Researchers are developing mobile sensing platforms to facilitate public awareness of environmental conditions. However, turning such awareness into practical community action and political change requires more than just collecting and presenting data. To inform research on mobile environmental sensing, we conducted design fieldwork with government, private, and public interest stakeholders. In parallel, we built an environmental air quality sensing system and deployed it on street sweeping vehicles in a major U.S. city; this served as a research vehicle"by grounding our interviews and affording us status as environmental action researchers. In this paper, we present a qualitative analysis of the landscape of environmental action, focusing on insights that will help researchers frame meaningful technological interventions.

Poor air quality is a global health issue, causing serious problems like asthma, cancer, and heart disease around the world. Earlier this decade, the World Health Organization estimated that three million people die each year from the effects of air pollution [6]. Unfortunately, while variations in air quality are significant, today's air quality monitors are very sparsely deployed. To address this visibility gap, the Common Sense project is developing participatory sensing systems that allow individuals to measure their personal exposure, groups to aggregate their members' exposure, and activists to mobilize grassroots community action.

Habitat and environmental monitoring is a driving application for wireless sensor networks. We present an analysis of data from a second generation sensor networks deployed during the summer and autumn of 2003. During a 4 month deployment, these networks, consisting of 150 devices, produced unique datasets for both systems and biological analysis. This paper focuses on nodal and network performance, with an emphasis on lifetime, reliability, and the the static and dynamic aspects of single and multi-hop networks. We compare the results collected to expectations set during the design phase: we were able to accurately predict lifetime of the single-hop network, but we underestimated the impact of multi-hop traffic overhearing and the nuances of power source selection. While initial packet loss data was commensurate with lab experiments, over the duration of the deployment, reliability of the backend infrastructure and the transit network had a dominant impact on overall network performance. Finally, we evaluate the physical design of the sensor node based on deployment experience and a post mortem analysis. The results shed light on a number of design issues from network deployment, through selection of power sources to optimizations of routing decisions.

Virtual networks provide applications with the illusion of having their own dedicated, high-performance networks, although network interfaces posses limited, shared resources. We present the design of a large-scale virtual network system and examine the integration of communication programming interface, system resource management, and network interface operation. Our implementation on a cluster of 100 workstations quantifies the impact of virtualization on small message latencies and throughputs, shows full hardware performance is delivered to dedicated applications and time-shared workloads, and shows robust performance under demanding workloads that overcommit interface resources.

Implicit coscheduling is a distributed algorithm for time-sharing communicating processes in a cluster of workstations. By observing and reacting to implicit information, local schedulers in the system make independent decisions that dynamically coordinate the scheduling of communicating processes. The principal mechanism involved is two-phase spin-blocking: a process waiting for a message response spins for some amount of time, and then relinquishes the processor if the response does not arrive.In this paper, we describe our experience implementing implicit coscheduling on a cluster of 16 UltraSPARC I workstations; this has led to contributions in three main areas. First, we more rigorously analyze the two-phase spin-block algorithm and show that spin time should be increased when a process is receiving messages. Second, we present performance measurements for a wide range of synthetic benchmarks and for seven Split-C parallel applications. Finally, we show how implicit coscheduling behaves under different job layouts and scaling, and discuss preliminary results for achieving fairness.

Clusters of multiprocessors, or Clumps, promise to be the supercomputers of the future, but obtaining high performance on these architectures requires an understanding of interactions between the multiple levels of interconnection. In this paper, we present the first multi-protocol implementation of a lightweight message layer---a version of Active Messages-II running on a cluster of Sun Enterprise 5000 servers connected with Myrinet. This research brings together several pieces of high-performance interconnection technology: bus backplanes for symmetric multiprocessors, low-latency networks for connections between machines, and simple, user-level primitives for communication. The paper describes the shared memory message-passing protocol and analyzes the multi-protocol implementation with both microbenchmarks and Split-C applications. Three aspects of the communication layer are critical to performance: the overhead of cache-coherence mechanisms, the method of managing concurrent access, and the cost of accessing state with the slower protocol. Through the use of an adaptive polling strategy, the multi-protocol implementation limits performance interactions between the protocols, delivering up to 160 MB/s of bandwidth with 3.6 microsecond end-to-end latency. Applications within an SMP benefit from this fast communication, running up to 75% faster than on a network of uniprocessor workstations. Applications running on the entire Clump are limited by the balance of NIC's to processors in our system, and are typically slower than on the NOW. These results illustrate several potential pitfalls for the Clumps architecture.

This paper presents a network mapping algorithm and proves its correctness assuming a traffic-free network.Respecting well-defined parameters, the algorithm produces a graph isomorphic to N -F, where N is the network of switches and hosts and F is the set of switches connected by a switch-bridge to the set of hosts I-I.We show its performance on a Myrinet system-area network with a fat-tree-like topology.It can map 36 nodes, 13 switches and 64 links in 248 ms and 100 nodes, 40 switches, and 193 linksin981 rns.From such maps, the system computes mutually deadlock-free routes and distributes them to all network interfaces.Switched, multi-gigabyte per second, system area networks are the enabling building-blocks for networks of workstations.Because of their core role, these networks should be dynamically recontigurable, automatically adapting to the addition or removal of hosts, switches and links. IntroductionSystem area networks [1] move switched, low-latency, high-speed networks away from the backplanes and cabinets of massively parallel processors into the traditional territory of local area networks.These networks commonly use source-based message routing through anonymous switches.In this regime, their topologies may no longer be the static, well-defined, and well-understood [2] graphs such as hypercubes, meshes.etc., and instead may be arbitrary graphs that change over time.Therefore, systems must periodically discover their topologies rather than assuming one a priori.Lacking an out-of-band mechanism for directly querying switches for their identities, systems must use in-band messaging to disambiguate switch identities when discovering the network topology.From the resulting maps, systems can compute mutually deadlock-free routes without relying upon properties of traditional multicomputer networks, e.g., static topologies, that may now be transient.