Finer Garbage Collection in Lindacap

Author:

Udzir Nur Izura1,Ibrahim Hamidah1,Demesie Sileshi2

Affiliation:

1. Universiti Putra Malaysia, Malaysia

2. Bahir Dar University, Ethiopia

Abstract

As open systems persist, garbage collection (GC) can be a vital aspect in managing system resources. Although garbage collection has been proposed for the standard Linda, it was a rather course-grained mechanism. This finer-grained method is offered in Lindacap, a capability-based coordination system for open distributed systems. Multicapabilities in Lindacap enable tuples to be uniquely referenced, thus providing sufficient information on the usability of tuples (data) within the tuple-space. This paper describes the garbage collection mechanism deployed in Lindacap, which involves selectively garbage collecting tuples within tuple-spaces. The authors present the approach using reference counting, followed by the tracing (mark-and-sweep) algorithm to garbage collect cyclic structures. A time-to-idle (TTI) technique is also proposed, which allows for garbage collection of multicapability regions that are being referred to by agents but are not used in a specified length of time. The performance results indicate that the incorporation of garbage collection techniques adds little overhead to the overall performance of the system. The difference between the average overhead caused by the mark-and-sweep and reference counting is small, and can be considered insignificant if the benefits brought by the mark-and-sweep is taken into account.

Publisher

IGI Global

Reference28 articles.

1. Accurate garbage collection in uncooperative environments revisited

2. Bryce, C., Oriol, M., & Vitek, J. (1999). A coordination model for agents based on secure spaces (LNCS 1594, pp. 4-20). Berlin: Springer Verlag.

3. Carriero, N., & Gelernter, D. (1986). The S/Net’s Linda kernel. ACM Transactions on Computer Systems.

4. A method for overlapping and erasure of lists

5. A new architecture for concurrent lazy cyclic reference counting on multi-processor systems.;A.de Araújo Formiga;Journal of Universal Computer Science,2007

同舟云学术

1.学者识别学者识别

2.学术分析学术分析

3.人才评估人才评估

"同舟云学术"是以全球学者为主线,采集、加工和组织学术论文而形成的新型学术文献查询和分析系统,可以对全球学者进行文献检索和人才价值评估。用户可以通过关注某些学科领域的顶尖人物而持续追踪该领域的学科进展和研究前沿。经过近期的数据扩容,当前同舟云学术共收录了国内外主流学术期刊6万余种,收集的期刊论文及会议论文总量共计约1.5亿篇,并以每天添加12000余篇中外论文的速度递增。我们也可以为用户提供个性化、定制化的学者数据。欢迎来电咨询!咨询电话:010-8811{复制后删除}0370

www.globalauthorid.com

TOP

Copyright © 2019-2024 北京同舟云网络信息技术有限公司
京公网安备11010802033243号  京ICP备18003416号-3