|
Second Workshop on Real, Large Distributed Systems—Preliminary Abstract
Pp. 67–72 of the Proceedings
Using PlanetLab for Network Research: Myths, Realities, and Best Practices
Neil Spring, University of Maryland; Larry Peterson, Andy Bavier, and Vivek S. Pai, Princeton University
Abstract
PlanetLab is a research testbed that supports 428 experiments
on 276 sites, with 583 nodes in 30 countries. It has
lowered the barrier to distributed experimentation in network
measurement, peer-to-peer networks, content distribution,
resource management, authentication, distributed
file systems, and many other areas.
PlanetLab did not become a useful network testbed
overnight. It started as little more than a group of Linux
machines with a common password file, which scaled
poorly and suffered under load. However, PlanetLab was
conceived as an evolvable system under the direction of a
community of researchers. With their help, PlanetLab version
3.0 has since corrected many previous faults through
virtualization and substantial performance isolation. This
paper is meant to guide those considering developing a
network service or experiment on PlanetLab by separating
widely-held myths from the realities of service and
experiment deployment.
Building and maintaining a testbed for the research
community taught us lessons that may shape its continued
evolution and may generalize beyond PlanetLab to
other systems. First, users do not always search out Òbest
practiceÓ approaches: they expect the straightforward approach
to work. Second, users rarely report failed attempts:
we learned of the perceived shortcomings described
in this paper through conversations, not through
messages to the mailing lists. Third, frustration lingers:
users hesitate to give another chance to a system that was
recently inadequate or difficult to use. These experiences
are especially challenging for an evolvable system, which
relies on user feedback to evolve so that more users can
be supported by features they desire.
We organize the myths in decreasing order of veracity:
those that are realities in Section 2, that were once true
in Section 3, and those that are false if best practices are
employed in Section 4. We summarize the discussion in
Section 5.
- View the full text of this paper in HTML and PDF.
Until December 2006, you will need your USENIX membership identification in order to access the full papers. The Proceedings are published as a collective work, © 2005 by the USENIX Association. All Rights Reserved. Rights to individual papers remain with the author or the author's employer. Permission is granted for the noncommercial reproduction of the complete work for educational or research purposes. USENIX acknowledges all trademarks within this paper.
- If you need the latest Adobe Acrobat Reader, you can download it from Adobe's site.
|
Need help? Use our Contacts page.
