Speaker: Ion Stoica
Professor, UC Berkeley; Co-founder, Databricks and Conviva
Title: Conquering Big Data with Spark and BDAS
Abstract: Today, big and small organizations alike collect huge amounts of data, and they do so with one goal in mind: extract "value" through sophisticated exploratory analysis, and use it as the basis to make decisions as varied as personalized treatment and ad targeting. Unfortunately, existing data analytics tools are slow in answering queries, as they typically require to sift through huge amounts of data stored on disk, and are even less suitable for complex computations, such as machine learning algorithms. These limitations leave the potential of extracting value of big data unfulfilled.
To address this challenge, we are developing Berkeley Data Analytics Stack (BDAS), an open source data analytics stack that provides interactive response times for complex computations on massive data. To achieve this goal, BDAS supports efficient, large-scale in-memory data processing, and allows users and applications to trade between query accuracy, time, and cost. In this talk, I'll present the architecture, challenges, results, and our experience with developing BDAS, with a focus on Apache Spark, an in-memory cluster computing engine that provides support for a variety of workloads, including batch, streaming, and iterative computations. In a relatively short time, Spark has become the most active big data project in the open source community, and is already being used by over one hundred of companies and research institutions.
Bio: Ion Stoica is a Professor in the EECS Department at University of California at Berkeley. He received his PhD from Carnegie Mellon University in 2000. He does research on cloud computing and networked computer systems. Past work includes the Dynamic Packet State (DPS), Chord DHT, Internet Indirection Infrastructure (i3), declarative networks, replay-debugging, and multi-layer tracing in distributed systems. His current research focuses on resource management and scheduling for data centers, cluster computing frameworks, and network architectures. He is an ACM Fellow and has received numerous awards, including the SIGCOMM Test of Time Award (2011), and the ACM doctoral dissertation award (2001). In 2006, he co-founded Conviva, a startup to commercialize technologies for large scale video distribution, and in 2013, he co-founded Databricks as startup to commercialize technologies for Big Data processing.
Speaker: Francois Baccelli
Professor, UT Austin
Title: Euclidean Content Delivery Networks
Abstract: The talk will present a new model for content delivery networking which accounts for the fact that the rate between two peers may depend on their distance.
In its simplest incarnation, the associated dynamics is that of a spatial birth and death process of the Euclidean space where the birth rate is constant and the death rate of a given point is the shot noise created at its location by the other points of the current configuration. The response function of this shot noise is determined by the way the rate depends on distance.
The unique time-stationary regime of this class of point processes will be derived by a coupling from the past argument. Two asymptotic regimes will be analyzed: the fluid regime and the hard-core regime. The results on these regimes are based on a mix of moment measure analysis and dimensional analysis and have important design implications. The first of them is super-scalability, namely a situation where the equilibrium mean latency is a decreasing function of the load.
Bio: François Baccelli is Simons Math+X Chair in Mathematics and ECE at UT Austin.
His research directions are at the interface of Applied Mathematics
(Probability, Stochastic Geometry, Dynamical Systems) and Communications
(Network Science, Information Theory).
He is co-author of research monographs on point processes and queues
(with P. Brémaud); max plus algebras and network dynamics (with G. Cohen,
G. Olsder and J.P. Quadrat); stationary queuing networks (with P. Brémaud);
stochastic geometry and wireless networks (with B. Blaszczyszyn).
He is a member of the French Academy of Sciences and part time
researcher at INRIA.
Speaker: Florian Simatos
Title: Performance analysis of wireless networks with mobile users
Abstract: In this talk I will discuss a model for a wireless network with a fixed infrastructure and mobile users. Because the network’s capacity is fixed, when the system is overloaded users move within the network much faster than they arrive to or depart from it, and the coexistence of these two time scales significantly complicates the analysis of such systems. I will explain how to establish fluid limits, and how this makes it possible to analyze the performance of fair bandwidth-sharing algorithms. In the last part of the talk I will also present results on heavy-traffic approximations. These results rely on a new method to establish heavy traffic approximations which I will present.
This talk will be based on joint works with S. Borst, A. Lambert and D. Tibi.
Bio: Florian Simatos is a researcher at Inria. His research lies in applied probability, with a strong emphasis on the performance analysis of communication networks. Prior to joining Inria, he did two post-docs: one at Eindhoven University of Technology and one at CWI (Amsterdam). He did his PhD at INRIA under the supervision of Philippe Robert. Together with A. Ganesh, S. Lilenthal, D. Manjunath and A. Proutiere he received the best paper award at the ACM SIGMETRICS 2010 conference.