Disseminating Updates on Broadcast Disks
Disseminating Updates on Broadcast Disks
复制标题
在广播磁盘上传播更新
DOI：
--
复制
发表时间：
1996
期刊：
影响因子：
--
通讯作者：
S. Zdonik
中科院分区：
文献类型：
--
作者：
S. Acharya;M. Franklin;S. Zdonik
最近，人们越来越关注使用数据传播作为在有线和无线环境中将数据从服务器传递到客户端的手段。使用数据传播，数据的传输由服务器发起，从而扭转了客户端和服务器之间的传统关系。在以前的论文中，我们已经提出了一个模型，用于结构化的重复传输的数据在广播媒体中的广播传输。由于物理或应用程序相关的原因，广播网关旨在用于客户端和服务器之间的通信容量不对称的环境中。这种环境的例子包括具有移动的客户端的无线网络、有线和直接卫星广播以及信息传播应用。我们最初的研究主要集中在广播数据不变时的机制性能。在本文中，我们扩展这些结果，将更新的影响。我们首先提出了几种替代模型的更新，并检查的货币之间的数据和性能产生的基本权衡。然后，我们提出并分析实现这些不同的模型的机制。性能结果表明，即使在更新必须立即传输的模型中，通过使用简单的技术来传播和预取数据项，也可以使广播重传技术的性能变得相当复杂。
Lately there has been increasing interest in the use of data dissemination as a means for delivering data from servers to clients in both wired and wireless environments. Using data dissemination, the transfer of data is initiated by servers, resulting in a reversal of the traditional relationship between clients and servers. In previous papers, we have proposed Broadcast Disks as a model for structuring the repetitive transmission of data in a broadcast medium. Broadcast Disks are intended for use in environments where, for either physical or application-dependent reasons, there is asymmetry in the communication capacity between clients and servers. Examples of such environments include wireless networks with mobile clients, cable and direct satellite broadcast, and information dispersal applications. Our initial studies of Broadcast Disks focused on the performance of the mechanism when the data being broadcast did not change. In this paper, we extend those results to incorporate the impact of updates. We first propose several alternative models for updates and examine the fundamental tradeoff that arises between the currency of data and performance. We then propose and analyze mechanisms for implementing these various models. The performance results show that, even in a model where updates must be transmitted immediately, the performance of the Broadcast Disks technique can be made quite mbust through the use of simple techniques for propagating and prefetching data items.