[Ns-developers] GSoC 2012 projects discussion
q5frc at unb.ca
Sun Mar 18 05:25:02 PDT 2012
I'm a second year Ph.D student in Canada and plan to apply several NS-3
projects in GSoC 2012 which are very close to my current research project.
Two questions here:
1, Are all projects in idea page available for student to apply? I know
that not all projects in idea page were accepted by NS-3 in 2009 and 2010.
2, Can a student apply several projects in NS-3 GSoC 2012? If so, can I
apply those projects in one proposal? Or I can must submit multiple
proposals for each project I'm interesting.
In the rest part of this email, I first state the motivation for me to
apply NS-3 GSoC. Then I list projects I'm interesting and how those
projects relate to my current research.
The main motivation for me to apply GSoC is that my research area
is very close to several projects in NS-3. Currently, I'm doing the
multipath video delivery optimization in wireless network. Specifically,
I focus on the multipath adaptive HTTP streaming and multipath transport
layer protocol (e.g., multipath TCP, SCTP). I both develop my solution
in NS-2 and linux testbed now. Based on above research topic, currently,
I'm developing the adaptive HTTP streaming traffic model in NS-2.34 and
use multipath protocol in transport layer. Therefore, several projects
in NS-3 this year, such as SCTP, PackMime-HTTP and Tmix, are closely
related to my research work.
Another motivation is that I want to move my simulation from NS-2
to NS-3 and use NS-3 in my future research. I used NS-2 in my research
for at least 3 years and feel that the maintenance of NS-2 is a big
issue. Also, two features in NS-3 attracts me a lot:
a) multiple interface mobile node support
b) LTE support
Those two features are very helpful for my research now.
Interested projects list:
Although I'm doing the research on concurrent TCP now, I will extend my
algorithm to SCTP in my future ICC/WCNC 2013 paper. I think implementing
a standard SCTP based on related RFC drafts is enough now.
Currently, PackMime-HTTP in NS-2 does not consider video traffic. I
think we can consider this features in NS-3 because web-based video
increases a lot in recent years.
In the idea page, it said that the delaybox work is already started. So
do we need to coordinate the developer of DelayBox in NS-3 in GSoC? Or
we just develop delaybox independently?
4, NS-3 in cloud
This is a very interesting project. I'm also doing some research on how
to develop mobile program in cloud environment. Usually, two ways are
proposed to utilize the cloud resource for mobile program: 1) moving
part of computing model to cloud side and returning the results to
client. 2) directly call the API provided by cloud, e.g., HAWAII
projects in Microsoft. However, in idea page, it seems that we will run
all the NS-3 program in cloud. But the computing resource assignment is
usually done by cloud itself. So what specific missions can NS-3 do in
4, LTE scheduler
Because I'm doing the research about multipath transmission for video in
wireless network. Scheduler algorithm is a key part in my work. Although
this project is in MAC layer, there are still some common parts between
schedulers in the upper layer and MAC layer. However, there are varied
scheduling algorithms now. Personally, I have a suggestion that we can
develop a scheduler framework which can be easily further modified to
the specific new algorithm. This is because NS-3 is a general simulation
tool and should not focus on a non-standard algorithm in the main source
Looking forward to your suggestions.
Faculty of Computer Science
University of New Brunswick
540 Windsor Street
E. q5frc at unb.ca
More information about the Ns-developers