This is a recording of Global CENTRA Webinar on 24 January 2018 (at 1-2pm UTC), presented by Prof Christopher Stewart, ReRout Lab, The Ohio State University. The webinar *series is hosted by CENTRA Project (globalcentra.org, supported by US NSF ACI Award 1550126), headquartered at the **ACIS Lab, University of Florida.
00:00 Welcome (all the 33 attendees from the U.S., Taiwan, Portugal, Japan, Myanmar, Malaysia, the Philippines and China) by Jose Fortes, Chair of CENTRA Steering Committee.
01:48 Dr. Chris Stewart's talk begins
03:10 Larus' Virtuous Cycle - Edge Edition - trends and motivation underlined autonomous systems and how Edge Computing plays into that
11:14 Autonomous Systems
18:02 Edge computing
22:54 Research Directions
23:11 Q&A Part I - (1) different types of data and categorizing workloads; (2) capacity planning (slide #15): transient vs. fundamental
30:13 Platform Support: Characterize Autonomous Workloads
39:42 Platform Support: Cost-effective Parallelism [including Q&A Part II: are ground control measurements in place to figure out the depth and size of the plants at the time of capturing image (via the drone)?]
50:18 Platform Support: Efficient Code - Approximation for OLDI Services
57:30 Q&A Part III
[Abstract of the Presentation]
Cyber-physical devices, from image sensors to microphones to appliance controllers to consumer drones, are revolutionizing cities, businesses and homes. The era of autonomous systems is at hand. It is time to update research agendas across computer science, especially in computer systems. Challenges facing autonomous systems cut across traditional domains and introduce new metrics of merit that traditional evaluations can't capture.
This talk describes active research at ReRout Lab at The Ohio State University (web.cse.ohio-state.edu/~stewart.962/rerout.html) Our research aims to democratize autonomous systems, creating programming paradigms that lead to robust app stores. Our research also improves processors, operating systems and distributed platforms underlying autonomous systems. In our vision, infrastructure must provide sustained speedup even without Moore's Law. We have designed and implemented autonomous systems in the cloud and using unmanned aerial vehicles. We have also made seminal contributions for problems related to infrastructure and sustained speedup. While our solutions employ queuing theory, online algorithms, benchmarking and other established techniques in computer systems, we have learned key abstractions to combine these techniques effectively.
*For information of CENTRA's Spring 2018 Webinar Series (January throughout April 2018), please visit globalcentra.org/spring2018webinars/ for more information. Registration is available now.
**Advanced Computing and Information Systems Laboratory (ACIS Lab): acis.ufl.edu/
Follow us on Facebook: facebook.com/GlobalCENTRA and facebook.com/acis.lab