Kooperativer Bibliotheksverbund

Berlin Brandenburg

and
and

Your email was sent successfully. Check your inbox.

An error occurred while sending the email. Please try again.

Proceed reservation?

Export
  • 1
    Language: English
    In: Procedia Computer Science, 2011, Vol.3, pp.401-407
    Description: Packet compression is a well-known technique for improving the performance of low speed networks such as WANs. This technique is also effective in networks with a high cost per transmitted byte, namely wireless networks. Most implementations of this technique as a network service, like IPComp, are not transparent and require modifications either to applications or to operating systems. Some other implementations of this technique as a network service, e.g. WANProxy, run as user space processes that impose extra overhead due to unnecessary packet hooking mechanisms, involving frequent switches between user and kernel spaces. This paper presents a transparent packet compression network service using virtualization technology. This service transparently compresses network packet payloads and reduces communication overhead of applications running inside virtual machines. This service requires no modifications to applications, operating systems or virtual machine monitors. In addition, it has no extra overhead for switching between user and kernel spaces, because it is implemented as a configurable kernel module that can be activated or deactivated dynamically. A proof-of-concept transparent packet compression service has been implemented on Xen hypervisor. Evaluation results show the feasibility of development and dynamic configuration of such a transparent packet compression network service. Results also show approximately 25% improvement in network performance of applications over slow or congested links. In addition, the results show that this service can tremendously reduce transmitted bytes over high cost networks like wireless networks, in a transparent manner.
    Keywords: Packet Compression ; Virtualization Technology ; Transparent Network Service ; Computer Science
    ISSN: 1877-0509
    E-ISSN: 1877-0509
    Library Location Call Number Volume/Issue/Year Availability
    BibTip Others were also interested in ...
  • 2
    Language: English
    In: Procedia Computer Science, 2011, Vol.3, pp.484-490
    Description: Task allocation is a critical issue in proper engineering of cooperative applications in embedded systems with latency and energy constraints, as in wireless sensor and actor networks (WSANs). Existing task allocation algorithms are mostly concerned with energy savings and ignore time constraints and thus increase the makespan of tasks in the network as well as the probability of malfunctioning of the network. In this paper we take both energy awareness and reduction of actor tasks’ times to completion in WSANs into account and propose a two-phase task allocation technique based on Queuing theory. In the first phase, tasks are equally assigned to actors just to measure the capability of each actor to perform the assigned tasks. Tasks are then allocated to actors according to their measured capabilities in such a way to reduce the total completion times of all tasks in the network. The results of simulations on typical scenarios shows 45% improvement in the makespan of tasks in a network compared to the wellknown opportunistic load balancing (OLB) task allocation algorithm that is generally used in distributed systems. It is shown that our algorithms provide better tradeoffs between load balancing and completion times of all tasks in a WSAN compared to OLB.
    Keywords: Wireless Sensor Actor Network ; Task Allocation ; Task Completion Time ; Load Balancing ; Queuing System ; Computer Science
    ISSN: 1877-0509
    E-ISSN: 1877-0509
    Library Location Call Number Volume/Issue/Year Availability
    BibTip Others were also interested in ...
Close ⊗
This website uses cookies and the analysis tool Matomo. Further information can be found on the KOBV privacy pages