Bluecouts: A catternet Formation Bluecouts: A catternet Formation ergio González-Valenzuela 1 on T. Vuong 2 Victor C. M. Leung 1 1 Department of Electrical and Computer Engineering 2 Department of Computer cience The University of British Columbia Vancouver, Canada 4 th Workshop on Applications and ervices in Wireless Networks Boston, Massachusetts, UA August 2004 IEEE AWN 2004, August 10 th, 2004
Bluecouts: A catternet Formation Presentation Outline Introduction catternet formation through mobile processing Bluecouts: On-demand scatternet formation imulation results Conclusions IEEE AWN 2004, August 10 th, 2004
Bluecouts: A catternet Formation Introduction What is Bluetooth? Global standard for short-range wireless communications 10 meters @ 721 Kbps / 3 Mbps (Enhanced Data Rate - EDR) Enables data and voice communications - WPAN High level of hardware integration & low power consumption Ideal for PDAs, cell phones, etc. Economic impact: shipping currently estimated at 2 million Bluetooth-enabled products per week worldwide IEEE AWN 2004, August 10 th, 2004
Bluecouts: A catternet Formation Introduction Bluetooth Primitives ample Bluetooth catternet Bluetooth device discovers neighbouring devices via inquiry process Devices assume either a master or slave role. A master handles up to seven slaves in active communications connected in a star-shaped topology - piconet Piconets may be interconnected via bridges to create a scatternet IP Phone Tablet PC Cell Phone Piconet 1 PDA1 Internet Access Point Projector Printer PDA2 Laptop Computer Fax Piconet 2 IEEE AWN 2004, August 10 th, 2004
Bluecouts: A catternet Formation Introduction Issues Problem definition: How do we create scatternets efficiently? everal existing proposals: Bluestar, Bluemesh, Bluenet, Bluetrees, TFP, FP, DTC, etc. Assumptions often made by existing FPs: ynchronous start/operation: Devices are somehow able to initialize the protocol at the same time, and must wait for partial computations from other devices to proceed with own decisions All BDs must be within radio range of each other Additional BDs cannot join the scatternet at a later time IEEE AWN 2004, August 10 th, 2004
Bluecouts: A catternet Formation catternet Formation Through Mobile Processing Our proposed solution Limitations of existing approaches attributed to the communication models they employ We propose a novel mobile agent-based solution Contrary to what existing schemes do, we decouple device discovery from actual topology formation Eliminates constraints often seen in existing approaches: Protocol runs in a fully asynchronous fashion Absolute radio coverage among all BDs no longer a constraint catternet can grow dynamically IEEE AWN 2004, August 10 th, 2004
Bluecouts: A catternet Formation catternet Formation Through Mobile Processing Our proposed solution (cont d) Employed Wave as our mobile processing implementation tool to code light-weight mobile agents Wave s key features: An internal mechanism for automated spatial coordination of mobile agents (Track Layer) An scripting-like language that enables highly compact agent code, leading to reduced bandwidth consumption External interfacing, enabling the interpreter to utilize existing resources (e.g. Bluetooth HCI APIs) Other agent platforms can be used, but would require much more programming to accomplish the same objective IEEE AWN 2004, August 10 th, 2004
Bluecouts: A catternet Formation Bluecouts: On-Demand catternet Formation Agent spreading mechanism Wave agents spread through the existing links in a controlled fashion and recursively signal back the state of the last computation s outcome (false, done, true, abort), leading up to the further replication of the mobile process or its termination. WI Agent Echo WI Process start Process termination A 1 A 2 A 3 A 4 E 1 E 2 E 3 E 4 E 5 IEEE AWN 2004, August 10 th, 2004
Bluecouts: A catternet Formation Bluecouts: On-Demand catternet Formation Bluecouts in action Case 1: A BD is discovered by a master and becomes slave M9 (nothing to do) IEEE AWN 2004, August 10 th, 2004
Bluecouts: A catternet Formation Bluecouts: On-Demand catternet Formation Bluecouts in action Case 2: A BD is discovered by a slave and becomes master. Agents are launched in an attempt to reconfigure the new BD s role. M M M or M uccessful reconfiguration Unsuccessful reconfiguration (a scatternet is formed) IEEE AWN 2004, August 10 th, 2004
Bluecouts: A catternet Formation Bluecouts: On-Demand catternet Formation Bluecouts in action Case 3: Agents conduct a coordinated spatial depth-first search over a logical backbone (i.e. excluding leaf nodes) in an attempt to reconfigure the new BD s role. Piconet 4 M Piconet 2 M Piconet 3 M Piconet 5 M M Piconet 1 IEEE AWN 2004, August 10 th, 2004
Bluecouts: A catternet Formation imulation Results imulation parameters Periodic node arrivals Nodes are uniformly distributed 10, 20 & 40 square meter areas A Wave agent (204 bytes) fits in a single DM5 ACL packet (224 Bytes) 50 simulation runs per test area Reasonably large scatternets 200 nodes IEEE AWN 2004, August 10 th, 2004
Bluecouts: A catternet Formation imulation Results Piconets probed 16 14 12 10 8 6 4 2 0 10x10 20X20 40X40 Delay (sec) 0.35 0.3 0.25 0.2 0.15 0.1 0.05 0 10x10 20x20 40x40 10 30 50 70 90 110 130 150 170 190 10 30 50 70 90 110 130 150 170 190 catternet size (nodes) catternet size (nodes) Number of piconets probed Process completion delay IEEE AWN 2004, August 10 th, 2004
Bluecouts: A catternet Formation imulation Results laves per master 8 7 6 5 4 3 2 1 10x10 20x20 40x40 Diameter 100 10 10x10 20x20 40x40 0 1 21 41 61 81 101 121 141 161 181 1 10 30 50 100 200 catternet size (nodes) lave/master ratio catternet size(nodes) catternet diameter IEEE AWN 2004, August 10 th, 2004
Bluecouts: A catternet Formation imulation Results 35 25 Number of piconets 30 25 20 15 10 5 0 10x10 20x20 40x40 Bandwidth (Kilobytes) 20 15 10 5 0 10x10 20x20 40x40 1 11 21 31 41 51 61 71 81 91 101 111 121 131 141 151 161 171 181 191 10 30 50 70 90 110 130 150 170 190 catternet size (nodes) Total number of piconets catternet size (nodes) Consumed bandwidth IEEE AWN 2004, August 10 th, 2004
Bluecouts: A catternet Formation imulation Results Discussion Changes to the Baseband specification are transparent to our scheme: agents employ APIs available at the nodes Existing schemes greatly emphasize on results at the Baseband layer: incompatible performance metrics deem a direct comparison mostly impractical Topology optimality not degraded as scatternets grow: slaves/master ratio performance is comparable or better Bandwidth consumption very low and linear ecurity issues attributed to mobile agents are less of a concern here: the interpreter lives around L3, not L7 IEEE AWN 2004, August 10 th, 2004
Bluecouts: A catternet Formation Conclusions First mobile agent-based scatternet formation protocol (to our knowledge) Agent approach helps decouple scatternet formation from device discovery, which greatly facilitates the topology reconfiguration process Agent approach enables fully asynchronous protocol operation and helps to eliminate constraints observed in existing schemes Programmable approach introduces unmatched flexibility by allowing context-aware topology formation IEEE AWN 2004, August 10 th, 2004
Bluecouts: A catternet Formation Acknowledgements This research is supported by the National ciences and Engineering Research Council of Canada (NERC) through grant TPGP 257684, the Canadian Institute for Telecommunications Research (CITR) Network of Centres of Excellence, and the Advanced ystems Institute (AI) of British Columbia. Thanks to the anonymous reviewers of our paper! Thank you! IEEE AWN 2004, August 10 th, 2004