Running data-parallel jobs across geo-distributed sites has emerged as a promising direction due to the growing need for geo-distributed cluster deployment. A key difference between geo-distributed and intra-cluster jobs is the heterogeneous (and often constrained) nature of compute and network resources across the sites. We propose Tetrium, a system for multi-resource allocation in geo-distributed clusters, that jointly considers both compute and network resources for task placement and job scheduling. Tetrium significantly reduces job response time, while incorporating several other performance goals with simple control knobs. Our EC2 deployment and trace-driven simulations suggest that Tetrium improves the average job response time by up to 78% compared to existing data-locality-based solutions, and up to 55% compared to Iridium, the recently proposed geo-distributed analytics system.

Abstract In this paper, we investigate the problem of joint sink deployment and association (JSDA) in multi‐sink wireless camera networks (WCNs). Each camera in a WCN requires a different streaming rate of delivering its stream back to an access point (sink) because of various surveillance requirements, for example, event detection or target tracking. In a WCN where multiple channels are supported, the sinks must be placed in suitable locations (sink deployment) to collect streams from cameras over nonoverlapping channels, and in addition, each camera must associate with the appropriate sink (sink association) so that its demand rate can be optimally satisfied. To achieve this goal, we first formulate the JSDA problem as an optimization model using mixed‐integer linear programming and prove its NP‐completeness. Two approaches, branch‐and‐bound and our heuristics, iterative sink deployment and association (ISDA), are then developed to solve JSDA. We evaluate the performance via simulations with the traces collected by real measurements and show that ISDA can effectively satisfy cameras' demands with a reasonable computational cost. Copyright © 2014 John Wiley & Sons, Ltd.

With growing data volumes generated and stored across geo-distributed datacenters, it is becoming increasingly inefficient to aggregate all data required for computation at a single datacenter. Instead, a recent trend is to distribute computation to take advantage of data locality, thus reducing the resource (e.g., bandwidth) costs while improving performance. In this trend, new challenges are emerging in job scheduling, which requires coordination among the datacenters as each job runs across geo-distributed sites. In this paper, we propose novel job scheduling algorithms that coordinate job scheduling across datacenters with low overhead, while achieving near-optimal performance. Our extensive simulation study with realistic job traces shows that the proposed scheduling algorithms result in up to 50% improvement in average job completion time over the Shortest Remaining Processing Time (SRPT) based approaches.

ABSTRACT The utilization of limited energy in wireless sensor networks (WSNs) is the critical concern, whereas the effectiveness of routing mechanisms substantially influence energy usage. We notice that two common issues in existing specific routing schemes for WSNs are that (i) a path may traverse through a specific set of sensors, draining out their energy quickly and (ii) packet retransmissions over unreliable links may consume energy significantly. In this paper, we develop an energy‐efficient routing scheme (called EFFORT) to maximize the amount of data gathered in WSNs before the end of network lifetime. By exploiting two natural advantages of opportunistic routing, that is, the path diversity and the improvement of transmission reliability, we propose a new metric that enables each sensor to determine a suitable set of forwarders as well as their relay priorities. We then present EFFORT, a routing protocol that utilizes energy efficiently and prolongs network lifetime based on the proposed routing metric. Simulation results show that EFFORT significantly outperforms other routing protocols. Copyright © 2011 John Wiley & Sons, Ltd.

In this paper, we investigate the Access Point (AP) deployment problem in Wireless Surveillance Systems (WSSs). Cameras in a WSS could have various demanding upload streaming rates to the AP (usually deemed as a sink or data collector) according to their surveillance requirements, e.g. object detection frequency, video resolution, or the successful ratio of object tracking. Therefore, a suitable AP location must be determined in order to satisfy each camera's demand. We first provide a mathematical formulation that defines the best location of the AP, and then propose a novel mechanism, called Spring-Cam, to exploit the concept of mass-spring systems in Physics to locate the suitable AP location with consideration of camera demands as well as transmission quality of the link between each camera and the AP. Simulation results show that Spring-Cam can use only a limited measurement overhead to find an AP location that achieves a performance comparable with exhausted searching the best location.

Lifetime-maximization is the critical concern for wireless sensor networks (WSNs). We notice that two common issues in existing routing schemes for WSNs are that (1) a path may traverse through a fixed set of sensors, draining out their energy quickly, and (2) packet retransmissions over an unreliable link of any fixed-path may consume energy significantly. In this paper, we exploit two natural advantages of opportunistic routing, i.e., path diversity and the improvement of transmission reliability, to develop a distributed routing scheme (called EFFORT) for prolonging the network-lifetime of a WSN. Specifically, a new metric is proposed to assist each sensor in determining a suitable set of forwarders as well as their priorities and, thus, enables EFFORT to extend the network-lifetime. Simulation results show that EFFORT effectively achieves network-lifetime extension compared to other routing protocols.

Since the effectiveness of the routing scheme directly affects resource (e.g. energy, bandwidth, and memory) usage in wireless sensor networks, research works on routing protocols have received considerable attention recently. In this paper, we propose a location-free point-to-point routing scheme, called virtual domain and coordinate routing (VDCR), in wireless sensor networks. VDCR applies the concepts of the virtual domain and the virtual coordinate to design its distributed assignment protocol and the location-free routing protocol. That is, the assignment protocol assigns each sensor node the virtual coordinate to represent the location in a network and the virtual domain to maintain the information about the existing shortest path that has been established during the coordinate assignment. Our VDCR is able to efficiently discover the routing path based on the assigned domains and coordinates. The simulations show that the routing path of VDCR is shorter than ABVCap by 62%-70%. Besides, VDCR require 63%-89% preprocess overhead of ABVCap. For realistic network topologies with obstacles, VDCR can improve the routing path by 64% and up to 96%, compared with GPSR that requires physical location information.

Due to the vehicle's movement and inherent characteristics of wireless channel, how to discover and retrieve the required contents has become a challenging issue in a hybrid network including vehicular ad hoc networks (VANETs) and the Internet. In this paper, we propose a social cluster-based P2P framework that estimates similarity and connection condition among peers to provide mobile peers efficient resource discovery and retrieval. The idea of our framework is to consider a mobile peer's preference and its connectivity such as the lifetime or the bandwidth for a wireless link. Specifically, there are two major components in our framework. First, a social cluster-based overlay structure and lifetime-aware flooding scheme enable mobile peers to discover resource through the Internet and VANETs. Second, a connectivity-aware retrieval scheme considers connection lifetime and bandwidth schedules and determines how to retrieve resource from available peers. The simulation results show that, compared with other existing P2P schemes, our social cluster-based P2P framework is able to quickly locate more available peers holding the required contents and achieve a higher retrieval ratio.