One decade XLDB

One decade XLDB
展开查看详情

1. Session 3 - Scaling Cloud to Big Data Dirk Duellmann, CERN XLDB 2017: 10th Extremely Large Databases Conference 11 October 2017 | Clermont-Ferrand, France

2. One decade XLDB • Since the first meeting at SLAC • formula still works • DB academics +
 
 Data intensive science +
 
 Data intensive industry

3. One decade XLDB • Since the first meeting at SLAC • formula still works • DB academics + 
 
 Data intensive science +
 
 Data intensive industry also one decade…

4. One decade of 
 data access “cloud style” • 10 years since amazon dynamo paper • Scaling can be more important than single cluster with strong consistency • CAP theorem, KV-stores, Hadoop and friends • Scaling stable deployment can be more important than ultimate performance • Good match for data analysis (stable, read-only) • additional challenge for other applications • Cloud style data access has fully arrived in science infrastructures

5. 3 different aspects • Dan van der Steer, CERN • Building Scale-Out Storage Infrastructures with RADOS and Ceph • Michael Mueller, Container Solutions • Cloud Native Databases • Mike Hildreth, University of Notre Dame • Data, Software, and Knowledge Preservation of Scientific Results