CNRL logo

Computer Networking Research Laboratory

Dept. of Electrical & Computer Engineering, Colorado State University

ResQue - Multi-attribute Resource and Query Generator

Multi-attribute Resource and Query (ResQue) generator can be used to generate large synthetic traces of computing resources and range queries. Such traces are useful in large-scale performance studies of resource discovery systems, job schedulers, etc., in collaborative peer-to-peer systems, volunteer computing, and grid and cloud computing. ResQue generates correlated multivariate vectors of static attributes and time series of dynamic attributes as well as multi-attribute range queries. It preserves the statistical properties of real-world computing resources such as distributions of attributes, complex correlation between static and/or dynamic attributes, contemporaneous correlation among dynamic attributes, and popularity of attributes. Users may use the provided datasets or their own ones as the basis to generate large synthetic traces. Several tools are also provided to simplify the conversion of other datasets to the format supported by ResQue.

Please contact us if you would like to share your dataset with other ResQue users.

Resources

Use the following links to access source code, datasets, user guide, and utilities.

  • Executable files − Complied using MATLAB 2012a on a 64-bit Windows machine.
  • Source files
  • Datasets − Several pre-processed datasets are already available. See datasets for specific details. We would like to thank Dr. J. Albrecht for providing the SWORD dataset, Dr. V.S. Pai for providing CoMon dataset, and Grid Observatory team for providing the grid dataset.
    • Dummy − Dummy dataset used to demonstrate how ResQue works (1 KB).
    • PlanetLab − Resource dataset with both static and dynamic attributes as well as multi-attribute range queries. Resource datasets were collected from CoMon and query datasets from SWORD (24 MB).
    • SETI@home − Static attributes only. Data collected from SETI@home (10 MB).
    • GCO grid − Resource dataset with both static and dynamic attributes from Green Computing Observatory which is a sub-project within Grid Observatory (75 MB).
    • CSU - Resource dataset with both static and dynamic attributes from our campus (41 MB).
  • User guide
  • Utilities − Used to preprocess CoMon, SWORD, SETI, and GCO data, split time series according to structural changes, etc. (8 MB)

Publications

  1. H. M. N. Dilum Bandara and Anura P. Jayasumana, "On Characteristics and Modeling of P2P Resources with Correlated Static and Dynamic Attributes," In Proc. IEEE GLOBECOM '11, Dec. 2011.
  2. H. M. N. Dilum Bandara and Anura P. Jayasumana, "Characteristics of Multi-Attribute Resources/Queries and Implications on P2P Resource Discovery," In Proc. 9th ACS/IEEE Int. Conf. On Computer Systems And Applications (AICCSA 2011), Sharm El-Sheikh, Egypt, Dec. 2011.
  3. H. M. N. D. Bandara, “Enhancing collaborative peer-to-peer systems using resource aggregation and caching: A multi-attribute resource and query aware approach,” PhD Dissertation, Colorado State University, Fall 2012.