Pub Date : 1998-12-14DOI: 10.1109/ICPADS.1998.741028
Loon-Been Chen, I-Chen Wu
In a distributed system, identifying consistent checkpoints is essential for error recovery and debugging. We design an efficient incremental algorithm capable of identifying all the consistent and removable checkpoints each time a new checkpoint is reported. By doing so, the required memory space can be minimized by removing those removables. While minimizing the memory space, the algorithm requires only O(p/sup 2/M) time in total, where p is the number of processes and M is the number of checkpoints.
{"title":"An efficient incremental algorithm for identifying consistent checkpoints","authors":"Loon-Been Chen, I-Chen Wu","doi":"10.1109/ICPADS.1998.741028","DOIUrl":"https://doi.org/10.1109/ICPADS.1998.741028","url":null,"abstract":"In a distributed system, identifying consistent checkpoints is essential for error recovery and debugging. We design an efficient incremental algorithm capable of identifying all the consistent and removable checkpoints each time a new checkpoint is reported. By doing so, the required memory space can be minimized by removing those removables. While minimizing the memory space, the algorithm requires only O(p/sup 2/M) time in total, where p is the number of processes and M is the number of checkpoints.","PeriodicalId":226947,"journal":{"name":"Proceedings 1998 International Conference on Parallel and Distributed Systems (Cat. No.98TB100250)","volume":"58 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1998-12-14","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129362351","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 : 1998-12-14DOI: 10.1109/ICPADS.1998.741091
T. Kanezuka, H. Higaki, M. Takizawa
This paper discusses how to make a distributed object system flexible so as to satisfy applications' requirements when changing the system environment. The system change is modeled to be the change of not only types of service but also quality of service (QoS) supported by the objects. There are two types of methods changing the objects, one for manipulating the states of the objects and another for changing QoS of the objects. We discuss new relations among methods with respect to QoS. By using the QoS-based relations, we newly discuss a QoS-based compensating way to recover the object from the less qualified state.
{"title":"Flexible distributed systems for multimedia applications","authors":"T. Kanezuka, H. Higaki, M. Takizawa","doi":"10.1109/ICPADS.1998.741091","DOIUrl":"https://doi.org/10.1109/ICPADS.1998.741091","url":null,"abstract":"This paper discusses how to make a distributed object system flexible so as to satisfy applications' requirements when changing the system environment. The system change is modeled to be the change of not only types of service but also quality of service (QoS) supported by the objects. There are two types of methods changing the objects, one for manipulating the states of the objects and another for changing QoS of the objects. We discuss new relations among methods with respect to QoS. By using the QoS-based relations, we newly discuss a QoS-based compensating way to recover the object from the less qualified state.","PeriodicalId":226947,"journal":{"name":"Proceedings 1998 International Conference on Parallel and Distributed Systems (Cat. No.98TB100250)","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1998-12-14","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"130016148","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 : 1998-12-14DOI: 10.1109/ICPADS.1998.741156
F. Chen, T.-W. Hou
Discusses how a Java execution environment, named Gabi, is designed and implemented. It includes the design and implementation of the interpreter of a Java Virtual Machine (JVM), the class loader, the frame and Java stack, the hash table (class table, native function table and Java string table), the handling of classes.zip, the interface to the native function, the interface to the just-in-time (JIT) compiler and its compiled code for Java methods and the multithreading support, and approaches to boost the performance of the bytecode interpreter by programming skills. Benchmarks are made to compare Gabi with Sun's JRE (Java Runtime Environment).
{"title":"Design, and implementation of a Java execution environment","authors":"F. Chen, T.-W. Hou","doi":"10.1109/ICPADS.1998.741156","DOIUrl":"https://doi.org/10.1109/ICPADS.1998.741156","url":null,"abstract":"Discusses how a Java execution environment, named Gabi, is designed and implemented. It includes the design and implementation of the interpreter of a Java Virtual Machine (JVM), the class loader, the frame and Java stack, the hash table (class table, native function table and Java string table), the handling of classes.zip, the interface to the native function, the interface to the just-in-time (JIT) compiler and its compiled code for Java methods and the multithreading support, and approaches to boost the performance of the bytecode interpreter by programming skills. Benchmarks are made to compare Gabi with Sun's JRE (Java Runtime Environment).","PeriodicalId":226947,"journal":{"name":"Proceedings 1998 International Conference on Parallel and Distributed Systems (Cat. No.98TB100250)","volume":"63 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1998-12-14","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"130459584","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 : 1998-12-14DOI: 10.1109/ICPADS.1998.741124
Chao-Chin Wu, Der-Lin Pean, Cheng Chen
We propose a hardware-centric look-ahead memory consistency model that makes the data consistent according to the special ordering requirement of memory accesses for critical sections. The novel model imposes fewer restrictions on event ordering than previously proposed models thus offering the potential of higher performance. The architecture has the following features: blocking and waking up processes by hardware; allowing instructions to be executed out-of-order; until having acquired the lock can the processor allow the requests for accessing the protected data to be evicted to the memory subsystem. The advantages of the look-ahead model include: more program segments are allowed parallel execution; locks can be released earlier, resulting in reduced waiting times for acquiring locks; and less network traffic because more write requests are merged by using two write caches.
{"title":"Look-ahead memory consistency model","authors":"Chao-Chin Wu, Der-Lin Pean, Cheng Chen","doi":"10.1109/ICPADS.1998.741124","DOIUrl":"https://doi.org/10.1109/ICPADS.1998.741124","url":null,"abstract":"We propose a hardware-centric look-ahead memory consistency model that makes the data consistent according to the special ordering requirement of memory accesses for critical sections. The novel model imposes fewer restrictions on event ordering than previously proposed models thus offering the potential of higher performance. The architecture has the following features: blocking and waking up processes by hardware; allowing instructions to be executed out-of-order; until having acquired the lock can the processor allow the requests for accessing the protected data to be evicted to the memory subsystem. The advantages of the look-ahead model include: more program segments are allowed parallel execution; locks can be released earlier, resulting in reduced waiting times for acquiring locks; and less network traffic because more write requests are merged by using two write caches.","PeriodicalId":226947,"journal":{"name":"Proceedings 1998 International Conference on Parallel and Distributed Systems (Cat. No.98TB100250)","volume":"52 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1998-12-14","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"126808400","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 : 1998-12-14DOI: 10.1109/ICPADS.1998.741128
Sung-Min Lee, Tai-Yun Kim
The World Wide Web has a lot of distributed digital information. When the size of the Web is increased beyond a few sites and a small number of documents, finding information requires time-consuming searching and is too tedious when a user retrieves specific information every day. We propose a news on demand (NOD) service system that gathers daily news information using a robot agent and delivers integrated news to users. Once a user registers his information and preferences, he can get the news information that he is most interested in via multimedia e-mail.
{"title":"A news on demand service system based on robot agent","authors":"Sung-Min Lee, Tai-Yun Kim","doi":"10.1109/ICPADS.1998.741128","DOIUrl":"https://doi.org/10.1109/ICPADS.1998.741128","url":null,"abstract":"The World Wide Web has a lot of distributed digital information. When the size of the Web is increased beyond a few sites and a small number of documents, finding information requires time-consuming searching and is too tedious when a user retrieves specific information every day. We propose a news on demand (NOD) service system that gathers daily news information using a robot agent and delivers integrated news to users. Once a user registers his information and preferences, he can get the news information that he is most interested in via multimedia e-mail.","PeriodicalId":226947,"journal":{"name":"Proceedings 1998 International Conference on Parallel and Distributed Systems (Cat. No.98TB100250)","volume":"43 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1998-12-14","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131453950","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 : 1998-12-14DOI: 10.1109/ICPADS.1998.741096
Tzung-Shi Chen, Neng-Chung Wang, Chih-Ping Chu
In this paper, we propose four efficient multicast routing schemes in wormhole-routed star networks with multidestination routing capability. All of the four proposed schemes are path-based and deadlock-free. The first scheme, dual-path routing, sends the message in parallel through two independent paths. The second one, shortcut-node-based datapath routing, is similar to dual-path routing except that the routing tries to find a shortcut node to route the message as soon as possible to reduce the length of transmission path. The third one, multipath routing, is a multiple dual-path routing strategy that includes source-to-relay and relay-to-destination phases. The last scheme, proximity grouping routing, is similar to multipath routing except that in the partitioning step of source and destination nodes the relation of spatial locality of nodes is also taken into account to reduce the length of transmission paths. Finally, the experimental results are given to show that the performance based on unicast-based and traditional Hamiltonian-path routing schemes can be improved significantly by the four proposed routing schemes respectively.
{"title":"Path-based multicast communication in wormhole-routed star graph multicomputers","authors":"Tzung-Shi Chen, Neng-Chung Wang, Chih-Ping Chu","doi":"10.1109/ICPADS.1998.741096","DOIUrl":"https://doi.org/10.1109/ICPADS.1998.741096","url":null,"abstract":"In this paper, we propose four efficient multicast routing schemes in wormhole-routed star networks with multidestination routing capability. All of the four proposed schemes are path-based and deadlock-free. The first scheme, dual-path routing, sends the message in parallel through two independent paths. The second one, shortcut-node-based datapath routing, is similar to dual-path routing except that the routing tries to find a shortcut node to route the message as soon as possible to reduce the length of transmission path. The third one, multipath routing, is a multiple dual-path routing strategy that includes source-to-relay and relay-to-destination phases. The last scheme, proximity grouping routing, is similar to multipath routing except that in the partitioning step of source and destination nodes the relation of spatial locality of nodes is also taken into account to reduce the length of transmission paths. Finally, the experimental results are given to show that the performance based on unicast-based and traditional Hamiltonian-path routing schemes can be improved significantly by the four proposed routing schemes respectively.","PeriodicalId":226947,"journal":{"name":"Proceedings 1998 International Conference on Parallel and Distributed Systems (Cat. No.98TB100250)","volume":"20 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1998-12-14","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131489697","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 : 1998-12-14DOI: 10.1109/ICPADS.1998.741149
M. Guo, I. Nakata, Yoshiyuki Yamashita
Array redistribution is required often in programs on distributed memory parallel computers. It is essential to use efficient algorithms for redistribution, otherwise the performance of the programs may degrade considerably. The redistribution overheads consist of two parts: index computation and interprocessor communication. If there is no communication scheduling in a redistribution algorithm, the communication contention may occur, which increases the communication waiting time. In order to solve this problem, we propose a technique to schedule the communication so that it becomes contention-free. Our approach initially generates a communication table to represent the communication relations among sending nodes and receiving nodes. According to the communication table, we then generate another table named communication scheduling table. Each column of the communication scheduling table is a permutation of receiving node numbers in each communication step. Thus the communications in our redistribution algorithm are contention-free. Our approach can deal with multi-dimensional shape changing redistribution.
{"title":"Contention-free communication scheduling for array redistribution","authors":"M. Guo, I. Nakata, Yoshiyuki Yamashita","doi":"10.1109/ICPADS.1998.741149","DOIUrl":"https://doi.org/10.1109/ICPADS.1998.741149","url":null,"abstract":"Array redistribution is required often in programs on distributed memory parallel computers. It is essential to use efficient algorithms for redistribution, otherwise the performance of the programs may degrade considerably. The redistribution overheads consist of two parts: index computation and interprocessor communication. If there is no communication scheduling in a redistribution algorithm, the communication contention may occur, which increases the communication waiting time. In order to solve this problem, we propose a technique to schedule the communication so that it becomes contention-free. Our approach initially generates a communication table to represent the communication relations among sending nodes and receiving nodes. According to the communication table, we then generate another table named communication scheduling table. Each column of the communication scheduling table is a permutation of receiving node numbers in each communication step. Thus the communications in our redistribution algorithm are contention-free. Our approach can deal with multi-dimensional shape changing redistribution.","PeriodicalId":226947,"journal":{"name":"Proceedings 1998 International Conference on Parallel and Distributed Systems (Cat. No.98TB100250)","volume":"55 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1998-12-14","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"121853603","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 : 1998-12-14DOI: 10.1109/ICPADS.1998.741162
Jiro Sekiba, G. Kitagata, T. Suganuma, Tetsuo Kinoshita, Ken-ichi Okada, N. Shiratori
Asynchronous messaging systems like e-mail systems need some advanced characteristics, such as intelligence, controllability and scalability, to accomplish more effective and sophisticated message handling. We propose a framework of next generation messaging systems, Flexible Asynchronous Messaging System (FAMES), which consists of autonomous and collaborative software agents. In FAMES, various messaging functions composing of agents can be utilized to integrate heterogeneous user environment. Moreover, FAMES operates message flow in an intelligent manner, considering the receiver's own convenience and that flexible message delivery can be achieved. We designed and implemented the proposed system based on multi agent technology, and we show its effectiveness through experimental studies using a prototype system.
{"title":"Design and implementation of agent-based flexible asynchronous messaging system","authors":"Jiro Sekiba, G. Kitagata, T. Suganuma, Tetsuo Kinoshita, Ken-ichi Okada, N. Shiratori","doi":"10.1109/ICPADS.1998.741162","DOIUrl":"https://doi.org/10.1109/ICPADS.1998.741162","url":null,"abstract":"Asynchronous messaging systems like e-mail systems need some advanced characteristics, such as intelligence, controllability and scalability, to accomplish more effective and sophisticated message handling. We propose a framework of next generation messaging systems, Flexible Asynchronous Messaging System (FAMES), which consists of autonomous and collaborative software agents. In FAMES, various messaging functions composing of agents can be utilized to integrate heterogeneous user environment. Moreover, FAMES operates message flow in an intelligent manner, considering the receiver's own convenience and that flexible message delivery can be achieved. We designed and implemented the proposed system based on multi agent technology, and we show its effectiveness through experimental studies using a prototype system.","PeriodicalId":226947,"journal":{"name":"Proceedings 1998 International Conference on Parallel and Distributed Systems (Cat. No.98TB100250)","volume":"145 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1998-12-14","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127593412","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 : 1998-12-14DOI: 10.1109/ICPADS.1998.741093
Yomin Hou, Chien-Min Wang, Ming Tsai, Lih-Hsing Hsu
For distributed memory parallel computers, broadcast operations are widely used in a variety of applications. In this paper we propose an efficient algorithm for broadcasting on an all-port wormhole-routed 2D torus with arbitrary size. The underlying network is assumed to support only the dimension-ordered unicast. By taking the advantage of the all-port model and the distance insensitivity of the wormhole routing, the proposed algorithm can greatly reduce the number of message-passing steps. In addition, it can be proved to be depth contention-free. The performance study in this paper clearly shows the advantage of the proposed algorithm.
{"title":"Broadcasting on wormhole-routed 2D tori with arbitrary size","authors":"Yomin Hou, Chien-Min Wang, Ming Tsai, Lih-Hsing Hsu","doi":"10.1109/ICPADS.1998.741093","DOIUrl":"https://doi.org/10.1109/ICPADS.1998.741093","url":null,"abstract":"For distributed memory parallel computers, broadcast operations are widely used in a variety of applications. In this paper we propose an efficient algorithm for broadcasting on an all-port wormhole-routed 2D torus with arbitrary size. The underlying network is assumed to support only the dimension-ordered unicast. By taking the advantage of the all-port model and the distance insensitivity of the wormhole routing, the proposed algorithm can greatly reduce the number of message-passing steps. In addition, it can be proved to be depth contention-free. The performance study in this paper clearly shows the advantage of the proposed algorithm.","PeriodicalId":226947,"journal":{"name":"Proceedings 1998 International Conference on Parallel and Distributed Systems (Cat. No.98TB100250)","volume":"17 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1998-12-14","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131867928","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 : 1998-12-14DOI: 10.1109/ICPADS.1998.741106
S. Petrovic
Performances of two explicit rate based congestion control schemes for ABR service in ATM networks: Explicit Rate Indication for Congestion Avoidance (ERICA) and proposed Adaptive Stochastic (AS), are compared under heavy load conditions. Neither of the schemes dictates a particular switch architecture. The requirements tested include: utilization, queuing delay, queuing delay variance and queue size. A simple network topology involves an ATM switching node modeled as a single server queuing system, fed by a number of highly bursty traffic sources. Small and large bursts are used in simulations. The controllable traffic sources' allowed cell rates (ACR) are dynamically shaped by the explicit feedback messages from the switch. The results of a simulation study suggest that AS scheme can provide for higher priority traffic in ATM LANs, shorter queuing delay, queuing delay variance, maximum and mean queue size, while retaining ERICA scheme's utilization.
{"title":"Comparative evaluation of adaptive stochastic and ERICA switch algorithms for ABR traffic management in ATM networks","authors":"S. Petrovic","doi":"10.1109/ICPADS.1998.741106","DOIUrl":"https://doi.org/10.1109/ICPADS.1998.741106","url":null,"abstract":"Performances of two explicit rate based congestion control schemes for ABR service in ATM networks: Explicit Rate Indication for Congestion Avoidance (ERICA) and proposed Adaptive Stochastic (AS), are compared under heavy load conditions. Neither of the schemes dictates a particular switch architecture. The requirements tested include: utilization, queuing delay, queuing delay variance and queue size. A simple network topology involves an ATM switching node modeled as a single server queuing system, fed by a number of highly bursty traffic sources. Small and large bursts are used in simulations. The controllable traffic sources' allowed cell rates (ACR) are dynamically shaped by the explicit feedback messages from the switch. The results of a simulation study suggest that AS scheme can provide for higher priority traffic in ATM LANs, shorter queuing delay, queuing delay variance, maximum and mean queue size, while retaining ERICA scheme's utilization.","PeriodicalId":226947,"journal":{"name":"Proceedings 1998 International Conference on Parallel and Distributed Systems (Cat. No.98TB100250)","volume":"107 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1998-12-14","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131912156","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}