Pub Date : 1989-06-05DOI: 10.1109/ICDCS.1989.37975
A. Viterbi
Models and analytical techniques are developed to evaluate the performance of time-synchronous packet networks with priority queueing disciplines. In cases which require approximation, the model was validated through extensive computer simulation and compared to analytical results. A local first-order Markov model based on a burstiness criterion is used to approximate the delay incurred by each class of traffic in a network. The application of this model to selected networks shows that it accurately predicts behavior of tandem (i.e. with and without departures), unidirectional, and bidirectional loop networks.<>
{"title":"Performance analysis of synchronous packet networks with priority queueing disciplines","authors":"A. Viterbi","doi":"10.1109/ICDCS.1989.37975","DOIUrl":"https://doi.org/10.1109/ICDCS.1989.37975","url":null,"abstract":"Models and analytical techniques are developed to evaluate the performance of time-synchronous packet networks with priority queueing disciplines. In cases which require approximation, the model was validated through extensive computer simulation and compared to analytical results. A local first-order Markov model based on a burstiness criterion is used to approximate the delay incurred by each class of traffic in a network. The application of this model to selected networks shows that it accurately predicts behavior of tandem (i.e. with and without departures), unidirectional, and bidirectional loop networks.<<ETX>>","PeriodicalId":266544,"journal":{"name":"[1989] Proceedings. The 9th International Conference on Distributed Computing Systems","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"1989-06-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"133354518","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 : 1989-06-05DOI: 10.1109/ICDCS.1989.37966
A. Bouloutas, P. Gopal
The problem of partitioning a large computer network into clusters in order to reduce the amount of network resources consumed by the routing algorithm is addressed. The clustering problem is formulated as a general graph partitioning problem. It is shown that the problem of partitioning a graph into a minimum number of clusters with unit weight vertices and a given weight bound on the cluster size is NP-complete if each cluster is required to be internally connected. It is also shown that if a diameter bound is imposed on the cluster instead of the weight bound, then the problem is NP-complete, even when cluster connectivity is not required. An optimum partitioning algorithm is presented for the latter problem when the graph is a tree. An optimum partitioning algorithm is presented for another problem in which each cluster is required to contain exactly one of a set of specified vertices called cluster heads.<>
{"title":"Some graph partitioning problems and algorithms related to routing in large computer networks","authors":"A. Bouloutas, P. Gopal","doi":"10.1109/ICDCS.1989.37966","DOIUrl":"https://doi.org/10.1109/ICDCS.1989.37966","url":null,"abstract":"The problem of partitioning a large computer network into clusters in order to reduce the amount of network resources consumed by the routing algorithm is addressed. The clustering problem is formulated as a general graph partitioning problem. It is shown that the problem of partitioning a graph into a minimum number of clusters with unit weight vertices and a given weight bound on the cluster size is NP-complete if each cluster is required to be internally connected. It is also shown that if a diameter bound is imposed on the cluster instead of the weight bound, then the problem is NP-complete, even when cluster connectivity is not required. An optimum partitioning algorithm is presented for the latter problem when the graph is a tree. An optimum partitioning algorithm is presented for another problem in which each cluster is required to contain exactly one of a set of specified vertices called cluster heads.<<ETX>>","PeriodicalId":266544,"journal":{"name":"[1989] Proceedings. The 9th International Conference on Distributed Computing Systems","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"1989-06-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"122118357","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 : 1989-06-05DOI: 10.1109/ICDCS.1989.37941
Philip M. Thambidurai, You-Keun Park, Kishor S. Trivedi
The problem of predicting the reliability of a distributed system based on the principles of Byzantine agreement is addressed. The system is considered inoperable or failed if Byzantine agreement cannot be guaranteed. The reliability models depend on a unified model of interactive consistency, which is based on a unique fault taxonomy appropriate for distributed systems. The unified model takes advantage of the fact that some faults may not be of an arbitrary nature, while still allowing for the fact that some faults may be arbitrary. A closed-form expression for the reliability and the mean time to failure of systems base on the unified model is derived. Each processor is allowed to have multiple failure modes, and the contribution of the interactive consistency algorithm is explicitly taken into account. The practical value of this unified model in designing ultrareliable systems is demonstrated by several examples.<>
{"title":"On reliability modelling of fault-tolerant distributed systems","authors":"Philip M. Thambidurai, You-Keun Park, Kishor S. Trivedi","doi":"10.1109/ICDCS.1989.37941","DOIUrl":"https://doi.org/10.1109/ICDCS.1989.37941","url":null,"abstract":"The problem of predicting the reliability of a distributed system based on the principles of Byzantine agreement is addressed. The system is considered inoperable or failed if Byzantine agreement cannot be guaranteed. The reliability models depend on a unified model of interactive consistency, which is based on a unique fault taxonomy appropriate for distributed systems. The unified model takes advantage of the fact that some faults may not be of an arbitrary nature, while still allowing for the fact that some faults may be arbitrary. A closed-form expression for the reliability and the mean time to failure of systems base on the unified model is derived. Each processor is allowed to have multiple failure modes, and the contribution of the interactive consistency algorithm is explicitly taken into account. The practical value of this unified model in designing ultrareliable systems is demonstrated by several examples.<<ETX>>","PeriodicalId":266544,"journal":{"name":"[1989] Proceedings. The 9th International Conference on Distributed Computing Systems","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"1989-06-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"125160242","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 : 1989-06-05DOI: 10.1109/ICDCS.1989.37939
Raymond C. Chen, P. Dasgupta
An object/thread based paradigm is presented that links data consistency with object/thread semantics. The paradigm can be used to achieve a wide range of consistency semantics from strict atomic transactions to standard process semantics. The paradigm supports three types of data consistency. Object programmers indicate the type of consistency desired on a per-operation basis, and the system performs automatic concurrency control and recovery management to ensure that those consistency requirements are met. This allows programmers to customize consistency and recovery on a per-application basis without having to supply complicated, custom recovery management schemes. The paradigm allows robust and nonrobust computation to operate concurrently on the same data in a well-defined manner. The operating system need support only one vehicle of computation-the thread.<>
{"title":"Linking consistency with object/thread semantics: an approach to robust computation","authors":"Raymond C. Chen, P. Dasgupta","doi":"10.1109/ICDCS.1989.37939","DOIUrl":"https://doi.org/10.1109/ICDCS.1989.37939","url":null,"abstract":"An object/thread based paradigm is presented that links data consistency with object/thread semantics. The paradigm can be used to achieve a wide range of consistency semantics from strict atomic transactions to standard process semantics. The paradigm supports three types of data consistency. Object programmers indicate the type of consistency desired on a per-operation basis, and the system performs automatic concurrency control and recovery management to ensure that those consistency requirements are met. This allows programmers to customize consistency and recovery on a per-application basis without having to supply complicated, custom recovery management schemes. The paradigm allows robust and nonrobust computation to operate concurrently on the same data in a well-defined manner. The operating system need support only one vehicle of computation-the thread.<<ETX>>","PeriodicalId":266544,"journal":{"name":"[1989] Proceedings. The 9th International Conference on Distributed Computing Systems","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"1989-06-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"120866229","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 : 1989-06-05DOI: 10.1109/ICDCS.1989.37955
B. Jin, Lan Jin
A bit-positional notation is proposed for representing the interconnection topology of a hypercube. This approach greatly helps in the analysis of basic properties of a hypercube network. Some systematic procedures are derived to expand binary trees and map them into higher-dimensional hypercubes. The authors show that an n-cube network can be enhanced by adding at most 2/sup n-2/ diagonal links, which makes it possible to map a complete binary tree into a hypercube with any one of its nodes as the root. The problem of mapping mesh networks into hypercube networks is discussed.<>
{"title":"A new approach to hypercube network analysis","authors":"B. Jin, Lan Jin","doi":"10.1109/ICDCS.1989.37955","DOIUrl":"https://doi.org/10.1109/ICDCS.1989.37955","url":null,"abstract":"A bit-positional notation is proposed for representing the interconnection topology of a hypercube. This approach greatly helps in the analysis of basic properties of a hypercube network. Some systematic procedures are derived to expand binary trees and map them into higher-dimensional hypercubes. The authors show that an n-cube network can be enhanced by adding at most 2/sup n-2/ diagonal links, which makes it possible to map a complete binary tree into a hypercube with any one of its nodes as the root. The problem of mapping mesh networks into hypercube networks is discussed.<<ETX>>","PeriodicalId":266544,"journal":{"name":"[1989] Proceedings. The 9th International Conference on Distributed Computing Systems","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"1989-06-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"121577243","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 : 1989-06-05DOI: 10.1109/ICDCS.1989.37976
W. B. Watson
Using a hardware monitor, the capacity of a mature (over 10 years old), heavily loaded (5 Cray supercomputers, 3 Ethernet gateways, 13 terminal concentrators, 5000 terminal, 40 node) two-trunk, HYPERchannel-based, high-data-rate local area network was tested. Results indicate that the current load could be intensified by 40% before the network reaches saturation. As the load approaches saturation, overly persistent, low-level protocols begin thrashing, effectively destabilizing the network and destroying much of its potential capacity. The steps needed to extend its lifespan and improve its performance under heavy loads are recommended.<>
{"title":"Capacity testing a HYPERchannel-based local area network","authors":"W. B. Watson","doi":"10.1109/ICDCS.1989.37976","DOIUrl":"https://doi.org/10.1109/ICDCS.1989.37976","url":null,"abstract":"Using a hardware monitor, the capacity of a mature (over 10 years old), heavily loaded (5 Cray supercomputers, 3 Ethernet gateways, 13 terminal concentrators, 5000 terminal, 40 node) two-trunk, HYPERchannel-based, high-data-rate local area network was tested. Results indicate that the current load could be intensified by 40% before the network reaches saturation. As the load approaches saturation, overly persistent, low-level protocols begin thrashing, effectively destabilizing the network and destroying much of its potential capacity. The steps needed to extend its lifespan and improve its performance under heavy loads are recommended.<<ETX>>","PeriodicalId":266544,"journal":{"name":"[1989] Proceedings. The 9th International Conference on Distributed Computing Systems","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"1989-06-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"115733150","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 : 1989-06-05DOI: 10.1109/ICDCS.1989.37974
Vikram V. Karmarkar, J. G. Kuhl
A new integrated demand-assignment multiple-access (DAMA) method for multiple-bus local networks (MBLNs) is proposed and evaluated. The performance of this method is shown to be superior to independent token-passing as well as to previously proposed integrated access schemes. V-STIA, a virtual token-passing extension of a modified explicit token-passing scheme called single-token integrated access (STIA), delivers the best overall performance at medium to heavy loads while achieving good light-load performance without collision detection. Consequently, V-STIA makes the fewest possible demands on interface capabilities for DAMA support in MBLNs. Additionally, V-STIA implementation does not require simultaneous transmit capability for correct operation and greatly reduces design complexity. Performance advantages of MBLNs over an equivalent bandwidth single-bus DAMA system are established.<>
{"title":"A high performance virtual token-passing multiple-access method for multiple-bus local networks","authors":"Vikram V. Karmarkar, J. G. Kuhl","doi":"10.1109/ICDCS.1989.37974","DOIUrl":"https://doi.org/10.1109/ICDCS.1989.37974","url":null,"abstract":"A new integrated demand-assignment multiple-access (DAMA) method for multiple-bus local networks (MBLNs) is proposed and evaluated. The performance of this method is shown to be superior to independent token-passing as well as to previously proposed integrated access schemes. V-STIA, a virtual token-passing extension of a modified explicit token-passing scheme called single-token integrated access (STIA), delivers the best overall performance at medium to heavy loads while achieving good light-load performance without collision detection. Consequently, V-STIA makes the fewest possible demands on interface capabilities for DAMA support in MBLNs. Additionally, V-STIA implementation does not require simultaneous transmit capability for correct operation and greatly reduces design complexity. Performance advantages of MBLNs over an equivalent bandwidth single-bus DAMA system are established.<<ETX>>","PeriodicalId":266544,"journal":{"name":"[1989] Proceedings. The 9th International Conference on Distributed Computing Systems","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"1989-06-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"115329909","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 : 1989-06-05DOI: 10.1109/ICDCS.1989.37984
I. A. Cimet, C. Cheng, Srikanta P. R. Kumar
A method to design resilient protocols for spanning tree problems is proposed. The method consists of a generic resilient protocol that can be used for any spanning tree problem. A resilient protocol for a particular spanning tree problem can be obtained by adjusting various run-time parameters of the generic protocol. The generic protocol is designed by using the concepts of replacement sets that contain the links that can be used to replace a link that fails and replacement values that measure the change in the spanning tree when a link in a replacement set is used to recreate a spanning tree. The approach is as general as the common approach (for the case of spanning trees) because the generic protocol can solve a new spanning tree problem without an individual design. The protocols obtained with this method are efficient and easy to implement and verify because they are all derived from the generic protocol.<>
{"title":"On the design of resilient protocols for spanning tree problems","authors":"I. A. Cimet, C. Cheng, Srikanta P. R. Kumar","doi":"10.1109/ICDCS.1989.37984","DOIUrl":"https://doi.org/10.1109/ICDCS.1989.37984","url":null,"abstract":"A method to design resilient protocols for spanning tree problems is proposed. The method consists of a generic resilient protocol that can be used for any spanning tree problem. A resilient protocol for a particular spanning tree problem can be obtained by adjusting various run-time parameters of the generic protocol. The generic protocol is designed by using the concepts of replacement sets that contain the links that can be used to replace a link that fails and replacement values that measure the change in the spanning tree when a link in a replacement set is used to recreate a spanning tree. The approach is as general as the common approach (for the case of spanning trees) because the generic protocol can solve a new spanning tree problem without an individual design. The protocols obtained with this method are efficient and easy to implement and verify because they are all derived from the generic protocol.<<ETX>>","PeriodicalId":266544,"journal":{"name":"[1989] Proceedings. The 9th International Conference on Distributed Computing Systems","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"1989-06-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"125724844","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 : 1989-06-05DOI: 10.1109/ICDCS.1989.37973
P. Leu, B. Bhargava
Concurrent checkpointing and recovery using a concurrent transaction processing model which consists of four types of atomic operation and five types of conflict is developed. Each checkpoint/rollback transaction is executed by multiple processes in the system. They can be executed concurrently. It is shown that the consistency of recovery lines and rollback lines established by checkpoint transactions and rollback transactions can be achieved by enforcing serializability on the corresponding transactions. There are two advantages in using a transaction model for concurrent checkpointing and recovery: (1) it is easier to find algorithms to solve a transaction processing problem; and (2) based on this model, related issues of the two corresponding problems can be thought of uniformly. This model clarifies the concepts of concurrent checkpointing and recovery, and brings more ideas for designing algorithms.<>
{"title":"A model for concurrent checkpointing and recovery using transactions","authors":"P. Leu, B. Bhargava","doi":"10.1109/ICDCS.1989.37973","DOIUrl":"https://doi.org/10.1109/ICDCS.1989.37973","url":null,"abstract":"Concurrent checkpointing and recovery using a concurrent transaction processing model which consists of four types of atomic operation and five types of conflict is developed. Each checkpoint/rollback transaction is executed by multiple processes in the system. They can be executed concurrently. It is shown that the consistency of recovery lines and rollback lines established by checkpoint transactions and rollback transactions can be achieved by enforcing serializability on the corresponding transactions. There are two advantages in using a transaction model for concurrent checkpointing and recovery: (1) it is easier to find algorithms to solve a transaction processing problem; and (2) based on this model, related issues of the two corresponding problems can be thought of uniformly. This model clarifies the concepts of concurrent checkpointing and recovery, and brings more ideas for designing algorithms.<<ETX>>","PeriodicalId":266544,"journal":{"name":"[1989] Proceedings. The 9th International Conference on Distributed Computing Systems","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"1989-06-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"126583594","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 : 1989-06-05DOI: 10.1109/ICDCS.1989.37965
H. Garcia-Molina, Annemarie Spauster
Three ordering properties are characterized, namely, single-source, multiple-source, and multiple-group ordering, and their solutions are discussed. The multiple-group ordering property, which guarantees that two messages destined to two processes are delivered in the same relative order, even if they originate at different sources and are addressed to different multicast groups, is concentrated on. A protocol that solves the multiple-group ordering problem is presented. The issues of performance and reliability are addressed by providing comparisons with other techniques for ordering multicasts. In many cases this new algorithm solves the problem with greater efficiency than previous solutions without sacrificing reliability. It is pointed out that one disadvantage of the technique is that sometimes sites are required to handle messages which they do not need to deliver locally. These so-called extra nodes, however, do not occur frequently according to the experiments presented.<>
{"title":"Message ordering in a multicast environment","authors":"H. Garcia-Molina, Annemarie Spauster","doi":"10.1109/ICDCS.1989.37965","DOIUrl":"https://doi.org/10.1109/ICDCS.1989.37965","url":null,"abstract":"Three ordering properties are characterized, namely, single-source, multiple-source, and multiple-group ordering, and their solutions are discussed. The multiple-group ordering property, which guarantees that two messages destined to two processes are delivered in the same relative order, even if they originate at different sources and are addressed to different multicast groups, is concentrated on. A protocol that solves the multiple-group ordering problem is presented. The issues of performance and reliability are addressed by providing comparisons with other techniques for ordering multicasts. In many cases this new algorithm solves the problem with greater efficiency than previous solutions without sacrificing reliability. It is pointed out that one disadvantage of the technique is that sometimes sites are required to handle messages which they do not need to deliver locally. These so-called extra nodes, however, do not occur frequently according to the experiments presented.<<ETX>>","PeriodicalId":266544,"journal":{"name":"[1989] Proceedings. The 9th International Conference on Distributed Computing Systems","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"1989-06-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129376073","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}