PENG9630 Internet Architecture and Measurements Course description

This course will give the student insight into the different parts that comprise the internet's architecture and how one can monitor, assess and characterise them. This involves a diverse set of topics that includes but is not limited to routing and addressing, content distribution, data centre networks, key services and application such as DNS and web and mobile broadband. The course will focus particularly on quantification of the robustness and reliability of the internet's architecture and services. Furthermore, the course will draw upon new advancments in the fields of machine learning and network science to extend and expand the toolset available for anlayzing Internet measurements.

The course will be offered once a year, provided 3 or more students sign up for the course. If less than 3 students sign up for a course, the course will be cancelled for that year.

None.

None.

Knowledge

On successful completion of the course, the student:

• has an overview of the different elements that comprise the architecture of today’s internet.
• has a good understanding about the approaches for conducting internet measurements and the latest advances in this field.
• be familiar of a broad set of tools that can help analyzing Internet measurments. Of a particular relevance here are tools that originate in other disciplines like Machine Learning and Statisitcal Physics. This will not only expand the available toolset but also increases the potential for interdisciplinory collaboration going forward.

Skills

On successful completion of the course, the student can:

• plan and carry out state-of-the-art measurement tasks
• can formulate research questions on the robustness and performance of operational networks, and design measurements for evaluating these questions.
• will have a general practical understanding of how different parts of the internet's architecture interplay to offer a performant end-to-end service.

General competence

On successful completion of the course, the student can:

• participate in debates and present aspects of his/her expertise in a way that promotes such discussions.
• drive innovation

Module 1 will take the form of lectures. Module 2 will take the form of lab and homework assignments. Module 3 will take the form of seminars. In module 3, the student will present a case to the other students. We will also invite guest lecturers from research groups that focuses on machine learning and network science to introduce the students to potential tools and analysis methods.

Practical training

The students will participate in lab experiments to explore how once can measure various aspects of internet's robustness and performance. The students will write a summary of one of the tools that were introduced in the lab and discuss its benefits and limitations.

None.

Both the presentation of the case in Module 3 of the course and the tool summary document in the practical training part the course will form basis of assessment.

Both exams must be passed in order to pass the course.

The oral presentation cannot be appealed.

All aids are permitted.

Pass or fail.

Students taking the course must have a thorough knowledge of advanced calculus, including ordinary and partial differential equations. The student should also be familiar with linear algebra and Fourier and Laplace transform theory. In terms of programming, the candidate should have some experience in implementing numerical methods, including schemes for solving partial differential equations.

The candidate should also have a certain knowledge of mathematical analysis, modern physics or physiology – depending on specialization.

The course will be offered once a year, provided 3 or more students sign up for the course. If less than 3 students sign up for a course, the course will be cancelled for that year.