PlanetLab is a continuously-evolving global network research testbed that is simultaneously used by hundreds of researchers for diverse tasks, ranging from shortterm self-contained experiments among PlanetLab nodes to continuously-running Web-accessible services with tens of thousands of non-PlanetLab users. While PlanetLab cannot provide a perfectly-customized environment for every experiment, it has been changing over time, and the base of knowledge of how to best utilize it has also been growing. As a result, many of the early observations researchers made about PlanetLab would change if rechecked today. In this paper, we discuss these issues and explain whether they remain, have been addressed via PlanetLab’s evolution, or can be avoided by the use of best practices. Where possible, we provide quantitative evidence showing the realities of PlanetLab and possible research avenues to further broaden the opportunities for using PlanetLab in network research. 1

Abstract. This paper reports our experiences building PlanetLab over the last four years. It identifies the requirements that shaped PlanetLab, explains the design decisions that resulted from resolving conflicts among these requirements, and reports our experience implementing and supporting the system. Due in large part to the nature of the “PlanetLab experiment, ” the discussion focuses on synthesis rather than new techniques, balancing system-wide considerations rather than improving performance along a single dimension, and learning from feedback from a live system rather than controlled experiments using synthetic workloads. 1

Network accountability, forensic analysis, and failure diagnosis are becoming increasingly important for network management and security. Such capabilities often utilize network provenance – the ability to issue queries over network meta-data. For example, network provenance may be used to trace the path a message traverses on the network as well as to determine how message data were derived and which parties were involved in its derivation. This paper presents the design and implementation of ExSPAN, a generic and extensible framework that achieves efficient network provenance in a distributed environment. We utilize the database notion of data provenance to “explain ” the existence of any network state, providing a versatile mechanism for network provenance. To achieve such flexibility at Internet-scale, ExSPAN uses declarative networking in which network protocols can be modeled as continuous queries over distributed streams and specified concisely in a declarative query language. We extend existing data models for provenance developed in database literature to enable distribution at Internet-scale, and investigate numerous optimization techniques to maintain and query distributed network provenance efficiently. The ExSPAN prototype is developed using Rapid-Net, a declarative networking platform based on the emerging ns-3 toolkit. Experiments over a simulated network and an actual deployment in a testbed environment demonstrate that our system supports a wide range of distributed provenance computations efficiently, resulting in significant reductions in bandwidth costs compared to traditional approaches.

PlanetLab is a geographically distributed platform for deploying, evaluating, and accessing planetary-scale network services. PlanetLab is a shared community effort by a large international group of researchers, each of whom gets access to one or more isolated slices of PlanetLab’s global resources. Because we deployed Planet-Lab and started supporting users before we fully understood what its architecture would be, being able to evolve the system became a requirement. This paper examines the set of design principles that guided this evolution. Some of these principles were explicit at the project outset, and others have become crystallized as the platform has developed. 1.

With the introduction of cellular phones, mobile wireless communication has become an integral part of day-to-day life in this century. A testbed is crucial in continuing cuttingedge wireless technology research and development since physical and MAC layers in mobile wireless communication have many parameters pertinent to performance that are often not tractable in analysis or simulation alone. Although many wireless testbeds have been built, most are not general purpose, have a limited access policy, or focus on single technologies like WiFi. In this paper, we present the implementation and evaluation of a mobile testbed node. This node will be the founding block of a mobile wireless testbed that we envision to be as open as PlanetLab to researchers. 1.

In the past decade, distributed systems have rapidly evolved and gained significant traction in the research community, with an increasing interest concentrated on developing and analyzing secure distributed systems. In this paper, we present DS2 (Declarative Secure Distributed Systems), a unified platform for specifying, implementing, and analyzing large-scale secure distributed systems. First, we propose the Secure Network Datalog (SeNDlog) language that enables distributed systems and their security policies to be specified and implemented within a same declarative framework. We show that the existing semi-naïve evaluation can be extended to execute SeNDlog programs that incorporate authenticated communication among untrusted nodes. Second, we demonstrate that network provenance – the metadata that explains the derivation of network state – can be naturally and concisely captured within the DS2 system. We extend existing data models for provenance to enable distribution at Internet-scale, and present techniques for efficient and customizable maintenance and querying of network provenance. Finally, the future research plans on secure provenance and its integration with legacy applications are presented for discussion.

Surveillance of Internet communications is increasingly common. As a greater and greater percentage of communication occurs over the Internet, the desire by law enforcement, intelligence agencies, criminals, and others to access these communications grows. In recent years, motivated by updated legislation, we have seen the first large-scale systems for intercepting Internet communications deployed, and there is increasing pressure for more such systems to be developed and put to use. Such systems raise a number of obvious questions for the security research community. Unfortunately, nearly all the systems that have been developed are closed and proprietary, and their inner workings closely guarded for commercial and “security ” reasons. Very little research exists in the open academic literature exploring the technical aspects of Internet surveillance, and (to our knowledge) none which focuses on security or reliability. In this work we examine one specific problem, that of performing reliable capture of Internet communications. This work has three main contributions which address some, but by no means all, of the open questions relating to reliable capture in Internet surveillance. First, we provide a survey of the current state of practice for Internet capture in the public literature. Second, we examine a number of ways in which existing capture solutions fall short of perfect capture, and the consequences, namely theoretical vulnerabilities as well as practical attacks on the accuracy and completeness of information analyzed. Finally, we construct a set of improved capture tools which provide stronger, more reliable results when used in conjunction with existing tools. This document represents a dissertation in progress.

PlanetLab is a seasoned Internet testbed for distributed applications consisting of donated nodes located at more than 350 remote locations spread across the globe. It is particularly appropriate for peer-to-peer application research due its large-scale, distributed operation, and the availability of nodes with edge characteristics. This talk describes the basic testbed offering and suggests appropriate use-cases. Test facilities; Distributed algorithms