{"title":"A comparison of priority-based decentralized load balancing policies","authors":"K. J. Lee, D. Towsley","doi":"10.1145/317499.317540","DOIUrl":"https://doi.org/10.1145/317499.317540","url":null,"abstract":"Load balancing policies in distributed systems divide jobs into two classes; those processed at their of origination (local jobs) and those processed at some other site in the system after being transfered through a communication network (remote jobs). This paper considers a class of decentralized load balancing policies that use a threshold on the local job queue length at each host in making decisions for remote processing. They differ from each other according to how they assign priorities to each of these job classes, ranging from one providing favorable treatment to local jobs to one providing favorable treatment to remote jobs. Under each policy, the optimal load balancing problem is formulated as an optimization problem with respect to the threshold parameter. The optimal threshold is obtained numerically using matrix-geometric formulation and an iteration method. Last, we consider the effects that the job arrival process can have on performance. One expects that load balancing for systems operating in an environment of bursty job arrivals should be more beneficial than for an environment with random job arrivals. This fact is observed through numerical examples.","PeriodicalId":445002,"journal":{"name":"SIGMETRICS '86/PERFORMANCE '86","volume":"3 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1986-05-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"114544291","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}
{"title":"Queueing analysis of nonpreemptive reservation priority discipline","authors":"H. Takagi, M. Murata","doi":"10.1145/317499.317556","DOIUrl":"https://doi.org/10.1145/317499.317556","url":null,"abstract":"Analysis is given to a nonpreemptive priority queueing system with P classes of messages where the class of message to be served next is the highest priority class waiting at the time of service start. (If this were the highest priority class waiting at the service completion epoch, we would have a classical nonpreemptive head-of-line priority queueing system.) We assume that the message service time distribution is identical for all classes. The mean message waiting time is obtained explicitly for each class, and numerically compared to the values in the corresponding head-of-line system. We have also proposed and evaluated a fairness measure to demonstrate the degree of discrimination. This model can be applied to the performance analysis of the prioritized token-ring scheme in local area computer networks when the propagation delay and bit latency are negligible compared to the frame transmission time.","PeriodicalId":445002,"journal":{"name":"SIGMETRICS '86/PERFORMANCE '86","volume":"20 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1986-05-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"125033837","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}
{"title":"Measurement and analysis of locality phases in file referencing behaviour","authors":"S. Majumdar, R. Bunt","doi":"10.1145/317499.317551","DOIUrl":"https://doi.org/10.1145/317499.317551","url":null,"abstract":"Recent research has demonstrated the existence of locality in short-term file referencing behaviour. A detailed study of the dynamic characteristics of file referencing is presented in this paper. The concept of Bounded Locality Intervals from the field of program behaviour has been used to model the locality phases of file referencing behaviour. The model is found to be powerful both from a descriptive point of view and from the perspective of understanding the performance implications of locality properties of file referencing behaviour on file system management.","PeriodicalId":445002,"journal":{"name":"SIGMETRICS '86/PERFORMANCE '86","volume":"27 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1986-05-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"125876934","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}
{"title":"Analysis of a conveyor queue in a flexible manufacturing system","authors":"E. Coffman, E. Gelenbe, E. Gilbert","doi":"10.1145/317499.317554","DOIUrl":"https://doi.org/10.1145/317499.317554","url":null,"abstract":"In a flexible manufacturing system stations are arranged along a common conveyor that brings items for processing to the stations and also carries away the processed items. At each station specialized robots automatically load and unload items on and off the conveyor. We examine here a single station in such a system. A new kind of queueing problem arises, with input-output dependencies that result because the same conveyor transports items both to and from the station. The paper analyzes two models of a station. Model 1 has one robot that cannot return a processed item to the conveyor while unloading a new item for processing. Model 2 has two robots to allow simultaneous loading and unloading of the conveyor. A principal goal of the analysis is the proper choice of the distance separating the two points at which items leave and rejoin the conveyor.","PeriodicalId":445002,"journal":{"name":"SIGMETRICS '86/PERFORMANCE '86","volume":"27 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1986-05-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127662491","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}
{"title":"Performance analysis of multiprocessor cache consistency protocols using generalized timed Petri nets","authors":"M. Vernon, M. A. Holliday","doi":"10.1145/317499.317534","DOIUrl":"https://doi.org/10.1145/317499.317534","url":null,"abstract":"We use an exact analytical technique, based on Generalized Timed Petri Nets (GTPNs), to study the performance of shared bus cache consistency protocols for multiprocessors. We develop a general framework within which the key characteristics of the Write-Once protocol and four enhancements that have been combined in various ways in the literature can be identified and evaluated. We then quantitatively assess the performance gains for each of the four enhancements. We consider three levels of data sharing in our workload models. One of the enhancements substantially improves system performance in all cases. Two enhancements are shown to have negligible effect over the range of workloads analyzed. The fourth enhancement shows a small improvement for low levels of sharing, but shows more substantial improvement as sharing is increased, if we assume a “good access pattern”. The effects of two architectural parameters, the blocksize and the main memory cycle time are also considered.","PeriodicalId":445002,"journal":{"name":"SIGMETRICS '86/PERFORMANCE '86","volume":"13 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1986-05-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"122072205","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}
{"title":"Measuring operating system performance on modern micro-processors","authors":"R. Razouk, Terri Stewart, M. Wilson","doi":"10.1145/317499.317552","DOIUrl":"https://doi.org/10.1145/317499.317552","url":null,"abstract":"The use of micro-processors and commercial operating systems in real-time applications demands a good understanding of factors which influence software performance. Advances in micro-processor design (e.g. pipelining) make performance prediction based on instruction cycle counts difficult. In addition, the increasing complexity of operating systems raises doubts about our ability to ensure that their performance will meet system requirements. Performance measurement is more important than ever.\u0000This paper describes an ongoing project intended to use performance measurements to characterize the performance of real-time systems software. To date the project has conducted extensive experiments on an in-house operating system running on Intel's 286/10 micro-computer in order to test the feasibility of accurate and repeatable measurement of O/S performance. The measurement approach, which views the software from a resource-consumption standpoint, can be applied to both O/S and application level software. Some of the measurement results are presented here and are used to test the manufacturer's assumptions about the hardware's performance.","PeriodicalId":445002,"journal":{"name":"SIGMETRICS '86/PERFORMANCE '86","volume":"8 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1986-05-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127464290","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}
{"title":"A probabilistic analysis of a database system","authors":"W. A. Massey","doi":"10.1145/317499.317547","DOIUrl":"https://doi.org/10.1145/317499.317547","url":null,"abstract":"In Gray, Homan, Obermarck, and Korth [GHOK], the authors give many conjectures based on simulation for the probabilistic analysis of transaction lock-waits and transaction deadlocks. In this paper, we introduce a probabilistic model to explain their observations.","PeriodicalId":445002,"journal":{"name":"SIGMETRICS '86/PERFORMANCE '86","volume":"123 1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1986-05-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"124521987","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}
{"title":"Performance evaluation of multiversion with the Oracle synchronization","authors":"Andrew Witkowski","doi":"10.1145/317499.317548","DOIUrl":"https://doi.org/10.1145/317499.317548","url":null,"abstract":"In this paper we present a new analytical model for performance measurements of timestamp driven databases. The model is based on two-dimensional Poisson processes where one coordinate represents the real arrival time and the other the timestamp of an arriving messages. The notion of preemption is defined which serves as a model for synchronization. Preemption naturally implies such performance measures as response time and amount of abortion in the system. The concept of oracle is introduced which allows evaluation of a lower bound on the synchronization cost. Preemption and the oracle are then used to evaluate performance of the Multiversion synchronization. We present the distribution and the expectation of the synchronization cost. The analysis is then applied to a database with exponential communication delays (α) and the intensity of transaction λ. It is shown that for Multiversion, this cost depends linearly on l/α and logarithmically on λ.","PeriodicalId":445002,"journal":{"name":"SIGMETRICS '86/PERFORMANCE '86","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1986-05-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"130570876","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}
{"title":"Load-balancing heuristics and process behavior","authors":"W. Leland, T. Ott","doi":"10.1145/317499.317539","DOIUrl":"https://doi.org/10.1145/317499.317539","url":null,"abstract":"Dynamic load balancing in a system of loosely-coupled homogeneous processors may employ both judicious initial placement of processes and migration of existing processes to processors with fewer resident processes. In order to predict the possible benefits of these dynamic assignment techniques, we analyzed the behavior (CPU, disk, and memory use) of 9.5 million Unix* processes during normal use. The observed process behavior was then used to drive simulation studies of particular dynamic assignment heuristics.\u0000Let <italic>F</italic>(·) be the probability distribution of the amount of CPU time used by an arbitrary process. In the environment studied we found:<list><item>(1-<italic>F</italic>(<italic>x</italic>)) ≉ <italic>rx</italic><supscrpt>-<italic>c</italic></supscrpt>, 1.05<<italic>c</italic><1.25.\u0000</item><item><italic>F</italic>(·) is far enough from exponential to make exponential models of little use.\u0000</item><item>With a foreground-background process scheduling policy in each processor, simple heuristics for initial placement and processor migration can significantly improve the response ratios of processes that demand exceptional amounts of CPU, without harming the response ratios of ordinary processes.\u0000</item></list>","PeriodicalId":445002,"journal":{"name":"SIGMETRICS '86/PERFORMANCE '86","volume":"74 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1986-05-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"126170203","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}
{"title":"Modeling multiprocessor computer systems with unbalanced flows","authors":"S. Madnick, Richard Y. Wang","doi":"10.1145/317499.317536","DOIUrl":"https://doi.org/10.1145/317499.317536","url":null,"abstract":"A performance analysis methodology using certain aspects of queueing theory to evaluate computer system speed performance is presented. This methodology specifically focuses on modeling multiprocessor computer systems with unbalanced flows (i.e., number of transactions leaving a server is not the same as number of transactions entering that server) due to asynchronously spawned parallel tasks. This unbalanced flow phenomenon, which has a significant effect on performance, cannot be solved analytically by classical queueing network models.\u0000A decomposition method is applied to decompose the unbalanced flows. Formulae for open queueing networks with unbalanced flows due to asynchronously spawned tasks are developed. Furthermore, an algorithm based on Buzen's convolution algorithm is developed to test the necessary and sufficient condition for closed system stability as well as to compute performance measures. An average of less than four iterations is reported for convergence with this algorithm.\u0000A Study of the INFOPLEX multiprocessor data storage hierarchy, comparing this rapid solution algorithm with simulations, has shown highly consistent results. A cost effective software tool, using this methodology, has been developed to analyze an architectural design, such as INFOPLEX, and to produce measures such as throughput, utilization, and response time so that potential performance problems can be identified.","PeriodicalId":445002,"journal":{"name":"SIGMETRICS '86/PERFORMANCE '86","volume":"11 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1986-05-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"132409165","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}