This year, the BigDataOcean project presided over the DEBS conference’s Grand Challenge, providing the data and platform for the participants to compete to win the Grand Prize of 1000 USD.
What is DEBS and the Grand Challenge?
On a yearly basis since 2011, the DEBS conference runs an event-processing-themed Grand Challenge for anyone who wishes to compete, be it research groups, industry, or anything in between. The purpose is to come up with a solution to the problem defined by the Grand Challenge organisers (this year, this role was played by BigDataOcean). Each year, the participants get to explore and familiarise themselves with a new problem and dataset, and get judged by a set of evaluation criteria: correctness of solution, throughput, and latency.
This year’s challenge was two-fold: given a data stream of comma-separated tuple representing AIS signals from various ships, a) predict their destination and b) predict their arrival time at their destination. The solution must give one prediction per AIS position, i.e., comma-separated tuple.
The Hobbit Project’s benchmarking system was utilised as the evaluation platform, and the Grand Challenge data was provided by MarineTraffic and hosted on the BigDataOcean platform.