Pub Date : 2001-08-15DOI: 10.1109/MASCOT.2001.948892
S. Galmés, R. Puigjaner
The performance of a queueing system fed by an aggregate ATM model is considered. Particularly, a (batch- on/off)/D/1 queue is analyzed in terms of the complete response time distribution. The analysis follows an exact decomposition approach, where the response time is evaluated as the superposition of the contribution of single bursts (small time-scale effects) and the contribution of the interaction between bursts (large timescale effects). For the contribution of single bursts, an exact closed-formula is obtained. The interaction between bursts is modeled by means of a Markov chain, which in fact corresponds to a general random walk. The expressions obtained in this paper will help in providing a better understanding of the relationships between traffic and performance parameters.
{"title":"Performance evaluation based on an aggregate ATM model","authors":"S. Galmés, R. Puigjaner","doi":"10.1109/MASCOT.2001.948892","DOIUrl":"https://doi.org/10.1109/MASCOT.2001.948892","url":null,"abstract":"The performance of a queueing system fed by an aggregate ATM model is considered. Particularly, a (batch- on/off)/D/1 queue is analyzed in terms of the complete response time distribution. The analysis follows an exact decomposition approach, where the response time is evaluated as the superposition of the contribution of single bursts (small time-scale effects) and the contribution of the interaction between bursts (large timescale effects). For the contribution of single bursts, an exact closed-formula is obtained. The interaction between bursts is modeled by means of a Markov chain, which in fact corresponds to a general random walk. The expressions obtained in this paper will help in providing a better understanding of the relationships between traffic and performance parameters.","PeriodicalId":375127,"journal":{"name":"MASCOTS 2001, Proceedings Ninth International Symposium on Modeling, Analysis and Simulation of Computer and Telecommunication Systems","volume":"145 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2001-08-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"132451363","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 : 2001-08-15DOI: 10.1109/MASCOT.2001.948860
Kyungtae Woo, Chansu Yu, Dongman Lee, H. Youn, Ben Lee
As mobile computing requires more computation as well as communication activities, energy efficiency becomes the most critical issue for battery-operated mobile devices. Specifically, in ad hoc networks where each node is responsible for forwarding neighbor nodes' data packets, care has to be taken not only to reduce the overall energy consumption of all relevant nodes but also to balance individual battery levels. Unbalanced energy usage will result in earlier node failure in overloaded nodes, and in turn may lead to network partitioning and reduced network lifetime. This paper presents a new routing algorithm, called local energy-aware routing (LEAR), which achieves a trade-off between balanced energy consumption and shortest routing delay, and at the same time avoids the blocking and route cache problems. Our performance study based on GloMoSim simulator shows that compared to DSR the proposed LEAR improves the energy balance 1.0-35% depending on node mobility.
{"title":"Non-blocking, localized routing algorithm for balanced energy consumption in mobile ad hoc networks","authors":"Kyungtae Woo, Chansu Yu, Dongman Lee, H. Youn, Ben Lee","doi":"10.1109/MASCOT.2001.948860","DOIUrl":"https://doi.org/10.1109/MASCOT.2001.948860","url":null,"abstract":"As mobile computing requires more computation as well as communication activities, energy efficiency becomes the most critical issue for battery-operated mobile devices. Specifically, in ad hoc networks where each node is responsible for forwarding neighbor nodes' data packets, care has to be taken not only to reduce the overall energy consumption of all relevant nodes but also to balance individual battery levels. Unbalanced energy usage will result in earlier node failure in overloaded nodes, and in turn may lead to network partitioning and reduced network lifetime. This paper presents a new routing algorithm, called local energy-aware routing (LEAR), which achieves a trade-off between balanced energy consumption and shortest routing delay, and at the same time avoids the blocking and route cache problems. Our performance study based on GloMoSim simulator shows that compared to DSR the proposed LEAR improves the energy balance 1.0-35% depending on node mobility.","PeriodicalId":375127,"journal":{"name":"MASCOTS 2001, Proceedings Ninth International Symposium on Modeling, Analysis and Simulation of Computer and Telecommunication Systems","volume":"47 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2001-08-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"130105676","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 : 2001-08-15DOI: 10.1109/MASCOT.2001.948856
Yang Xiao, C. L. P. Chen
Adaptive multimedia is promising in wireless/mobile networks since it mitigates the fluctuation of resources caused by the mobility in wireless/mobile networks. However, its bandwidth adaptation causes bandwidth degradation. In order to fully characterize the bandwidth degradation and to provide better QoS to service users, we propose two new QoS parameters for adaptive multimedia: the degradation ratio (DR) and the degradation degree (DD), which characterize both the frequency of degradation and the degree of degradation. A measurement-based call admission control framework and a K-level bandwidth adaptation algorithm (KL-BAA) are also proposed to satisfy the proposed QoS parameters' requirements, and to utilize the resource efficiently. Moreover, the measurement-based formulas for DR and DD are provided using the time averaging method. Simulation results show that QoS requirements are satisfied, and KL-BAA outperforms the two-level bandwidth adaptation algorithm. Simulation results are also compared with those from the non-adaptive multimedia services and reveal that the adaptive multimedia framework outperforms the non-adaptive multimedia services.
{"title":"QoS for adaptive multimedia in wireless/mobile networks","authors":"Yang Xiao, C. L. P. Chen","doi":"10.1109/MASCOT.2001.948856","DOIUrl":"https://doi.org/10.1109/MASCOT.2001.948856","url":null,"abstract":"Adaptive multimedia is promising in wireless/mobile networks since it mitigates the fluctuation of resources caused by the mobility in wireless/mobile networks. However, its bandwidth adaptation causes bandwidth degradation. In order to fully characterize the bandwidth degradation and to provide better QoS to service users, we propose two new QoS parameters for adaptive multimedia: the degradation ratio (DR) and the degradation degree (DD), which characterize both the frequency of degradation and the degree of degradation. A measurement-based call admission control framework and a K-level bandwidth adaptation algorithm (KL-BAA) are also proposed to satisfy the proposed QoS parameters' requirements, and to utilize the resource efficiently. Moreover, the measurement-based formulas for DR and DD are provided using the time averaging method. Simulation results show that QoS requirements are satisfied, and KL-BAA outperforms the two-level bandwidth adaptation algorithm. Simulation results are also compared with those from the non-adaptive multimedia services and reveal that the adaptive multimedia framework outperforms the non-adaptive multimedia services.","PeriodicalId":375127,"journal":{"name":"MASCOTS 2001, Proceedings Ninth International Symposium on Modeling, Analysis and Simulation of Computer and Telecommunication Systems","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2001-08-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"130447389","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 : 2001-08-15DOI: 10.1109/MASCOT.2001.948895
Jehan-Francois Pâris
Stream tapping-also known as patching-can reduce the bandwidth requirements of video-on-demand services by allowing new customer requests to "tap" the data streams of other requests for the same video. Previous studies have shown that stream tapping works best when the request arrival rate does not exceed ten to twenty requests per hour for a two-hour video. At higher arrival rates, it performs much worse than broadcasting protocols. To overcome this limitation, we propose a stream tapping protocol that preloads in the customer set-top box the first few minutes of all popular videos. To offset the cost of the additional buffer space, our protocol never requires the set-top box to receive data from the video server at more than twice video consumption rate. Our simulations indicate that preloading the first eight minutes of a two-hour video was enough to achieve lower bandwidth requirements than the best broadcasting protocols at any request arrival rate.
{"title":"A stream tapping protocol with partial preloading","authors":"Jehan-Francois Pâris","doi":"10.1109/MASCOT.2001.948895","DOIUrl":"https://doi.org/10.1109/MASCOT.2001.948895","url":null,"abstract":"Stream tapping-also known as patching-can reduce the bandwidth requirements of video-on-demand services by allowing new customer requests to \"tap\" the data streams of other requests for the same video. Previous studies have shown that stream tapping works best when the request arrival rate does not exceed ten to twenty requests per hour for a two-hour video. At higher arrival rates, it performs much worse than broadcasting protocols. To overcome this limitation, we propose a stream tapping protocol that preloads in the customer set-top box the first few minutes of all popular videos. To offset the cost of the additional buffer space, our protocol never requires the set-top box to receive data from the video server at more than twice video consumption rate. Our simulations indicate that preloading the first eight minutes of a two-hour video was enough to achieve lower bandwidth requirements than the best broadcasting protocols at any request arrival rate.","PeriodicalId":375127,"journal":{"name":"MASCOTS 2001, Proceedings Ninth International Symposium on Modeling, Analysis and Simulation of Computer and Telecommunication Systems","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2001-08-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129874503","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 : 2001-08-15DOI: 10.1109/MASCOT.2001.948865
Donghua Xu, G. Riley, M. Ammar, R. Fujimoto
We introduce and discuss a methodology for heterogeneous simulations of computer networks using the dynamic simulation backplane. This methodology allows for exchanging of protocol information between simulators across layers of the protocol stack. For example, the simulationist may wish to construct a simulation using the rich set of TCP models found in the ns network simulator, and at the same time using the highly detailed wireless MAC models found in the GloMoSim simulator. The backplane provides an interface between heterogeneous simulators which allows these simulators to exchange meaningful information across layers of the protocol stack, without detailed knowledge of internal representation in the foreign simulator. With this method of heterogeneous simulation, new and experimental protocols can be validated and tested in conjunction with existing and accepted simulations of lower protocol layers. We discuss the particular problems presented by the split protocol stack model, and present our solutions. We give results of our implementation of the split protocol backplane, using the ns simulator for the higher protocol stack layers, and the GloMoSim simulator for the lower layers.
{"title":"Split protocol stack network simulations using the dynamic simulation backplane","authors":"Donghua Xu, G. Riley, M. Ammar, R. Fujimoto","doi":"10.1109/MASCOT.2001.948865","DOIUrl":"https://doi.org/10.1109/MASCOT.2001.948865","url":null,"abstract":"We introduce and discuss a methodology for heterogeneous simulations of computer networks using the dynamic simulation backplane. This methodology allows for exchanging of protocol information between simulators across layers of the protocol stack. For example, the simulationist may wish to construct a simulation using the rich set of TCP models found in the ns network simulator, and at the same time using the highly detailed wireless MAC models found in the GloMoSim simulator. The backplane provides an interface between heterogeneous simulators which allows these simulators to exchange meaningful information across layers of the protocol stack, without detailed knowledge of internal representation in the foreign simulator. With this method of heterogeneous simulation, new and experimental protocols can be validated and tested in conjunction with existing and accepted simulations of lower protocol layers. We discuss the particular problems presented by the split protocol stack model, and present our solutions. We give results of our implementation of the split protocol backplane, using the ns simulator for the higher protocol stack layers, and the GloMoSim simulator for the lower layers.","PeriodicalId":375127,"journal":{"name":"MASCOTS 2001, Proceedings Ninth International Symposium on Modeling, Analysis and Simulation of Computer and Telecommunication Systems","volume":"64 6 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2001-08-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131508433","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 : 2001-08-15DOI: 10.1109/MASCOT.2001.948867
Hasyim Gautama, A. V. Gemund
Current analytic solutions to the execution time distribution of an N-ary parallel composition of tasks having independent and identically distributed execution times are computationally complex, except for a limited number of distributions. In this paper we introduce an analytical solution based on approximating the execution time distributions in terms of a limited number of statistical moments. This approach allows the parallel execution time to be approximated with O(1) solution complexity for a wide range of execution time distributions, while the approximation accuracy outperforms comparable techniques known to date. Experiments show that the error of the predicted mean value of the parallel execution time is even less than 4% for parallel loops comprising up to 10,000 tasks whose execution times are normally distributed. Measurements on real programs (NAS-EP benchmark, PSRS sorter, and WATOR simulator) confirm these results provided the task execution distributions are independent and unimodal.
{"title":"Low-cost performance prediction of data-dependent data parallel programs","authors":"Hasyim Gautama, A. V. Gemund","doi":"10.1109/MASCOT.2001.948867","DOIUrl":"https://doi.org/10.1109/MASCOT.2001.948867","url":null,"abstract":"Current analytic solutions to the execution time distribution of an N-ary parallel composition of tasks having independent and identically distributed execution times are computationally complex, except for a limited number of distributions. In this paper we introduce an analytical solution based on approximating the execution time distributions in terms of a limited number of statistical moments. This approach allows the parallel execution time to be approximated with O(1) solution complexity for a wide range of execution time distributions, while the approximation accuracy outperforms comparable techniques known to date. Experiments show that the error of the predicted mean value of the parallel execution time is even less than 4% for parallel loops comprising up to 10,000 tasks whose execution times are normally distributed. Measurements on real programs (NAS-EP benchmark, PSRS sorter, and WATOR simulator) confirm these results provided the task execution distributions are independent and unimodal.","PeriodicalId":375127,"journal":{"name":"MASCOTS 2001, Proceedings Ninth International Symposium on Modeling, Analysis and Simulation of Computer and Telecommunication Systems","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2001-08-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131211846","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 : 2001-08-15DOI: 10.1109/MASCOT.2001.948850
S. De, Sajal K. Das
To provide quality of service (QoS) guarantee for real-time traffic in conventional wireline networks, the initial source-to-destination route is maintained throughout a session unless there is any failure in the route. Because of time-varying network congestion, the initially selected route may not be optimal in terms of resource utilization. We propose a dynamic multipath routing (DMPR) scheme to improve resource utilization of a network carrying real-time traffic by re-routing on-going calls through shorter routes. Through an analytical model we characterize performance of the proposed DMPR scheme in circuit-switched networks. Results show that the more the network connectivity and the larger the network size, the lesser (better) is the network delay performance. Since the algorithm acts in parallel with the on-going calls, contrary to the popular belief it does not increase network latency while improving resource utilization.
{"title":"Dynamic multipath routing (DMPR): an approach to improve resource utilization in networks for real-time traffic","authors":"S. De, Sajal K. Das","doi":"10.1109/MASCOT.2001.948850","DOIUrl":"https://doi.org/10.1109/MASCOT.2001.948850","url":null,"abstract":"To provide quality of service (QoS) guarantee for real-time traffic in conventional wireline networks, the initial source-to-destination route is maintained throughout a session unless there is any failure in the route. Because of time-varying network congestion, the initially selected route may not be optimal in terms of resource utilization. We propose a dynamic multipath routing (DMPR) scheme to improve resource utilization of a network carrying real-time traffic by re-routing on-going calls through shorter routes. Through an analytical model we characterize performance of the proposed DMPR scheme in circuit-switched networks. Results show that the more the network connectivity and the larger the network size, the lesser (better) is the network delay performance. Since the algorithm acts in parallel with the on-going calls, contrary to the popular belief it does not increase network latency while improving resource utilization.","PeriodicalId":375127,"journal":{"name":"MASCOTS 2001, Proceedings Ninth International Symposium on Modeling, Analysis and Simulation of Computer and Telecommunication Systems","volume":"30 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2001-08-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"116918631","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 : 2001-08-15DOI: 10.1109/MASCOT.2001.948862
C. Kiddle, R. Simmonds, David K. Wilson, B. Unger
This paper presents a modelling language designed to satisfy the requirements for describing large and complex networks. The language is called ANML ("animal") which stands for ANother Modelling Language. ANML supports hierarchical modelling, reuse and validation. Schemas are used to define component types and constraints, databases act as repositories for reusable components and models define specific scenarios. We give an overview of ANML's properties and provide examples of its use for describing networks.
{"title":"ANML: A language for describing networks","authors":"C. Kiddle, R. Simmonds, David K. Wilson, B. Unger","doi":"10.1109/MASCOT.2001.948862","DOIUrl":"https://doi.org/10.1109/MASCOT.2001.948862","url":null,"abstract":"This paper presents a modelling language designed to satisfy the requirements for describing large and complex networks. The language is called ANML (\"animal\") which stands for ANother Modelling Language. ANML supports hierarchical modelling, reuse and validation. Schemas are used to define component types and constraints, databases act as repositories for reusable components and models define specific scenarios. We give an overview of ANML's properties and provide examples of its use for describing networks.","PeriodicalId":375127,"journal":{"name":"MASCOTS 2001, Proceedings Ninth International Symposium on Modeling, Analysis and Simulation of Computer and Telecommunication Systems","volume":"105 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2001-08-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"123502427","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 : 2001-08-15DOI: 10.1109/MASCOT.2001.948894
S. Yilmaz, I. Matta
The congestion control mechanisms of TCP make it vulnerable in an environment where flows with different congestion-sensitivity compete for scarce resources. With the increasing amount of unresponsive UDP traffic in today's Internet, new mechanisms are needed to enforce fairness in the core of the network. We propose a scalable Diffserv-like architecture, where flows with different characteristics are classified into separate service queues at the routers. Such class-based isolation provides protection so that flows with different characteristics do not negatively impact one another. We examine different aspects of UDP and TCP interaction and possible gains from segregating UDP and TCP into different classes. We also investigate the utility of further segregating TCP flows into two classes, which are class of short and class of long flows. Results are obtained analytically for both tail-drop and random early drop (RED) routers. Class-based isolation have the following salient features: (1) better fairness, (2) improved predictability for all kinds of flows, (3) lower transmission delay for delay-sensitive flows, and (4) better control over quality of service (QoS) of a particular traffic type.
{"title":"On class-based isolation of UDP, short-lived and long-lived TCP flows","authors":"S. Yilmaz, I. Matta","doi":"10.1109/MASCOT.2001.948894","DOIUrl":"https://doi.org/10.1109/MASCOT.2001.948894","url":null,"abstract":"The congestion control mechanisms of TCP make it vulnerable in an environment where flows with different congestion-sensitivity compete for scarce resources. With the increasing amount of unresponsive UDP traffic in today's Internet, new mechanisms are needed to enforce fairness in the core of the network. We propose a scalable Diffserv-like architecture, where flows with different characteristics are classified into separate service queues at the routers. Such class-based isolation provides protection so that flows with different characteristics do not negatively impact one another. We examine different aspects of UDP and TCP interaction and possible gains from segregating UDP and TCP into different classes. We also investigate the utility of further segregating TCP flows into two classes, which are class of short and class of long flows. Results are obtained analytically for both tail-drop and random early drop (RED) routers. Class-based isolation have the following salient features: (1) better fairness, (2) improved predictability for all kinds of flows, (3) lower transmission delay for delay-sensitive flows, and (4) better control over quality of service (QoS) of a particular traffic type.","PeriodicalId":375127,"journal":{"name":"MASCOTS 2001, Proceedings Ninth International Symposium on Modeling, Analysis and Simulation of Computer and Telecommunication Systems","volume":"8 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2001-08-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"130268369","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 : 2001-08-15DOI: 10.1109/MASCOT.2001.948880
A. Amer, D. Long
We introduce the aggregating cache, and demonstrate how it can be used to reduce the number of file retrieval requests made by a caching client, improving storage system performance by reducing the impact of latency. The aggregating cache utilizes predetermined groupings of files to perform group retrievals. These groups are maintained by the server, and built dynamically using observed inter-file relationships. Through a simple analytical model we demonstrate how this mechanism has the potential to reduce average latencies by 75% to 82%. Through trace-based simulation we demonstrate that a simple aggregating cache can reduce the number of demand fetches by almost 50%, while simultaneously improving cache hit ratios by up to 5%.
{"title":"Aggregating caches: A mechanism for implicit file prefetching","authors":"A. Amer, D. Long","doi":"10.1109/MASCOT.2001.948880","DOIUrl":"https://doi.org/10.1109/MASCOT.2001.948880","url":null,"abstract":"We introduce the aggregating cache, and demonstrate how it can be used to reduce the number of file retrieval requests made by a caching client, improving storage system performance by reducing the impact of latency. The aggregating cache utilizes predetermined groupings of files to perform group retrievals. These groups are maintained by the server, and built dynamically using observed inter-file relationships. Through a simple analytical model we demonstrate how this mechanism has the potential to reduce average latencies by 75% to 82%. Through trace-based simulation we demonstrate that a simple aggregating cache can reduce the number of demand fetches by almost 50%, while simultaneously improving cache hit ratios by up to 5%.","PeriodicalId":375127,"journal":{"name":"MASCOTS 2001, Proceedings Ninth International Symposium on Modeling, Analysis and Simulation of Computer and Telecommunication Systems","volume":"4 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2001-08-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"116284740","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：