Search

Regulated industries, such as Healthcare and Finance, are starting to move parts of their data and workloads to the public cloud. However, they are still reluctant to trust the public cloud with their most sensitive records, and hence leave them in their premises, leveraging the hybrid cloud architecture. We address the security and performance challenges of big data analytics using a hybrid cloud in a real-life use case from a hospital. In this use case, the hospital collects sensitive patient data and wants to run analytics on it in order to lower antibiotics resistance, a significant challenge in healthcare. We show that it is possible to run large-scale analytics on data that is securely stored in the public cloud encrypted using Apache Parquet Modular Encryption (PME), without significant performance losses even if the secret encryption keys are stored on-premises. PME is a standard mechanism for data encryption and key management, not specific to any public cloud, and therefore helps prevent vendor lock-in. It also provides privacy and integrity guarantees, and enables granular access control to the data. We also present an innovation in PME for lowering the performance hit incurred by calls to the Key Management Service. Our solution therefore enables protecting large amounts of sensitive data in hybrid clouds and still allows to efficiently gain valuable insights from it., Comment: 5 pages, 3 figures

Soft real-time applications require timely delivery of messages conforming to the soft real-time constraints. Satisfying such requirements is a complex task both due to the volatile nature of distributed environments, as well as due to numerous domain-specific factors that affect message latency. Prompt detection of the root-cause of excessive message delay allows a distributed system to react accordingly. This may significantly improve compliance with the required timeliness constraints. In this work, we present a novel approach for distributed performance monitoring of soft-real time distributed systems. We propose to employ recent distributed algorithms from the statistical signal processing and learning domains, and to utilize them in a different context of online performance monitoring and root-cause analysis, for pinpointing the reasons for violation of performance requirements. Our approach is general and can be used for monitoring of any distributed system, and is not limited to the soft real-time domain. We have implemented the proposed framework in TransFab, an IBM prototype of soft real-time messaging fabric. In addition to root-cause analysis, the framework includes facilities to resolve resource allocation problems, such as memory and bandwidth deficiency. The experiments demonstrate that the system can identify and resolve latency problems in a timely fashion., Comment: Submitted to the 29th Int'l Conference on Distributed Computing Systems (ICDCS 2009)