This talk focuses on emerging challenges and opportunities related to the scalability of data access in the context of increasing convergence between big data analytics and high performance computing approaches. It emphasizes the importance of scalability not only from the perspective of performance, but also resource utilization. Several broader patterns are used to illustrate this point: (1) versioning-based data models; (2) on-demand broadcast; (3) wide transformations. The talk concludes with several open questions that reveal exploratory future work.
 
Speaker Biography:
Bogdan Nicolae is a principal research scientist with Huawei Research Germany. He specializes in scalable storage, data management and fault tolerance for large scale distributed systems, with a focus on cloud computing and high performance architectures. He holds a PhD from University of Rennes 1, France and a Dipl. Eng. degree from Politehnica University Bucharest, Romania.  He is interested by and authored numerous papers in the areas of scalable I/O, storage elasticity and virtualization, data and metadata decentralization and availability, multi-versioning, checkpoint-restart, live migration. He is also actively involved in the organization of international conferences (e.g., SC, IPDPS, HPDC, PPoPP, CCGrid, CLUSTER, CLOUD) and editing of journals (e.g., IEEE TCC) relevant to these areas.