Network Monitoring Tools


Sponsor Ling Liu / Henrique Paques
{lingliu, paques}@cc.gatech.edu
104a / 223 CCB
Area Systems and Databases


Problem

The explosive growth of the Internet made a vast amount of diverse information (data sources) available online, but increasingly difficult to access. Search requests issued by search engines, for example, sometimes takes too long to complete mainly due to the heavy network traffic. The network links are getting too crowed. However, search engines can improve their performance by using network monitoring tools that periodically reports the current conditions of the network links. Having such information available, in the presence of failed links, the search engine could adapt the search request in order to improve its performance under the changed network conditions.

We have two possible projects that investigate network monitoring tools:

(1) Extending a network monitoring prototype: We currently have a prototype of a network monitoring tool, written in Java, that implements a high level version of the command ping. In this project, you will have to substitute that version of ping by a lower level implementation. We can discuss how this implementation can be done.

(2)  Performance evaluation of existing monitoring tools: In this project you are to run experiments with some existing network monitoring tools and compare their performance. We can discuss how to construct the experiments and how to compare the results.

Background

You are expected to have a solid grasp of Java programming for the first project.

For both projects, you should be knowledgeable in network concepts.

Deliverables

For the first project: (1) documented source code and (2) a report on the design decisions.

For the second project: a report consisting of (1) description of each network monitoring tool used, (2) description of the experiments, and (3) performance comparison.

Evaluation

The first project will be graded based on the quality of the report and the source code.

The second project will be graded based on the relevance of the experiments and the performance comparison.