Fault Tolerant Computing
Project Ideas
You must demonstrate
that you know how to research, are current in the chosen topic, have developed
significant expertise in it and can communicate effectively about your work.
Some thought on how to do research are here.
Requirements: see Term Project
Evaluation: Among the
factors that will be used for evaluation are: significance and originality,
thoroughness of research, depth of understanding displayed and presentation.
Topics: The topics
must be closely related to the course objectives. Here a list of suggested
topics.
Some possible ideas: These
are related to some of the problems that either we have worked on recently or
have thought about. You need to do your own research, however a related paper
is mentioned to serve as a starting point. Note that in addition to the
well-known ideas, you must become familiar with the current state of technology
and must have some novel ideas.
- Economic modeling of vulnerability markets A
related article, another.
- Computation of the quantitative reliability of a
distributed file system. A related article.
- Software test coverage models: Compare models that
have been proposed to relate test coverage and defects found. A
related paper.
- Impact of correlated failures on reliability. Some
related articles mine, Space
Shuttle O rings, wall-street
melt-down.
- Limits of practically achievable software reliability
or defect density. A
related article.
- Investigate the economics of software testing using
actual field data. Related
paper.
- Develop model for economic tradeoffs due to security
issues using actual data. Related
paper.
- Investigate quantitative methods for evaluating and
predicting human reliability using actual data. Related
paper, another.
- Investigate reliability evaluation in the presence of
correlated failures in software/hardware using data and analytical
methods. Related
paper.
- Develop models for risk evaluation due to
vulnerabilities, incorporating discovery and remediation. Related paper.
- Obtain an elegant algorithm for generating antirandom vectors, or prove that existing procedure
is an algorithm. Related
paper.
- Investigate antirandom
testing where the internal nodes have antirandom
values. Related
paper.
- Quantitative measurement of the impact of
security breaches. Related
article and another
article.
- Methods for quantitative modelling of corruption in
society. Related
paper..
- Estimation of RPO, RTO so that a proposed disaster
recovery scheme can be optimized Related
paper, another
paper
- Modelling the advantage of geographically
dispersed servers for disaster recovery. Related
paper. Related paper.
- Computation of the risk of security breaches:
Consider technology,
probabilities and costs.
Note that "risk" includes both the probability of an event and
the impact.
- Using fuzzing to discover zero-day vulnerabilities a related article, and another.
- Hypervisor security A related
article.
- Byzantine Fault-Tolerance A related article.