Welcome to the ADMT Publication Server

Quality Contracts for Real-Time Enterprises

DocUID: 2007-004 Full Text: PDF

Author: Alexandros Labrinidis, Huiming Qu, Jie Xu

Abstract: Real-time enterprises rely on user queries being answered in a timely fashion and using fresh data. This is relatively easy when systems are lightly loaded and both queries and updates can be finished quickly. However, this goal becomes fundamentally hard to achieve due to the high volume of queries and updates in real systems, especially in periods of flash crowds. In such cases, systems typically try to optimize for the average case, treating all users, queries, and data equally. In this paper, we argue that it is more beneficial for real-time enterprises to have the users specify how to balance such a tradeoff between Quality of Service (QoS) and Quality of Data (QoD), in other words, ``instructing'' the system on how to best allocate resources to maximize the overall user satisfaction. Specifically, we propose Quality Contracts (QC) which is a framework based on the micro-economic paradigm and provides an intuitive and easy to use, yet very powerful way for users to specify their preferences for QoS and QoD. Beyond presenting the QC framework, we present results of applying it in two different domains: scheduling in real-time web-databases and replica selection in distributed query processing.

Published In: Lecture Notes in Computer Science 4365: Post Proceedings of First International Workshop on Business Intelligence for the Real Time Enterprise

Pages: pp. 143-156

Year Published: 2007

Note: BIRTE'06 was held in conjunction with the VLDB'06 Conference, Seoul, Korea, Sept. 2006

Project: UserCentric,   S-CITI Subject Area: Web Databases

Publication Type: Book Chapter

Sponsor: NSF IIS-0534531, NSF ITR ANI-0325353

Citation:Text Latex BibTex XML Alexandros Labrinidis, Huiming Qu, and Jie Xu. Quality Contracts for Real-Time Enterprises, Lecture Notes in Computer Science 4365: Post Proceedings of First International Workshop on Business Intelligence for the Real Time Enterprise (BIRTE'06), pp. 143-156, September 2007.(BIRTE'06 was held in conjunction with the VLDB'06 Conference, Seoul, Korea, Sept. 2006)