Master Thesis Subjects

This is a (non-exhaustive) list of suggestions for master's thesis projects I would be willing to supervise.

Conditional probabilities in PRISM

We have a theoretical framework for computing conditional probabilities over Markov Chains and Markov Decision Processes (see this paper). A master's thesis subject would be to implement the described model checking algorithm and apply it to several case studies.

Attacking anonymizing networks

The anonyming network Tor works by rerouting TCP connections through a network of approximataly 1000 Tor routers. By encrypting the communication in several layers, called onion skins, anonymity of the communication is guaranteed. However, because of its low-latency nature, Tor can be attacked by analysing traffic patterns. By selectively moving bursts of traffic through the Tor network and by remotely measuring the load on Tor routers, it is possible to statistically infer the route packages are taking, thereby breaking anonymity. For more information, see for instance (S.J. Murdoch, G. Danezis, Low-Cost Traffic Analysis of Tor, 2005). A possible master thesis project is to implement this attack strategy and assess its practicallity.

Analysis of the Dutch election protocol

Recently the procedure for elections in The Netherlands has come unto larger scrutiny. In particular, voting machines have come under attack for leaking leaking information about what people vote via radio-magnetic waves and for not allowing the public at large to oversee the election process.

Dutch laws, or to be precise "de Kieswet", "het Kiesbesluit", and "aanvullende ministerieële regelingen", describe the whole protocol for holding elections. This ranges from the protocol to register political parties, through the actually happenings on election day, to the official establishment of the results two weeks after the elections by "de Kiesraad".

A possible master thesis project is to analyse attacks against this election protocol. This could, for instance, take the form of a (large) attack tree. In this form, this master thesis project is best suited for "Informatiekunde".

Measuring security

The ISO standards 27001 and 27002 specify (ISO 27001) what an organization should do to keep its information systems secure and give recommendations in the form of best practices (ISO 27002) how the specification can be implemented. For instance, in the section on ''User responsabilities'', ISO 27001 specifies the objective

To prevent unauthorized user access, and compromise or theft of information and information processing facilities.

One of the controls that ISO 27001 subsequently specifies is ''Password use'':

Users shall be required to follow good security practices in the selection and use of passwords.

ISO 27002 gives guidelines how this can be achieved:

All users should be advised to
1. keep passwords confidential;
2. avoid keeping a record [...];
3. chance passwords [when needed];
4. select quality passwords [...];
5. change passwords [regularly];
6. change temporary passwords [immediately];
7. not include passwords in any automated log-on process [...];
8. not share individual user passwords;
9. not use the same password for business and non-business purposes.

What is still lacking is a way to check if users are indeed keeping their passwords confidential, don't keep records, change passwords regularly, select quality passwords, etc. Other standard do suggest how security can be measured. For instance, the forthcoming standard ISO 27004 is supposed to address security metrics and also the NIST document 800-55 addresses ways to measure security. For instance, with regard to passwords, NIST 800-55 has the following example.

Critical element: Are data integrity and validation controls used [...]?
Subordinate question: Are procedures in place to determine compliance with password policies?
Metric: Percentage of systems that perform password policy verification.
Purpose: To determine whether procedures are performed [...].
Implementation evidence:
1. How many systems are there [...]?
2. Is a password policy documented?
3. If yes, how many systems test passwords for policy compliance?
4. Which [...] methods are used to test passwords [...]?
5. If passwords are checked, how often [...]?
Formula: [Question 3 divided by Question 1].

The goal of this project is to describe metrics for each of the requirements in ISO 27001 and for all (appropriate) guidelines in ISO 27002. For some metrics, the above mentioned NIST publication will be helpfull; the literature has many examples (for instance the book ``Security Metrics'' by Andrew Jaquith is a very rich source of metrics); and some metrics you will have to device yourself.

Contact

If you are interested in any of these, or related, master thesis subjects, feel free to email me at petervr@cs.ru.nl me or drop by in HG02.069.