RCRT : Rate-Controlled Reliable Transport
for Wireless Sensor Networks
Project Overview
Emerging high-rate applications (imaging, structural monitoring, acoustic localization) will need to transport large volumes of data concurrently from several sensors. These applications are also loss-intolerant. A key requirement for such applications, then, is a protocol that reliably transport sensor data from many sources to one or more sinks without incurring congestion collapse.
RCRT is "rate-controlled reliable transport" protocol that reliably transport sensor data from many sources to one or more sinks without incurring congestion collapse. RCRT allocates efficent data rates to each source nodes while guaranteeing 100% end-to-end reliable packet delivery.
The main components of RCRT are:
- Loss recovery: using end-to-end explicit NACKs and retransmissions
- Congestion detection: centralized, based on time to recover loss
- Rate adaptation: centralized, based on total aggregate traffic
- Rate allocation: centralized, based on application specific capacity allocation policy
Related Projects
People
Publications
Jeongyeup Paek, Ramesh Govindan, RCRT: Rate-Controlled Reliable Transport for Wireless Sensor Networks, In: Proceedings of the ACM Conference on Embedded Networked Sensor Systems (SenSys), Sydney, Australia, November 2007. [PDF] [Abstract]
Posters and Presentations
Jeongyeup Paek, RCRT: Rate-Controlled Reliable Transport Protocol for Wireless Sensor Networks, Sydney, Australia, November 2007. [PPT]
Jeongyeup Paek, Rate-Controlled Reliable Transport for Wireless Sensor Networks, University of California, Los Angeles, October 2007. [PPT]
Software
RCRT code can be found in Tenet software: ENL CVS
Results
- Rate r_i allocated to every node in the 40-node experiment with fair rate policy.
- Rate achieved by IFRC and RCRT in 30-node experiment
(Note: Some of the MAC layer configurations were different between IFRC and RCRT)
- Goodput achieved (vs. offered load) by best-effort transport and
reliable transport without congestion control. Y-error bar represents
maximum variation between nodes.
(40-node experiment)
- Goodput achieved (vs. requested rate) by RCRT. (40-node experiment)
Last Modified: 28 Aug 2007