Transactions have for decades provided the gold standard for writing data-driven applications. I will describe two scenarios where we want transactions, but where we can seemingly not really achieve them. In our first model, we want to make joint travel arrangements or jointly enroll in classes, i.e., two or more people want to cooperate to select a seat or a class. But a transaction gives the illusion of having the database by itself, preventing information flow. We show how to slightly change transactions to enable efficient cooperation.

In our second model, we can avoid cross-data center latencies when committing transactions in a distributed or replicated database system, such as across two data centers in Europe and the US. In our new model, we allow sites to be inconsistent during execution, as long as this inconsistency is bounded, thus avoiding costly round-trips.

This talk describes research done at Cornell University.

Johannes Gehrke is a Distinguished Engineer at Microsoft working on Delve, the Office Graph, and Big Data and Data Science in Office 365. Until 2015, Johannes was the Tisch University Professor in the Department of Computer Science at Cornell University where he graduated 24 PhD students. Johannes received an NSF Career Award, a Sloan Fellowship, a Humboldt Research Award, the 2011 IEEE Computer Society Technical Achievement Award, the 2011 Blavatnik Award from the New York Academy of Sciences, and he is an ACM Fellow. He co-authored the undergraduate textbook Database Management Systems (McGrawHill (2002), currently in its third edition), used at universities all over the world. Johannes was co-Chair of SIGKDD 2004, VLDB 2007, ICDE 2012, SOCC 2014, and ICDE 2015.