Pub Date : 1990-01-01DOI: 10.1109/ICDCS.1990.89270
L. Shang, C. Fan, Zhongxiu Sun
{"title":"Towards a Combinative Distributed Operating System in Cluster 86","authors":"L. Shang, C. Fan, Zhongxiu Sun","doi":"10.1109/ICDCS.1990.89270","DOIUrl":"https://doi.org/10.1109/ICDCS.1990.89270","url":null,"abstract":"","PeriodicalId":6300,"journal":{"name":"2012 IEEE 32nd International Conference on Distributed Computing Systems","volume":"16 1","pages":"175-182"},"PeriodicalIF":0.0,"publicationDate":"1990-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"72757931","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 1987-06-08DOI: 10.1007/3-540-18991-2_31
L. Brochard
{"title":"Communication and Control Costs of Domain Decomposition on Loosely Coupled Multiprocessors","authors":"L. Brochard","doi":"10.1007/3-540-18991-2_31","DOIUrl":"https://doi.org/10.1007/3-540-18991-2_31","url":null,"abstract":"","PeriodicalId":6300,"journal":{"name":"2012 IEEE 32nd International Conference on Distributed Computing Systems","volume":"198 1","pages":"544-551"},"PeriodicalIF":0.0,"publicationDate":"1987-06-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"76036519","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
This paper describes a process migration facility for the Sprite operating system. In order to provide location-transparent remote execution, Sprite associates with each process a distinguished home node, which provides kernel services to the process throughout the process''s lifetime. System calls that depend on the location of a process are forwarded to the process''s home node. Performance measurements based on a few simple benchmarks show that remote execution using the home node model is efficient as long as the number of system calls that must be forwarded home is small; this appears to be the case as long as file-system-related calls can be handled without involving the home node. The benchmarks also show that the cost of migrating a process can vary from a fraction of a second to many seconds; it is determined primarily by the number of dirty virtual memory pages and file blocks associated with the process.
{"title":"Process Migration in the Sprite Operating System","authors":"F. Douglis, J. Ousterhout","doi":"10.21236/ada619399","DOIUrl":"https://doi.org/10.21236/ada619399","url":null,"abstract":"This paper describes a process migration facility for the Sprite operating system. In order to provide location-transparent remote execution, Sprite associates with each process a distinguished home node, which provides kernel services to the process throughout the process''s lifetime. System calls that depend on the location of a process are forwarded to the process''s home node. Performance measurements based on a few simple benchmarks show that remote execution using the home node model is efficient as long as the number of system calls that must be forwarded home is small; this appears to be the case as long as file-system-related calls can be handled without involving the home node. The benchmarks also show that the cost of migrating a process can vary from a fraction of a second to many seconds; it is determined primarily by the number of dirty virtual memory pages and file blocks associated with the process.","PeriodicalId":6300,"journal":{"name":"2012 IEEE 32nd International Conference on Distributed Computing Systems","volume":"73 1","pages":"18-27"},"PeriodicalIF":0.0,"publicationDate":"1987-02-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"80658840","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
G. Roman, M. E. Ehlers, H. C. Cunningham, R. Lykins
A new approach to modelling distributed systems is presented. It uses sequential processes and event synchronization as building blocks to construct a cohesive picture of the interdependent requirements for the functionality, architecture, scheduling policies, and performance attributes of a distributed system. A language called CSPS (an extension of Hoare's CSP) is used in the illustration of the approach. Employing CSP as a base allows modelled systems to be verified using techniques already developed for verifying CSP programs and leads to the emergence of a uniform incremental strategy for verifying both logical and performance properties of distributed systems. Several small... Read complete abstract on page 2.
{"title":"Toward Comprehensive Specification of Distributed Systems","authors":"G. Roman, M. E. Ehlers, H. C. Cunningham, R. Lykins","doi":"10.7936/K76D5R8T","DOIUrl":"https://doi.org/10.7936/K76D5R8T","url":null,"abstract":"A new approach to modelling distributed systems is presented. It uses sequential processes and event synchronization as building blocks to construct a cohesive picture of the interdependent requirements for the functionality, architecture, scheduling policies, and performance attributes of a distributed system. A language called CSPS (an extension of Hoare's CSP) is used in the illustration of the approach. Employing CSP as a base allows modelled systems to be verified using techniques already developed for verifying CSP programs and leads to the emergence of a uniform incremental strategy for verifying both logical and performance properties of distributed systems. Several small... Read complete abstract on page 2.","PeriodicalId":6300,"journal":{"name":"2012 IEEE 32nd International Conference on Distributed Computing Systems","volume":"114 1","pages":"282-291"},"PeriodicalIF":0.0,"publicationDate":"1987-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"78087093","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
This paper describes the election algorithm that guarantees the reliability of TEMPO, a distributed clock synchronizer running on Berkeley UNIX 4.3BSD systems. TEMPO is a distributed program based on a master-slave scheme that is comprised of time daemon processes running on individual machines. The election algorithm chooses a new master from among the slaves after the crash of the machine on which the original master was running. When the master is working, it periodically resets an election timer in each slave. If the master disappears, the sl ave whose timer expires first will become a candidate for the new master. The election algorithm covers this normal case, as well as the infrequent case where there may be two or more simultaneous candidates. It also handles the case in which, due to a network partition that has been repaired, two masters are present at the same time.
{"title":"An Election Algorithm for a Distributed Clock Synchronization Program","authors":"R. Gusella, S. Zatti","doi":"10.21236/ada611781","DOIUrl":"https://doi.org/10.21236/ada611781","url":null,"abstract":"This paper describes the election algorithm that guarantees the reliability of TEMPO, a distributed clock synchronizer running on Berkeley UNIX 4.3BSD systems. TEMPO is a distributed program based on a master-slave scheme that is comprised of time daemon processes running on individual machines. The election algorithm chooses a new master from among the slaves after the crash of the machine on which the original master was running. When the master is working, it periodically resets an election timer in each slave. If the master disappears, the sl ave whose timer expires first will become a candidate for the new master. The election algorithm covers this normal case, as well as the infrequent case where there may be two or more simultaneous candidates. It also handles the case in which, due to a network partition that has been repaired, two masters are present at the same time.","PeriodicalId":6300,"journal":{"name":"2012 IEEE 32nd International Conference on Distributed Computing Systems","volume":"35 1","pages":"364-371"},"PeriodicalIF":0.0,"publicationDate":"1985-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"78937400","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Abstract : Networks of computers running Berkeley UNIX allow users to program and run multiple-process applications that execute concurrently on several machines. We present solutions to the problems of process tracking, administration, and control in this networked computing environment. We have designed and implemented a personal process manager for an enhanced Berkeley UNIX system that provides the user with much needed process management and process control capabilities not found elsewhere. The personal process manager is a distributed program based on a collection of user processes which make use of specialized system daemons. It provides on demand services, allows process control across machine boundaries, and may outlive the user login session in which it was created. When active, it becomes the process creation server for a user's remote processes, collects and preserves basic information about process activities, provides a notion of state of a distributed computation, and interfaces with several data analysis and data representation tools. The personal process manager also has crash recovery facilities.
{"title":"The Administration of Distributed Computations in a Networked Environment: An Interim Report","authors":"L. Cabrera, S. Sechrest, R. Cáceres","doi":"10.21236/ada611778","DOIUrl":"https://doi.org/10.21236/ada611778","url":null,"abstract":"Abstract : Networks of computers running Berkeley UNIX allow users to program and run multiple-process applications that execute concurrently on several machines. We present solutions to the problems of process tracking, administration, and control in this networked computing environment. We have designed and implemented a personal process manager for an enhanced Berkeley UNIX system that provides the user with much needed process management and process control capabilities not found elsewhere. The personal process manager is a distributed program based on a collection of user processes which make use of specialized system daemons. It provides on demand services, allows process control across machine boundaries, and may outlive the user login session in which it was created. When active, it becomes the process creation server for a user's remote processes, collects and preserves basic information about process activities, provides a notion of state of a distributed computation, and interfaces with several data analysis and data representation tools. The personal process manager also has crash recovery facilities.","PeriodicalId":6300,"journal":{"name":"2012 IEEE 32nd International Conference on Distributed Computing Systems","volume":"42 1","pages":"389-397"},"PeriodicalIF":0.0,"publicationDate":"1985-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"85762731","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Abstract : Prefix tables provide a mechanism for locating files in a system whose storage is distributed among many servers. The result is a single file system hierarchy visible uniformly and transparently to all clients. Each client of the filesystem maintains a local prefix table that identifies the server for a file based on the initial part of the file name. Prefix tables are built and modified using a simple broadcast protocol that is flexible enough to allow dynamic server reconfiguration and a simple form of replication.
{"title":"Prefix Tables: A Simple Mechanism for Locating Files in a Distributed System","authors":"B. Welch, J. Ousterhout","doi":"10.21236/ada611774","DOIUrl":"https://doi.org/10.21236/ada611774","url":null,"abstract":"Abstract : Prefix tables provide a mechanism for locating files in a system whose storage is distributed among many servers. The result is a single file system hierarchy visible uniformly and transparently to all clients. Each client of the filesystem maintains a local prefix table that identifies the server for a file based on the initial part of the file name. Prefix tables are built and modified using a simple broadcast protocol that is flexible enough to allow dynamic server reconfiguration and a simple form of replication.","PeriodicalId":6300,"journal":{"name":"2012 IEEE 32nd International Conference on Distributed Computing Systems","volume":"295 1","pages":"184-189"},"PeriodicalIF":0.0,"publicationDate":"1985-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"73682988","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Abstract : An architecture is described for a multi-processor implementation of real-time signal processing algorithms. A butterfly network is used to provide simultaneous, conflict-free interprocessor communication for multi-dimensional convolution and Fourier transformation. A hardware demonstration test-bed using four active processors was used to validate the concepts of shared algorithm execution, conflict-free data transfers, distributed network control, dynamic fault tolerance, and identical software in all processors. The Distributed Signal Processor (DISP) is designed to perform real-time signal processing on large data sets. Potential applications for the DISP include image processing (enhancement, restoration, segmentations, and coding), radar imaging, synthetic aperture radar, and infrared imaging (passive or active). These applications share three important characteristics. First, the data sets are large, generally multi-dimensional, with high computation rates. Second, the applications make extensive use of convolution and Fourier transformation algorithms. Finally, real-time signal processing applications generally have a regular, well defined time-line.
{"title":"A distributed signal processing architecture","authors":"A. Filip","doi":"10.21236/ada130227","DOIUrl":"https://doi.org/10.21236/ada130227","url":null,"abstract":"Abstract : An architecture is described for a multi-processor implementation of real-time signal processing algorithms. A butterfly network is used to provide simultaneous, conflict-free interprocessor communication for multi-dimensional convolution and Fourier transformation. A hardware demonstration test-bed using four active processors was used to validate the concepts of shared algorithm execution, conflict-free data transfers, distributed network control, dynamic fault tolerance, and identical software in all processors. The Distributed Signal Processor (DISP) is designed to perform real-time signal processing on large data sets. Potential applications for the DISP include image processing (enhancement, restoration, segmentations, and coding), radar imaging, synthetic aperture radar, and infrared imaging (passive or active). These applications share three important characteristics. First, the data sets are large, generally multi-dimensional, with high computation rates. Second, the applications make extensive use of convolution and Fourier transformation algorithms. Finally, real-time signal processing applications generally have a regular, well defined time-line.","PeriodicalId":6300,"journal":{"name":"2012 IEEE 32nd International Conference on Distributed Computing Systems","volume":"23 1","pages":"49-55"},"PeriodicalIF":0.0,"publicationDate":"1983-05-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"78224721","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
DPL-82 is a language for composing programs of concurrently- executing processes. Processes may be all on a single machine or may be distributed over a set of processors connected to a network. The semantics of the language is derived from the underlying interprocess communication facility (IPC) and from the dataflow model of computation. This paper discusses the major concepts of the language, namely nodes, arcs, connections, tokens, signals, and activations, and presents examples which illustrate the construction of distributed programs in DPL-82 with internal arcs, external arcs and child arcs. Features for process-to- processor mapping and dead process restart are mentioned. The paper concludes with some ideas for future research.
{"title":"DPL-82: A language for distributed processing","authors":"L. Ericson","doi":"10.21236/ada120122","DOIUrl":"https://doi.org/10.21236/ada120122","url":null,"abstract":"DPL-82 is a language for composing programs of concurrently- executing processes. Processes may be all on a single machine or may be distributed over a set of processors connected to a network. The semantics of the language is derived from the underlying interprocess communication facility (IPC) and from the dataflow model of computation. This paper discusses the major concepts of the language, namely nodes, arcs, connections, tokens, signals, and activations, and presents examples which illustrate the construction of distributed programs in DPL-82 with internal arcs, external arcs and child arcs. Features for process-to- processor mapping and dead process restart are mentioned. The paper concludes with some ideas for future research.","PeriodicalId":6300,"journal":{"name":"2012 IEEE 32nd International Conference on Distributed Computing Systems","volume":"76 1","pages":"526-531"},"PeriodicalIF":0.0,"publicationDate":"1982-07-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"86510367","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：