The battery life of mobile devices is one of their most important resources. Much of the literature focuses on accurately profiling the power consumption of device components or enabling application developers to develop energy-efficient applications through fine-grained power profiling. However, there is a lack of tools to enable users to extend battery life on demand. What can users do if they need their device to last for a specific duration in order to perform a specific task? To this extent, we developed BatteryExtender, a user-guided power management tool that enables the reconfiguration of the device's resources based on the workload requirement, similar to the principle of creating virtual machines in the cloud. It predicts the battery life savings based on the new configuration, in addition to predicting the impact of running applications on the battery life. Through our experimental analysis, BatteryExtender decreased the energy consumption between 10.03% and 20.21%, and in rare cases by up to 72.83%. The accuracy rate ranged between 92.37% and 99.72%.

Energy efficiency is an important issue for data centers given the amount of energy they consume yearly. However, there is still a gap of understanding of how exactly the application type and the heterogeneity of servers and their configuration impact the energy efficiency of data centers. To this end, we introduce the notion of Application Specific Energy Efficiency (ASEE) in order to rank energy efficiency of heterogeneous servers based on the hosted applications. We conducted extensive sets of experiments using three benchmarks: TPC-W, BS Seeker, and Matrix Stress mark. We observed that each server has different ASEE value based on the type of application running, the size of the virtual machine, the application load, and the scalability factor. In some cases, we witnessed 70% of ASEE improvement by changing the virtual machine size within the same node while keeping an identical load. In different cases, we witnessed up to 86% of ASEE improvement by running the same application with the same load within the same size of virtual machine but on different nodes. Our observation has many implications which include but are not limited to improving virtual machine scheduling based on the ASEE rank of the node. Another implication stresses on the importance of accurate prediction of application load and selecting the appropriate virtual machine size in order to improve the ASEE.

We describe QoSMap, an overlay construction mechanism which computes high quality overlay networks for applications having stringent constraints on hop-degrading QoS metrics and provides resilience against the Internetpsilas unpredictable network behavior. QoSMap implements three features to achieve the QoS and resiliency goals. In order to provide high QoS, it constructs overlay communication edges with short but efficient underlay paths. In case of QoS violations due to changing network conditions, QoSMap utilizes supplemental backup paths, which are specifically constructed in order to extend overlay lifetimes. QoSMap also avoids nodes that have experienced recent QoS failure, further improving the resilience of the overlays. PlanetLab experiments prove the ability of QoSMap to construct efficient and resilient overlays for applications with stringent QoS constraints of latency and packet loss for a variety of topologies. QoSMap-constructed overlays significantly outperform overlays constructed without the three QoSMap features. The results also reveal the usability and effectiveness of QoSMap to construct near-optimal solution in a short amount of time, compared to the optimal solution which has a very high time complexity.

We describe QoSMap, a novel approach for mapping overlay networks with stringent QoS requirements. QoSMap computes high quality overlay paths by utilizing direct underlay paths that fulfill the application's topology and link-specific QoS requirements. It also improves application dependability by increasing resiliency against QoS failures under dynamic network conditions. The increased resiliency is achieved through the use of supplemental indirect paths which have a single intermediate node and which also meet application QoS requirements. For evaluation, we compared QoSMap with a simple QoS approach for overlay mapping. We compared the two approaches for fulfilling varying application requests with strict QoS requirements of latency and packet loss on PlanetLab. We observed that QoSMap yielded high quality overlays. Further, the use of indirect paths upon the QoS failure of the direct path avoided overlay reconfiguration and resulted in increased dependability and longer lifetime for the application overlays.

Failure detectors represent a very important building block in distributed applications. The speed and the accuracy of the failure detectors is critical to the performance of the applications built on them. In a common implementation of failure detector based on heartbeats, there is a tradeoff between speed and accuracy so it is difficult to be both fast and accurate. Based on the observation that in many distributed applications, one process takes a special role as the leader, we propose a fast failure detection (FFD) algorithm that detects the failure of the leader both fast and accurately. Taking advantage of spatial multiple timeouts, FFD detects the failure of the leader within a time period of just a little more than one heartbeat interval, making it almost the fastest detection algorithm possible based on heartbeat messages. FFD could be used stand alone in a static configuration where the leader process is fixed at one site. In a dynamic setting, where the role of leader has to be assumed by another site if the current leader fails, FFD could be used in collaboration with a leader election algorithm to speed up the process of electing a new leader.

We describe the design of PSON, an overlay network, which improves the predictability of end-to-end communication latency on the Internet. Predictable Service Overlay Networks (PSON) monitor the paths between overlay nodes and continuously estimates an upper bound on communication latency to provide a soft synchrony guarantee to application. Through network measurements on PlanetLab nodes we observe that the upper bound estimation provides long durations of synchrony and rapidly adjusts the bound in response to an asynchronous event. Upper bound estimation is enhanced by the use of routing and overlay configuration to provide stable routes while reducing overhead. We describe the architecture, design and interface for the PSON and discuss various issues related to its performance.

Many applications need to pick servers with some desired distribution. For example, in probabilistic quorum systems, one method to generate quorums with high probability of intersection is to randomly pick k/spl radic/n nodes with a fixed probability distribution. Load balancing applications may need to take several samples of the servers with some desired distribution. Existing approaches realize a fixed stationary distribution by controlling the topology of the overlay graph and conducting random walks on it. In particular, existing approaches focus on achieving a uniform distribution. This paper proposes using the distributed Hastings-Metropolis algorithm to achieve any desired stationary distribution without control or global knowledge of the overlay graph. The new method facilitates good load balancing, since heterogeneous server capacity or other factors can be considered in deciding the appropriate distribution by which to pick servers.

Internet-based overlay networks are likely to demonstrate poor performance when deployed on the Internet without consideration of topological and link properties. We introduce a new method for tactical construction of unstructured overlay networks which exploits a minimal landmark infrastructure to align the overlay with the underlying network. Our framework results in well-connected overlays which can better meet the growing demands of current applications. We demonstrate, using simulation, that our approach yields reduced latency, better connectivity and higher availability of resources than conventional overlay systems and other landmark-based approaches, thereby providing a scalable and efficient approach for overlay construction.

No abstract available.

Distributed hash tables (DHT) bring the promise of increased availability of data to wide-area systems, under the assumption of uniform request load. However, they do not intrinsically offer protection against the hotspot phenomenon, when an item suddenly becomes popular and the node hosting the item must handle a high surge in demand. Therefore, most DHT-based systems employ techniques such as path caching or caching on the client to alleviate this problem. In this paper, we analyze the efficiency of path caching for DHT that offer convergence of routing paths, and compare it to probabilistic forwarding and congestion control techniques, which help balance the load among the nodes holding caches, while also providing superior availability and reducing the number of denied requests. We show that probabilistic forwarding can simultaneously reduce the request drop rate and the number of caches. We compare four algorithms for probabilistic forwarding and conclude that geometric and adaptive mechanisms are superior to the other approaches.

This paper describes a data aggregation and consensus algorithm for object location and tracking applications deployed on wireless sensor networks. This consensus algorithm permits ad-hoc, in-network, group formation in response to a detected event. By reaching a consensus in the network, only a single message indicating the detected event needs to be forwarded to the tracking application at a base station, leading to significant savings in communication costs and prolonging of the life of the network. Our ad-hoc approach offers superior detecting power to other approaches, which statically partition the network into detection regions. The use of a quorum mechanism to achieve consensus leads to higher availability under failure conditions. Moreover, the algorithm exploits the broadcast communication and synchrony properties of wireless sensor networks. Our experimental results show that the algorithm exhibits low overhead despite faults and converges quickly to a single consensus even if multiple nodes initiate the consensus protocol.

In this paper, we describe the design of a global infrastructure for monitoring and assertion-checking which is scalable and resilient to network and node failures, leveraging the benefits of distributed hash tables. Our mechanism distributes the event notification, predicate evaluation, and predicate reporting tasks across the Internet and is orthogonal to the specific monitoring techniques deployed. Our approach creates a hierarchy of event reporting and monitoring relationships, decoupling the event notification task from the monitored node and facilitating decomposition of the predicate evaluation and monitoring task if the semantics of the monitoring approach permits. The proposed infrastructure is designed for a best effort monitoring service, but may be strengthened to support stronger guarantees. The proposed has many desirable properties, including scalability, efficient routing, load balancing and good behavior under flash crowds, resiliency to node failure, including failure of monitored hosts and monitoring end-hosts, loose coupling of the monitored hosts and the monitoring task, and self-organizing behavior.

Evolve IV is an individual-based metabolically driven evolutionary ecosystem model that captures the reciprocal relationship between biota and their environment. In one series of experiments with Evolve IV, the initial population modifies the environment to its own disadvantage, i.e., it pollutes. Eventually a new nonpolluting population emerges. The length of time taken for this evolutionary learning is affected by the initial population's level of mobility.

Simulation and verification are two conventional techniques for the analysis of specifications of real-time systems. While simulation is relatively inexpensive in terms of execution time, it only validates the behavior of a system for one particular computation path. On the other hand, verification provides guarantees over the entire set of computation paths of a system, but is, in general, very expensive due to the state-space explosion problem. We introduce a new technique: simulation-verification combines the best of both worlds by synthesizing an intermediate analysis method. This method uses simulation to limit the generation of a computation graph to that set of computations consistent with the simulation. This limited computation graph, called a simulation-verification graph, can be one or more orders of magnitude smaller than the full computation graph. A tool, XSVT, is described which implements simulation-verification graphs. Three paradigms for using the new technique are proposed. The paper illustrates the application of the proposed technique via an example of a robot controller for a manufacturing assembly line.

This paper describes MTSim, an extensible, customizable simulation platform for the Modechart toolset (MT). MTSim provides support for "plugging in" user-defined viewers useful in simulating system behavior in different ways, including application-specific ways. MTSim also supports full user participation in the generation of simulations by allowing users to inject events into the execution trace. Moreover, MTSim provides monitoring and assertion checking of execution traces and the invocation of user-specified handlers upon assertion violation. This paper also introduces an MTSim component called WebSim, a suite of simulation tools for MT, and an application-specific component of MTSim which displays the cockpit of an F-18 aircraft and which responds to user inputs to model a bomb release function.