AMPLab has followed up a strong showing at the ACM Eurosys conference in April (where our students were authors and co-authors on 5 out of the 28 papers, including the Best Paper and Best Student Paper award winners) with another 5 papers and a demo to appear at the highly-competitive 2013 ACM SIGMOD Conference in New York City in June.   See the AMPLab Publications Page for these and other recent results from the lab.   Congratulations to all of the authors!

The papers are:

1) A paper on Shark by Reynold Xin et al., highlighting how Shark enables high-speed SQL processing to be combined seamlessly with equally-high-speed machine learning.

2) A PIQL paper by Michael Armbrust et al., which lays out a taxonomy of “scale independent” queries and describes how important classes of queries that were previously consider unscalable can be made scale independent using Materialized Views and Incremental Precomputlation.

3) A paper by Peter Bailis et al., on a surprising technique for adding Causal Consistency on top of existing distributed data stores that do not natively support such a strong consistency guarantee.

4) An extension to the CrowdER work on hybrid computer/crowdsourced entity resolution by former visitor and soon-to-be AMPLab Postdoc Jiannan Wang et al.   This work extends CrowdER to consider transitive relationships in matching, which greatly improves efficiency.

5) A paper with Jan Schaffner and other colleagues at SAP on Robust Tenant Placement for multi-tenant in memory databases.

In addition, there will be a demo of Probabilistically Bounded Staleness (PBS), which originally appeared in VLDB 2012 (where it was invited for a special “Best of” issue in ACM TODS)  and has since been integrated into the Cassandra system.