A Distributed Programming Infrastructure for Integrating Smart Sensors

A Distributed Programming Infrastructure for Integrating Smart Sensors Umakishore Ramachandran, Kenneth Mackenzie, Steve De. Weerth, Irfan Essa, Thad Starner College of Computing, Georgia Institute of Technology Problem: Octopus Apps! BASIC IDEA • emerging app class • tentacles: sensors, actuators • arms: data fusion, routing • head: cpu-intensive processing Stampede: Seamless Programming Sensor Stack: Application Data Fusion Layer Data Service Layer Medium Access Control BASIC IDEA • space-time memory • time-sequenced data streams • communication abstractions • channels, queues, registers • distributed garbage collection • computation as thread-channel graph Garbage collection 1 SAMPLE APPLICATIONS • distributed collaboration • aware spaces • smart environments • monitoring, control • surveillance • emergency response 2 3 4 5 6 7 Info Exchange Service Error Control Radio Control video Motion Mask Target Detection Histogram GOALS • in-stack fusion • logical naming • application-awareness Model 2 Location Registry Stampede share Programming Abstractions Producers. . . f() (actuators or other fusion channels) Fusion Module Resource Monitor, Routing Layer Operating System Hardware www Display Clients Elemental Events (Identity/Location) Data Events (Face, Moving Lip Detectors) Event Base Feature Extraction and Event Generators Media Streaming Engine Resource Management Feature Extraction Sensor Access and Management Media Warehouse Sensor Network “An Architecture for Event Web” Modahl, Bagrak, Wolenetz, Jain, Ramachandran IEEE FTDCS ’ 04, Suzhou, China (Minimize Transmission Cost) Placement Module www Results: Display Sources www Consumers S 1 S 3 www Ramesh Jain, Matthew Wolenetz Query Server Cost Function Filter D-Stampede Cluster Group Meeting Time: 10: 00 am-11: 00 am Location: CCB 201 Participants: Kishore Ramachandran, Applications Fusion Channel (a ‘Virtual Sensor’) (sensors or other fusion channels) simplified capture and rich access to structured media stores, organized around spatiotemporal events Experiential Event. Web Browser • federated data distribution • publish/subscribe model • internal data broker threads • type-lattice based transcoding Collage Workstations Domain Events (CS 6250 Lecture) Architecture Broker federation Task Graph • architecture and application • automates stream capture, feature extraction, correlation • identifies most related streams Event Web: Re-publish optimized fusion function placement in wired and wireless networks Stream Server transformation engine Model 1 Location DFuse: Web Results Client Media Capture Clients Info Collection derive Target Detection Video Display Client Functionality audio Video Frame Histogram Model Video Capture System Timing transform Digitizer Generic Workstation www Application: Smart. Kiosk People Tracker Change Detection distributed media analysis and correlation Location Service Media Broker: stream registration and S 2 REQUIREMENTS • support for physically distributed heterogeneous devices • easy access to compute-servers (clusters, grids) • diverse computation, communication and power capabilities • support for dynamic join/leave, registration, discovery • sophisticated stream management (fusion, type-based discovery, publication, discovery, filter framework) o_conn Key Radio, Sensors, Memory, CPU Stack Diagram . . . • seamless programming • across diverse hardware • of compelling applications • reveals middleware requirements Helper Service Layer OBSERVATIONS • concurrent apps • energy, net bandwidth constraints KT=2 FEATURES • distributed, pervasive infrastructure • of widely varying device capabilities • with control-loop flavor processing • on streams of varying bandwidths • requiring rapid response • at human perceptual speeds Approach: Smart Plumbing Access Routing Control, Attribute Filtering Scatter. Gather Translation, Persistent Medium. Access database Hardware i_conn 8 Deployment, Monitoring In-network Fusion Dynamic thread-channel graph domain channel thread TV Watcher: ROLE ASSIGNMENT • Naïve tree building • Optimization • Maintenance Testbed: IPAQ Farm “DFuse: A Framework for Distributed Data Fusion” Kumar, Wolenetz, Agarwalla, Shin, Hutto, Paul, Ramachandran ACM Sen. Sys ’ 03, Los Angeles, California 2003 “Media Broker: An Architecture for Pervasive Computing” Modahl, Bagrak, Wolenetz, Hutto, Ramachandran IEEE Per. Com ’ 04, Orlando, Florida Funded by NSF ITR/SY grant CCR-0121638 http: //www. cc. gatech. edu/~rama/ubiq-presence