The role of sunchrony in the implementation of shared memory objects

sunchrony在共享内存对象实现中的作用

• 批准号: 327672-2006
• 负责人: Brodsky, Alexander
• 金额: $ 1.02万
• 依托单位: University of Winnipeg
• 依托单位国家: 加拿大
• 项目类别: Discovery Grants Program - Individual
• 财政年份: 2008
• 资助国家: 加拿大
• 起止时间: 2008-01-01 至 2009-12-31
• 项目状态: 已结题
• 来源: https://www.nserc-crsng.gc.ca/ase-oro/Details-Detailles_eng.asp?id=397908
• 关键词: role; sunchrony; implementation; shared; memory

Presently, most of the algorithms for solving problems in a distributed fashion do not make any assumptions regarding the degree to which the components of the underlying distributed system are synchronized.  Although such assumptions allow for a very general design methodologies, they ignore a useful attribute of most distributed systems;  namely, that the components of many distributed are comparatively synchronized.  The amount of synchrony in a system is the degree to which the system's components operate in concert. I propose to investigate how this synchrony among a system's components can be leveraged in order to design faster and more efficient distributed algorithms.  Specifically, I am interested in determining the types of synchrony present in distributed systems, which forms are most useful, and how to design distributed algorithms so that they can leverage these forms of synchrony. This is useful for several reasons.  First, it improves our understanding of the foundations of computer science.  We know that that synchrony plays an important role in determining whether and how efficiently a particular problem can be solved in a distributed manner; the challenge is to determine what this role is. Second, by determining what forms of synchrony are useful, we can design more efficient distributed systems.  These lessons are applicable to both the hardware designers and the software architects, and would facilitate better integration between a system's hardware and software.  This would improve the operation, efficiency and safety of distributed systems, which are being utilized more and more in research and industry.

目前，大多数以分布式方式解决问题的算法都没有对底层分布式系统的组件同步程度做任何假设。尽管这样的假设允许使用非常通用的设计方法，但它们忽略了大多数分布式系统的一个有用属性；即，许多分布式的组件是相对同步的。系统中的同步量是系统组件协同工作的程度。我建议研究如何利用系统组件之间的这种同步，以便设计更快、更有效的分布式算法。具体来说，我感兴趣的是确定分布式系统中存在的同步类型，哪种形式最有用，以及如何设计分布式算法，以便它们能够利用这些同步形式。这很有用，原因有几个。首先，它提高了我们对计算机科学基础的理解。我们知道，在决定是否以分布式方式解决特定问题以及如何有效解决问题方面，同步发挥着重要作用；挑战在于确定这个角色是什么。其次，通过确定什么形式的同步是有用的，我们可以设计更高效的分布式系统。这些经验教训适用于硬件设计师和软件架构师，并且将促进系统硬件和软件之间更好的集成。这将提高分布式系统的运行、效率和安全性，分布式系统在研究和工业中得到越来越多的应用。