Zegura, Ellen W.

Associated Organization(s)
Organizational Unit
ArchiveSpace Name Record

Publication Search Results

Now showing 1 - 8 of 8
  • Item
    IC-Cloud: Computation Offloading to an Intermittently-Connected Cloud
    (Georgia Institute of Technology, 2013) Shi, Cong ; Pandurangan, Pranesh ; Ni, Kangqi ; Yang, Juyuan ; Ammar, Mostafa H. ; Naik, Mayur ; Zegura, Ellen W.
    Offloading computation-intensive components of mobile applications to the cloud is of great potential to speedup the execution and reduce the energy consumption for mobile devices. The gain from computation offloading is typically counterbalanced by communication costs and delays. It is, therefore, important to undertake offloading decisions based on future prediction of Internet access timeliness and quality. Previous approaches have considered this question under the assumption that network connectivity is relatively stable. In this paper, we present IC-Cloud, a computation-offloading system for mobile environments where Internet access to remote computation resources is of highly variable quality and often intermittent. IC-Cloud uses three key ideas: lightweight connectivity prediction, lightweight execution prediction and prediction use in a risk controlled manner tomake offloading decisions. Our connectivity-prediction method only uses the signal strength and user historical information to obtain a coarse-grained estimation of the network access quality. Our execution-prediction mechanism uses machine learning on dynamic program features to automatically, efficiently, and accurately predict the execution time of offloadable tasks, both on the phone and in the cloud. Acknowledging the uncertainties in these predictions, we propose a risk-control mechanism to reduce the impact of inaccurate predictions. We implemented IC-Cloud on Android and tested the system with different applications in various types of mobile environment. Results we obtained from the prototype show speedup and energy consumption reduction benefits in many computational contexts and intermittent connectivity environments.
  • Item
    Design and Analysis of Schedules for Virtual Network Migration
    (Georgia Institute of Technology, 2012) Lo, Samantha ; Ammar, Mostafa H. ; Zegura, Ellen W.
    The Internet faces well-known challenges in realizing modifications to the core architecture. To help overcome these limitations, virtual networks run over physical networks and use Internet paths and protocols as essentially a link layer in the virtual network. Effective use of the underlying network requires intelligent placement of virtual networks so that underlying resources do not incur over-subscription. Additionally, because virtual networks may come and go over time, and underlying networks may experience their own dynamic changes, virtual networks may need to be migrated— re-mapped to the physical network during active operation— to maintain good performance. In this paper we consider the problem of scheduling the sequence of node moves that take a virtual network from an original placement to a new placement. We build on prior work that achieves migration of a single node with minimal disruption to develop a model for the migration cost and latency for a given network migration schedule. We then develop algorithms for determining a single-node-at-a-time sequence of moves to minimize migration cost, and further consider multiple node moves in parallel to minimize migration time and cost. Our algorithms are the first we are aware of to systematically address the virtual network migration scheduling problem.
  • Item
    Serendipity: A Distributed Computing Platform for Disruption Tolerant Networks
    (Georgia Institute of Technology, 2011-01) Shi, Cong ; Lakafosis, Vasileios ; Ammar, Mostafa H. ; Zegura, Ellen W.
    The opportunistic or disruption tolerant networking (DTN) paradigm shows up in a variety of settings, from military to disasters to the developing world to deep space; anywhere that fixed infrastructure is either unavailable or expensive. Simple messaging applications have substantial value for communication and coordination. We posit that these settings can also leverage applications that are computationally complex and will benefit from distributed computing to take full advantage of nearby computational resources. Computing over these networks is not trivial, however, since network disconnections are common and persist over many time scales. In this paper we present the design and implementation of Serendipity, a general purpose distributed computing platform designed to run over a DTN. We have designed a simple but powerful job structure that is suitable for such an underlying network. As opposed to traditional distributed computing platforms in data centers and clusters, where a central master is used to allocate tasks and monitor the working nodes, Serendipity relies on the collaboration among DTN nodes on these functionalities. Smart task allocation algorithms are designed to disseminate tasks among mobile devices by accounting for the special properties of DTNs. The extensive evaluation of our system on Emulab demonstrates that Serendipity efficiently speeds up various kinds of distributed computing jobs by a factor of 2.3 to 10.1 in a diverse set of DTN environments.
  • Item
    Collaborative Research: NeTS-NBD: Construction of robust and efficient disruption tolerant networks
    (Georgia Institute of Technology, 2010-11-14) Zegura, Ellen W. ; Ammar, Mostafa H. ; Clark, Russ
  • Item
    Message Ferries as Generalized Dominating Sets in Intermittently Connected Mobile Networks
    (Georgia Institute of Technology, 2009) Ammar, Mostafa H. ; Polat, Bahadir K. ; Sachdeva, Pushkar ; Zegura, Ellen W.
    Message ferrying is a technique for routing data in wireless and mobile networks in which one or more mobile nodes are tasked with storing and carrying data between sources and destinations. To achieve connectivity between all nodes, message ferries may need to relay data to each other. While useful as a routing technique for wireless mobile networks in general, message ferrying is particularly useful in intermittently connected networks where traditional MANET routing protocols are not usable. A wireless and mobile network is said to possess intrinsic message ferrying capability if a subset of the nodes can act as message ferries by virtue of their own mobility pattern, without introducing additional nodes or modifying existing node mobility. Our goal in this work is to provide a formalism by which one can characterize intrinsic message ferrying capability. We first observe that the use of message ferries is the mobile generalization of the well-known use of connected dominating set-based routing in wireless networks. We next consider the problem of identifying the set of nodes in a mobile network which can act as message ferries by virtue of their mobility pattern. To this end, we define the concept of a connected message ferry dominating set (CMFDS) in a manner that achieves data delivery within certain performance bounds. We then develop algorithms that can be used to find such a set within a mobile, wireless network. The general CMFDS algorithm is built around a core algorithm that determines whether a single node in the network can act as a ferry. We provide some illustrative examples to show the application of our algorithm to several mobility patterns.
  • Item
    Hierarchical Power Management in Disruption Tolerant Networks with Traffic-Aware Optimization
    (Georgia Institute of Technology, 2006) Jun, Hyewon ; Ammar, Mostafa H. ; Corner, Mark D. ; Zegura, Ellen W.
    Disruption tolerant networks (DTNs) are wireless mobile networks that are characterized by frequent partitions and long delays. Such networks can be used in highly-challenged environment in which energy resources are limited. Efficient power management, therefore, is essential for their success. In this paper, we present a hierarchical power management in DTNs where nodes are equipped with two complementary radios: a long-range, high power radio and a short range, low-power radio. In this architecture, energy can be conserved by using the low-power radio to discover communication opportunities with other nodes and then wake up the high-power radio to undertake the data transmission. We develop a generalized power management framework and its variations around this idea and evaluate their relative performance. In addition, for the case in which traffic load can be predicted, we devise approximation algorithms to control the sleep/wake-up cycling to provide maximum energy conservation while discovering enough communication opportunities to handle a given traffic load. We evaluate our schemes and our choice of parameters through ns-2 simulations. Our simulation results show that the generalized power management mechanism could augment the usefulness of the low power radio and achieve better energy efficiency than mechanisms relying on one radio for discovery. In addition, our approximation algorithms reduce energy consumption from 73% to 93% compared with the case without power management. We also observe that while an additional low power radio does reduce the energy consumption needed for discovery, the improvement could be negligible in mobile DTNs due to the low density of nodes.
  • Item
    Capacity Enhancement Using Throw-Boxes in Mobile Delay Tolerant Networks
    (Georgia Institute of Technology, 2006) Zhao, Wenrui ; Chen, Yang ; Ammar, Mostafa H. ; Corner, Mark D. ; Levine, Brian ; Zegura, Ellen W.
    Delay tolerant networks (DTNs) are a class of emerging networks that are subject to frequent and long-duration partitions. Due to intermittent connectivity, DTNs might be significantly limited in supporting application needs, for example, leading to low throughput or high delay. To address this problem, we propose the use of throw-boxes to improve data delivery performance. Throw-boxes are small, inexpensive devices equipped with wireless interfaces and deployed to relay data between mobile nodes. Being small and inexpensive, throwboxes represent a flexible and cost-effective approach to enhance network capacity. In this paper, we systematically study two inter-related issues, namely deployment and routing, in using throw-boxes for throughput enhancement. Specifically, we develop algorithms for throw-box deployment and data forwarding under various routing strategies, including single path, multi-path and epidemic routing. Using extensive ns simulations, we evaluate the utility of throw-boxes and the impact of various routing and deployment strategies on network performance. Our objective is to guide the design and operations of throw-box-enhanced DTNs. We find that throw-boxes are very effective in improving both data delivery ratio and delay, especially for multi-path routing and environments with regular node movement.