Information Flow Control for Cloud and Internet of Things
Abstract: Because "security concerns hinder cloud adoption", cloud design focus has so far been on strong isolation between tenants. Inter-tenant interactions are not foreseen, and finer-grain protection e.g. between end-users of services, is left to the tenants.Requirements for data sharing between related applications are already emerging, and will increase, as cloud services become part of the Internet of Things (IoT).The keynote will outline our recent work on Information Flow Control (IFC) for cloud computing. IFC extends traditional access controls with continuous, data-centric, application-independent, dataflow control. IFC makes fine-grained, protected data sharing a possibility, rather than the current "all or nothing" approach.We have implemented IFC in a Linux Security Module, suitable for cloud deployment. Work is in progress on integrating IFC with middleware for IoT.
Speaker Bio: Jean Bacon is Emeritus Professor of Distributed Systems at the University of Cambridge, and leads the Opera research group, focussing on open, large-scale, secure, widely-distributed systems. She is a Fellow of the British Computer Society and IEEE and was an IEEE-Computer Society Governing Body member. She is a founding steering committee member of IEEE IC2E (International Conference on Cloud Engineering) and was PC chair in 2014. Her current EPSRC funding is to investigate Information Flow Control for cloud computing.
Addressing the Big Data Challenge Posed by the World's Largest Telescope
Abstract: I will begin by giving a brief overview of the upcoming Square Kilometre Array radio telescope and the scientific questions it aims to answer. Following this I will give an overview of a basic signal processing pipeline used to detect Pulsars and Fast Radio Bursts. Using a discussion focusing on the vast data rates associated with such observations I will show that the storage of data produced by the SKA is not feasible. This will then be used as a motivation for real-time Big Data processing. I will introduce the Astro-Accelerate Project which focuses on using many-core technologies such as GPUs, FPGAs and Intel’s Xeon-Phi to enable real-time digital signal processing for world leading radio-telescopes. I will present two case studies from this project demonstrating how these large data streams can be processed in real-time using computational accelerators. I will conclude by comparing current results from CPUs, FPGAs, Xeon-Phi and NVIDIA GPUs.