6.896 Topics in Computer Networks
Tclme LCS
(6.0
5.0
6.0)
Lecturer: H. Balakrishnan
Next Term: Not Offered
Lecturer's Rating: 6.0/7.0
Prerequisites: Probability background, 6.033 or 6.853

Response rate: 24 out of 35
Difficulty: 6.5/7.0
Overall Rating: 6.0/7.0
Term Evaluated: Fall 98

Lecturer's Comments:
6.896 covers recent and classic research topics in computer networks
and the Internet. Students should take this course to learn how to do
research, especially in networking, to understand the stateoftheart
in network protocols, architectures, and applications, and to
investigate novel ideas in the area via semesterlong research
projects. The prerequisite for this class is a systems background
with exposure to networking, elementary probabilistic analysis, and
the ability to critique work.
Students in 6.896 "became familiar with the stateoftheart in
networks." The students agreed that this is a survey course that
teaches "just about everything in (Internet) networking
research... but not any one topic in any great detail." Students also
commented that the course does not teach the fundamentals, or
background material, such as probability, for instance. Most students
were very interested in the subject material.
6.856 is mostly made up of M.Eng and graduate course VI students, with
about equal VI1, VI2, and VI3 representation. Students agreed that
much more time is spent on practicalities, rather than on any rigorous
theory. However, they still praised the depth of content in 6.896.
Students found lectures very useful in learning the material, often
ranking them as the best way to learn. Readings and problem sets were
second and third, respectively.
Students recommend taking 6.896 to learn about current research topics
in computer networks. Students interested in computer architecture
and implementations will also be interested. The class covers a lot
of material, so students should have a solid background in computer
architecture and networks. Students also suggested knowing awk, perl,
tcl, and C++. The only consistent complaint that students have of the
course is that it tends to gloss over important points in its
surveystyle treatment of the material. "The content is very good,
but there is too much jumping around between different areas. It
would be better to narrow the content and go in some areas with more
focus."
Lecturer H. Balakrishnan received positive comments for
encouraging students to pursue their own research projects outside of
class. Students also praised his enthusiasm and wellprepared
lectures: "His lecture is very wellprepared and full of insightful
comments." Most students suggested that the lecturer slow his pace
down considerably; "he does lecture at a very fast pace." Students
also commented that his pace quickens later into lecture. A few
students felt that he did a good job at conveying the material, but
that there was too much material to fit into one lecture. Students complimented his good use of the blackboard.
Some students found the 6.896 problem set very useful for learning the
material; all found it very long. Students also complained about
administrative problems, such as delays in assigning the problem set,
"poor timing", and "not returning" the problem sets. Some students
did not approve of the content of the problem sets, saying they "did
not shed much knowledge," and "needed much background" to complete.
Students estimated that about one third of the work involved "mindless
grunge." There was only one problem set at the time of the review,
and students spent 2050 hours working on it. The problem set was
assigned in teams of two, which students seemed to appreciate:
"without the psychological support, it would have been hard."
Students did not use bibles, as 6.896 was a new course this
semester.
"Hari is a networking god!"
"Basically, what this class needs is more support staff and it has the potential to become a classic."
Dated: December 1, 1998
Eta Kappa Nu, MIT