Information Centric Networking at MIT: Summer UROPS, 2013
Our research is in the area of a novel approach to the design of the Internet called “Information Centric Networking”, which assumes that the concept of “location” can be eliminated. The belief is that this will ease problems of both multi-party communication and mobility. At the basis of the of the approaches taken are mechanisms for identifying information to make it available and request instances of it, as well as a variety of designs for delivering the requested information. Our work focuses on two directions. The first is the question of identification of the information to have a particular set of characteristics – persistence and pervasiveness. We have begun a preliminary design using a simulation tool called ns-3 and plan to move to an experimental network to expand our ideas and experience. This UROP will be experimenting with two of the key designs in the NSF-funded, large-scale testbed called GENI.
The second is the question of how to evaluate these various designs, to figure out what works and what does not. The problem we will address this summer is that there are no uniform sets of data against which to evaluate all the different approaches. These datasets will be derived from existing data structures (such as Wikipedia) to bring reality to the studies, but will involve a combination of understanding the characteristics (size, frequency of pointers, etc.) to various parts of the data set as well as a canonicalization of the internal names, in order that they can be used in the different ICN systems, with a mapping to their own ids. This UROP will be building these large datasets and making them available for both our own work and that of the larger community.
Experience required: The first UROP will require systems programming skills and C++. 6.033 necessary. More advanced systems and programming courses a plus. The second UROP will require programming in Java or Python. 6.033 a plus. We are also considering ways to bring in a third UROP with less programming skills, but an interest and willingness to learn.
Funding: These two UROPs will be funded as National Science Foundation REU (Research for Undergraduate Opportunities) Fellowships, funded through MIT. The third will be through MIT Direct funding or the CSAIL Hackers Heaven Program.
Contact: Dr. Karen Sollins, firstname.lastname@example.org. Please send resume and a description of specific programming projects.
All candidates will be interviewed.