[Nrg-l] PhD Proposal Defense: Michael Ocean (Fri 2/29 @ 11am)

Azer Bestavros best at cs.bu.edu
Fri Feb 29 09:50:35 EST 2008

Reminder: In just about one hour... 



Computer Science Department
Boston University

Date: Friday February 29, 2008
Time: 11:00am
Place: Room MCS 135, 111 Cummington Street

snBench: Towards Modular, Functional Specification, Verification and
Development of Sensor Network Services

Michael Ocean


As the commoditization of sensors and actuators continues so does the
opportunity for increased Sensor Network (SN) deployments. Despite the
documented benefits of augmenting a variety of environments with SNs,
widespread adoption of SNs has yet to occur. While various solutions
have emerged to many individual SN-centric challenges (e.g., power
management, communication protocols, role assignment, code
dissemination, content-aware addressing) perhaps the largest obstacle to
widespread SN deployment is that those who wish to deploy and maintain a
SN lack the programming and systems expertise to do so. Many works have
emerged to provide reusable run-time components and programmatic
abstractions to ease SN development and deployment; these works
underscore the challenge of providing a system that is accessible to
novices without rendering the system powerless by over-simplification. 

As it is impossible to find a one-size-fits-all interface, we instead
recognize the value in tiered layers of abstraction to enable a variety
of high-level domain specific languages to be compiled to a common
(``narrow waist'') tasking language. This thin-waist tasking language
can in turn be re-translated to a variety of hardware platforms. The
goal of this thesis work is to provide (1) a common, SN tasking language
(Instruction Set Architecture) powerful enough to express complex SN
services while simple enough to be implemented on highly constrained
resources with predictable (soft) real-time execution, (2) a modular,
resource-aware SN run-time execution infrastructure, (3) a prototype
high-level language and compiler to illustrate the utility of the
tasking language, and (4) a type system to provide static verification
of program safety and the production of resource constraints. In this
talk I will present the guiding principles, salient features, and novel
components of this work. I will discus our current progress on this
path, and a road map for the work that is yet to be completed.


PhD Thesis Committee:

- Azer Bestavros
- Assaf Kfoury
- Ibrahim Matta
- Rich West

More information about the Nrg-l mailing list