For students interested in our graduate program in networking for a MS
or PhD degree:

I thank you for your interest in our networking research program at
RPI.  This note is intended to you more information about the
selection process. It is based upon REALISTIC data from the last
couple of years, and is about the best information I can give you at
this point. It should however give you enough information to decide
whether you should take a shot and apply or not.

We are a well-funded, growing group in the area of networking. Our
core competencies include the use of simulation and experimental
tools, and inter-disciplinary collaboration to attack high-impact,
next-generation research problems centered around the theme of traffic
management. We also specialize in quickly transitioning such
technology research to industry and standards bodies through
linux-based prototypes. Our goal is to build a top-class, high-impact
networking research group. This means that students who form the group
should be highly motivated, and independent top-class researchers. Our
selection process therefore reflects this search for top-quality
individuals who will lead our group to the next level.

Firstly, please note that since I am in the ECSE department, I DO NOT
offer financial aid to CS applicants (or applicants in other
departments) during the admissions process. But several CS students do
work with me later in their graduate program.

We consider students for TA/RA awards only after the admissions
process has been completed by the school. Further, please note there
are certain rough cutoffs we have in our initial screening process
which I mention below. If you are below the cutoffs on several
categories, you might frankly have trouble getting here with financial
aid (as I mentioned before, these are realistic numbers). This is also
because the school and department will prune the list of applicants
even before they reach me, and the department takes autonomous
decisions on TAs. 

If these cutoffs are not a problem, you should definitely take a shot
since the odds are at least not heavily stacked against you. You can
then contact Ms. Ann Bruno (ann@ecse.rpi.edu) for more details reg
paperwork for the applications process.

We will require students to satisfy both our TA and RA cutoffs; and
then we will offer aid (1/2 RA + 1/2 TA) to a prioritized list of
candidates. Our TA cutoffs are pretty high: GRE 2250+ and TOEFL
630+. Strong EE course performance is also considered (a minimum GPA
of 3.3/4 or 8.2/10 is usually required). We also look carefully at the
letters of recommendation for top 5-10% ratings (top 25% or less
ratings usually are considered a negative sign). 

For the 1/2 RA part, we are very focussed. First, we would more more
likely to give aid if the student shows a clear and unequivocal
interest in a PhD, and more so, if it is backed with evidence of prior
research work (preferably in networking). This is perhaps my
overriding criterion at this point, and I might choose not to bring in
students even if positions are open unless I find promising PhD
candidates based upon compelling evidence in the application files.

We then give a very high preference to candidates with a strong
background in all three of: networking, simulation/analysis, and
programming skills. Strong background (as evidenced by grades) in
probability, statistics, queueing theory, and any projects in
networking would be a big plus. Many of my projects are in the
Internet area and therefore TCP/IP knowledge, BSD Unix/Linux code
knowledge is something I specifically look for.

If you do meet these criteria, the remaining factors which will decide
selection are:
1) available funding (for RA and TA) and 
2) the level of competition you will encounter in peer applications.  

To give you a general idea of the numbers, this year we might have
about 3 or 4 positions open in networking, but we have about 75-100
applicants in the area.  However, we usually end up making at least 2
rounds of offers.

I hope this gives you a better idea of the internal review procedure
and competition. We tend to be very demanding in our admissions
process, but we reward successful students with an excellent graduate
experience in networking at RPI. Besides having 5 courses in
networking, 3-4 faculty members in the area and outstanding
lab/experimental facilities, we work closely with universities,
industry (both large and small/startup companies) and standards bodies
such as IETF and the ATM forum. Perhaps most important, I am
personally committed to the success and excellence of each of my
graduate students and will go the extra mile to help every one of them
succeed in their careers.

Please consider these issues carefully and realistically as you decide
to apply here. Due to sheer volume of email, I regret that I cannot
provide more information on this matter, and cannot respond to
follow on emails.

But rest assured that we will examine all qualified applicants very
closely and enthusiastically in early/mid-March 2000. All decisions
will be made purely upon verifiable information furnished in the
application, and not upon information mentioned in emails. So, please
take care in preparing your application to address the issues detailed
above.

Best of luck!
Shivkumar
[Updated: Nov 29th, 1999]