Managing and querying heterogenous and evolving data

近年来，现代信息系统在根本上发生了变化。与传统的客户端-服务器结构的数据仓库不同，现代信息系统由大量的分布式数据源组成，所有这些数据源都通过互联网连接。除了分布式之外，这些数据源也在不断演变，它们的模式也在频繁和异步地更新。今天的信息系统也是异构性的。除了传统的数据库之外，出现了新的数据源。无论是实时流数据，数据服务，还是数据服务，传感器和射频识别(RFID)读数。*建议的研究计划是开发在存在异步模式更新的情况下具有弹性的用户应用程序的技术。*我们将开发允许用户应用程序和查询自动适应异类数据集的不断演变的模式的技术。这些技术还将使任务关键型应用程序能够以最小的限制限制远程数据库的发展，以确保正确和不间断的数据访问。*建议的技术将考虑到新型数据源的独特特征，如Web服务、数据流、和传感器/RFID数据。我们的计划还将研究分布式信息系统中异构数据的自主存储和交换问题。我们的目标是为非结构化或半结构化的数据集或数据流设计高效的存储和交换布局。通过分析要存储或交换的数据，我们将开发算法来自动设计存储和交换布局，高效地利用现有的数据库系统和网络资源。除了静态数据集，我们特别感兴趣的是针对不断演变的数据集和数据流的数据存储和交换问题。

In recent years, modern information systems have evolved in fundamental ways. As opposed to the traditional client-server architecture of data warehouses, modern informations systems consists of large volumes of distributed data sources, all connected by the Internet.  Aside from being distributed, the data sources are also evolving, with their schemas updated frequently and asynchronously.  Today's information systems are also heterogeneous.  Beyond traditional databases, there is an emergence of novel sources of data.  They include real-time streamed data, data services, sensor and radio frequency identification (RFID) readings.  The proposed research program is on developing techniques for resilient user applications in the presence of asynchronous schema updates.  We will develop techniques to allow user applications and queries to automatically adapt to evolving schema of heterogeneous data sets.  These techniques will also enable mission critical applications to limit the evolution of remote databases with minimal restriction to ensure correct and uninterrupted data access.  The proposed techniques will take into account the unique characteristics of novel data sources such as web services, data streams, and sensor/RFID data.  The proposed program will also study the problem of autonomic storage and exchange of heterogeneous data in distributed information systems.  Our objective is to design efficient layouts for storage and exchangefor unstructured or semi-structured data sets or data streams.  By analyzing the data to be stored or exchanged, we will develop algorithms to automatically design the storage and exchange layout that efficiently make use of available database systems and network resources.  Aside from static data sets, we are particularly interested at data storage and exchange problem for evolving data sets and data streams.