The trouble with non-prepared SQL statements

Consider the following (amateur) PHP script for a login page:
$username = $_GET[user];
$password = $_GET[pwd];
$sql = "SELECT * FROM usertable
            WHERE username = ’$username’
            AND password = ’$password’ ";
$result = $db->query($sql);
if ($result->num_rows > 0) { /* Success */ }
else { /* Failure */ }
  1. Give a value that an attacker can enter in the field called user that will result in a successful login? Assume the attacker does not know any usernames or passwords for the site.
  2. Suppose we change lines 1 and 2 to
    $username = addslashes($ GET[user])
    $password = addslashes($ GET[pwd])
    Recall that the addslashes PHP function adds a slash before every quote. That is addslashes("a'b") will output the string "a\'b". Explain why this prevents the attack from part (a).
  3. Does addslashes() completely solve the problem? Consider the GBK Chinese unicode character set. Some characters in GBK are single bytes while others are double bytes. In particular, the following table shows a few GBK characters:
    0x 5c = \
    0x 27 = '
    0x bf 27 = 뼧
    0x bf 5c = 뽜
    That is, the database interprets 0x bf 5c as a single chinese character. Show that using a simple addslashes() that adds a slash (0x5c) before every quote (0x27) as in part (b) leads to a SQL injection attack. What value of user will result in a successful login?

Problem 6: DNSSEC

DNSSEC (DNS Security Extensions) is designed to prevent network attacks such as DNS record spoofing and cache poisoning. Generally, the DNSSEC server for example.com will posses the IP address of www.example.com. When queried about this record that it possesses, the DNSSEC server will return its answer with an associated signature. If the DNSEC server is queried about a host that does not exist, such as doesnotexist.example.com, the server uses NSEC or NSEC3 to show that the DNS server does not have an answer to the query.

Suppose a user R (a resolver, in DNS terminology) queries a DNSSEC server S, but all of the network traffic between R and S is visible to a network attacker N. The attacker N may read requests from R to S and may send packets to R that appear to originate from S.

  1. Why is authenticated denial of existence necessary? To answer this question, assume that S sends the same unsigned DOES-NOT-EXIST response to any query for which it has no matching record. Describe a possible attack.
  2. Assume now that S cryptographically signs its DOES-NOT-EXIST response, but the response does not say what query it is a response to. How is an attack still possible?
  3. A DNSSEC server may send a signed NSEC response to a query that does not have a matching record (such as doesnotexist.example.com). An NSEC response contains two names, corresponding to the existent record on the server that immediately precedes the query (in lexicographic order), and the existent record that immediately follows the query. For example, if a DNSSEC server has records for a.example.com, b.example.com, and c.example.com, the NSEC response to a query for (non-existent) abc.example.com contains a.example.com and b.example.com because these come just before and just after the requested name. To be complete, NSEC records also wrap-around, so a query for a non-existent name after the last existent name will receive an NSEC containing the last and first existent names.

    How should the resolver use the information contained in NSEC records to prevent the attacks you described in previous parts of this problem?

  4. NSEC leaks information that may be useful to attackers on the Internet. Describe how an attacker can use NSEC to enumerate all of the hosts sharing a common domain-name suffix. How is this information useful for attackers?
  5. NSEC3 is designed to prevent DNS responses from revealing unnecessary information. NSEC3 uses the lexicographic order of hashed records, instead of their unhashed order. In response to a query without a matching record, NSEC3 will return the hashed names that are just before and just after the hash of the query. For example, on a server containing a.example.com, b.example.com, and c.example.com, if a hashes to 30, b to 10, c to 20, and abc to 15, the NSEC3 response to a query for abc.example.com would contain 10.example.com and 20.example.com. Hashed names are also assumed to wrap around, in the same way as unhashed names in NSEC.

    Explain how a resolver should verify the validity of a response under NSEC3?