Dresden Đại học Công nghệ Đại học Erlangen-Nuremberg Cơ sở dữ liệu Tập đoàn Công nghệ cơ sở dữ liệu hệ thống 01062 Dresden, Đức Martensstr. 3, 91058 Erlangen, Đức cơ sở dữ liệu lehner@ schlesinger@ TÓM TẮT cảm biến đang được triển khai rộng rãi ứng dụng để phát hiện, đo lường và giám sát. Các dòng dữ liệu kết quả thường được thu thập trong một hệ thống cơ sở dữ liệu cảm biến tập trung tổ chức các dữ liệu thuộc cùng một chủ đề hoặc vị trí tương tự. Nó được phổ biến thực hành để. | Databases and Sensor Networks Copyright 2004 CRC Press LLC Querying Asynchronously Updated Sensor Data Sets under Quantified Constraints Lutz Schlesinger Wolfgang Lehner University of Erlangen-Nuremberg Dresden University of Technology Department of Database Systems Database Technology Group Martensstr. 3 91058 Erlangen Germany 01062 Dresden Germany schlesinger@ lehner@ ABSTRACT Sensor databases are being widely deployed for measurement detection and surveillance applications. The resulting data streams are usually collected in a centralized sensor database system holding the data belonging to the same subject or the same location. It is common practice to evaluate the data only locally. To discover new results modern requirements imply to combine data at different sensors focussing on different subjects. The natural way of realizing a homogenous view and a centralized access point to a distributed set of database systems is the concept of a federated or a data warehouse system. However the tremendously high data refresh rate disallows the usage of these traditional replicated approaches with its strong consistency property. Instead we propose a system that handles packaged updates to reduce the replication frequency and deals with weak consistency properties. Such a system may be based on grid computing technology This technique mainly used in the number crunching area provides an alliance of autonomous local computing systems working together to build a virtual supercomputer. In this paper we combine the grid idea with sensor database systems. A query to the system is sent to one of the grid members where each node holds cached data of other grid members. Due to the nature of sensor database applications cached information may become stale fast. In order to speed up query processing the access to outdated data may sometimes be acceptable if the user has knowledge about the degree of staleness or the degree of .
