PITA Fiscal Year 2007 Projects

Innovative Infrastructure System Assessment Technology

Mead-Lite: Embedded middleware platform for the rapid, portable development and deployment of distributed sensing applications
When multiple heterogeneous sensors are connected together for critical sensing applications (such as envisioned for Sensor Andrew), there is clearly a need for them to communicate with each other, regardless of the differences in the underlying platform, operating system, location, target application and usage. The proposed Mead-Lite embedded sensor middleware consists of a suite of tools and libraries that mask the intricate and complex details (such as protocol differences, byte-order differences, concurrency, asynchrony, consistency, etc.) of enabling client and server components running on various sensor nodes to work together seamlessly. This transparency is of great benefit to application programmers, who are most often domain experts in the specific target application (e.g., sensing chlorine levels in water, determining leaks and cracks in pipes), but not experts in the issues of networking or distributed programming. The application programmer does not need to re-write the application to suit every new sensor platform/architecture – by simply conforming to the middleware programming APIs, the application developer can re-locate his/her sensor application, as is, to run similarly on any target platform that Mead-Lite supports.

Given CenSCIR’s focus on building an array of large-scale distributed sensing applications, potentially distributed across campus and involving different kinds of sensors and capabilities, there is a compelling case for the researchers involved in CenSCIR to exploit Mead-Lite to develop their applications. This would free them from having to worry about (i) porting their applications to work with every new sensor platform, (ii) remote communication and distribution of sensing capability and data across heterogeneous platforms , and (iii) the impact of failures on their applications. The anticipated contributions and benefits of this proposed research will include (a) the identification of middleware needs within the Sensor Andrew target scenarios, (b) the development of the Mead-Lite reliable embedded middleware platform and tool-chain support, and (c) an initial CMU-driven contribution to the OMG industrial standardization efforts for embedded fault-tolerant middleware.