Abstract

Abstract—RPS is a publicly available toolkit that allows a practitioner to straightforwardly create flexible online and offline resource prediction systems in which resources are represented by independent, periodically sampled, scalar-valued measurement streams. The systems predict the future values of such streams from past values and are composed at runtime out of a large and extensible set of communicating components that are in turn constructed using RPS’s extensible sensor, prediction, wavelet, and communication libraries. This paper describes the design, implementation, and performance of RPS. We have used RPS extensively to evaluate predictive models and build online prediction systems for host load, Windows performance data, and network bandwidth. The computation and communication overheads involved in such systems are quite low. Index Terms—Distributed systems, performance of systems. æ 1

Citations

1324	The GRID: Blueprint for a new Computing Infrastructure - FOSTER, KESSELMAN - 1999
1181	Time Series Analysis: Forecasting and Control - Box, Jenkins, et al. - 1994
583	The Network Weather Service: a distributed resource performance forecasting service for metacomputing, Future Generation Comput - Wolski, Spring, et al. - 1999
296	Architectural Support for Quality of Service for CORBA Objects. Theory and Practice of Object Systems - Zinky, Bakken, et al. - 1997
214	Introduction to Time Series and Forecasting - Brockwell, Davis - 1996
205	An Introduction to LongMemory Time Series Models and Fractional Differencing - Granger, Joyeux - 1980
199	Dynamically forecasting network performance to support dynamic scheduling using the network weather service, in - Wolski - 1997
190	SPAND: Shared passive network performance discovery - Seshan, Stemm, et al. - 1997
149	Fractional differencing - Hosking - 1981
111	Threshold models in nonlinear time series analysis - Tong - 1983
99	The Remote Method Invocation Specification - Microsystems - 1997
86	Scheduling from the Perspective of the Application - Berman, Wolski - 1996
83	Space-time modelling with longmemory dependence: assessing Ireland’s wind power resource - Haslett, Raftery - 1989
82	Statistical Methods for Data with Long-Range Dependence - Beran - 1992
73	Predicting the cpu availability of timeshared unix systems on the computational grid - Wolski, Spring, et al. - 1999
50	O’Hallaron, Host load prediction using linear models, Cluster Computing 3 (2000) 265–280. 44 - Dinda, R
41	An evaluation of linear models for host load prediction - Dinda, O’Hallaron - 1999
40	Time series models for Internet traffic - -Basu, Mukherjee, et al. - 1996
38	Online prediction of the running time of tasks - Dinda
35	Automatic generation of parallel programs with dynamic load balancing - Siegell, Steenkiste - 1994
33	Preliminary report on the design of a framework for distributed visualization - Lopez, Aeschlimann, et al. - 1999
31	The performance of a service for network-aware applications - Obraczka, Georghiu - 1998
30	A time series model of long-term NSFNET backbone traffic - Groschwitz, Polyzos - 1994
29	Dynamic Monitoring of High-Performance Distributed Applications - Gunter, Tierney, et al. - 2002
24	The statistical properties of host load. Scientific Programming, 7(3,4 - Dinda - 1999
20	A resource monitoring system for network-aware applications - LOWEKAMP, MILLER, et al. - 1998
19	Design, Implementation and Evaluation of the Remos Network - Gross, Lowekamp, et al. - 2003
18	Direct queries for discovering network resource properties in a distributed environment - Lowekamp, O’Hallaron, et al. - 1999
17	A Prediction-Based Real-Time Scheduling Advisor - Dinda - 2002
16	Analysis of Observed Chaotic Data. Institute for Nonlinear Science - Abarbanel - 1996
16	Windows Performance Monitoring and Data Reduction using WatchTower - Knop, Schopf, et al. - 2002
15	The case for prediction-based best-effort real-time systems - DINDA, LOWEKAMP, et al. - 1998
15	A hierarchy of network performance characteristics for grid applications and services - Lowekamp, Tierney, et al. - 2003
12	On Distributed Scheduling Using Load Prediction from Past Information,” abstracts published - Samadani, Kalthofen - 1996
10	An Empirical Study of the Multiscale Predictability of Network Traffic - Qiao, Skicewicz, et al. - 2004
8	CORBA: Fundamentals and Programming - Siegal - 1996
7	Service location protocol. Internet RFC 2165 - VEIZADES, GUTTMAN, et al. - 1997
6	Fracdiff: Maximum Likelihood Estimation of the Parameters of a Fractionally Differenced ARIMA(p; d; q) Model,” Computer Program - Fraley - 1991
6	Mathcad User’s Guide - Mathsoft - 1995
6	Load Balancing Using Time Series Analysis for Soft Real-Time Systems with Statistically Periodic Loads - Hailperin - 1993
6	Multi-resolution resource behavior queries using wavelets - Skicewicz, Dinda, et al. - 2001
6	A Grid Monitoring Architecture. Technical report, Global Grid Forum - Tierney, Aydt, et al. - 2002
5	Work-based learning - Naylor - 1997
2	Exploiting packet header redundancy for zero cost dissemination of dynamic resource information - DINDA - 2002
2	Modeling and Taming - Lu - 2005
2	Tsunami: A Wavelet Toolkit for Distributed Systems - Skicewicz, Dinda - 2003
1	An Implementation of Diffusion in the Linux Kernel - Cornell, Lange, et al. - 2002
1	Characterizing and Predicting - Lu, Qiao, et al. - 2005
1	XML-RPC specification,” technical report - Winer - 1999