Recent research indicates that multi-hop wireless networks can suffer from extreme imbalances in the throughput achieved by simultaneous competing flows. We address this problem by designing a practical distributed algorithm to compute a time-slot based schedule that provides end-to-end max-min fairness. Our system uses randomized priorities based on local weights to arbitrate access between nodes that directly compete with each other (we call this weighted slot allocation or WSA). The local weights are in turn computed by a higher layer called end-to-end fairness using local weights (EFLoW). EFLoW implements an additive-increase multiplicative-decrease (AIMD) algorithm that can automatically adapt to changes in traffic demands and network conditions. In each iteration, EFLoW only uses state obtained from within a given node's contention region. We have implemented WSA and EFLoW in both a simulator and a real system by using the overlay MAC layer (OML). Unlike previous work on end-to-end fairness, our approach does not use a centralized coordinator and works for traffic patterns with any number of sources and sinks. Also, since we compute both the fair allocation and a schedule to achieve it, we do not make any assumptions about the efficiency of carrier-sense (CS) based MACs - this is very important in the light of recent work which shows that current CS-based MACs can be very unfair even when all nodes are limited to sending at their fair rate. Our results show that WSA and EFLoW can prevent starvation of flows and improve fairness without sacrificing efficiency for a wide variety of traffic patterns.
{"title":"Adaptive Distributed Time-Slot Based Scheduling for Fairness in Multi-Hop Wireless Networks","authors":"A. Rao, I. Stoica","doi":"10.1109/ICDCS.2008.108","DOIUrl":"https://doi.org/10.1109/ICDCS.2008.108","url":null,"abstract":"Recent research indicates that multi-hop wireless networks can suffer from extreme imbalances in the throughput achieved by simultaneous competing flows. We address this problem by designing a practical distributed algorithm to compute a time-slot based schedule that provides end-to-end max-min fairness. Our system uses randomized priorities based on local weights to arbitrate access between nodes that directly compete with each other (we call this weighted slot allocation or WSA). The local weights are in turn computed by a higher layer called end-to-end fairness using local weights (EFLoW). EFLoW implements an additive-increase multiplicative-decrease (AIMD) algorithm that can automatically adapt to changes in traffic demands and network conditions. In each iteration, EFLoW only uses state obtained from within a given node's contention region. We have implemented WSA and EFLoW in both a simulator and a real system by using the overlay MAC layer (OML). Unlike previous work on end-to-end fairness, our approach does not use a centralized coordinator and works for traffic patterns with any number of sources and sinks. Also, since we compute both the fair allocation and a schedule to achieve it, we do not make any assumptions about the efficiency of carrier-sense (CS) based MACs - this is very important in the light of recent work which shows that current CS-based MACs can be very unfair even when all nodes are limited to sending at their fair rate. Our results show that WSA and EFLoW can prevent starvation of flows and improve fairness without sacrificing efficiency for a wide variety of traffic patterns.","PeriodicalId":240205,"journal":{"name":"2008 The 28th International Conference on Distributed Computing Systems","volume":"3 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2008-06-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"130852468","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}
XML-based data dissemination networks are rapidly gaining momentum. In these networks XML content is routed from data producers to data consumers throughout an overlay network of content-based routers. Routing decisions are based on XPath expressions (XPEs) stored at each router. To enable efficient routing, while keeping the routing state small, we introduce an advertisement-based routing algorithm for XML content, present a novel data structure for managing XPEs, especially apt for the hierarchical nature of XPEs and XML, and develop several optimizations for reducing the number of XPEs required to manage the routing state. The experimental evaluation shows that our algorithms and optimizations reduce the routing table size by up to 90%, improve the routing time by roughly 85%, and reduce overall network traffic by about 35%. Experiments running on PlanetLab show the scalability of our approach.
{"title":"Routing of XML and XPath Queries in Data Dissemination Networks","authors":"Guoli Li, Shuang Hou, H. Jacobsen","doi":"10.1109/ICDCS.2008.31","DOIUrl":"https://doi.org/10.1109/ICDCS.2008.31","url":null,"abstract":"XML-based data dissemination networks are rapidly gaining momentum. In these networks XML content is routed from data producers to data consumers throughout an overlay network of content-based routers. Routing decisions are based on XPath expressions (XPEs) stored at each router. To enable efficient routing, while keeping the routing state small, we introduce an advertisement-based routing algorithm for XML content, present a novel data structure for managing XPEs, especially apt for the hierarchical nature of XPEs and XML, and develop several optimizations for reducing the number of XPEs required to manage the routing state. The experimental evaluation shows that our algorithms and optimizations reduce the routing table size by up to 90%, improve the routing time by roughly 85%, and reduce overall network traffic by about 35%. Experiments running on PlanetLab show the scalability of our approach.","PeriodicalId":240205,"journal":{"name":"2008 The 28th International Conference on Distributed Computing Systems","volume":"15 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2008-06-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129526869","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
This paper presents a theoretical and experimental study on the limitations of copy-on-write snapshots and incremental backups in terms of data recoverability. We provide mathematical proofs of our new findings as well as implementation experiments to show how data recovery is done in case of various failures. Based on our study, we propose a new system architecture that will overcome the problems of existing technologies. The new architecture can provide two-way data recovery capability with the same storage overheads and can be implemented fairly easily on existing systems. We show that the new architecture has maximum data recoverability and is practically feasible.
{"title":"Can We Really Recover Data if Storage Subsystem Fails?","authors":"Weijun Xiao, Qing Yang","doi":"10.1109/ICDCS.2008.38","DOIUrl":"https://doi.org/10.1109/ICDCS.2008.38","url":null,"abstract":"This paper presents a theoretical and experimental study on the limitations of copy-on-write snapshots and incremental backups in terms of data recoverability. We provide mathematical proofs of our new findings as well as implementation experiments to show how data recovery is done in case of various failures. Based on our study, we propose a new system architecture that will overcome the problems of existing technologies. The new architecture can provide two-way data recovery capability with the same storage overheads and can be implemented fairly easily on existing systems. We show that the new architecture has maximum data recoverability and is practically feasible.","PeriodicalId":240205,"journal":{"name":"2008 The 28th International Conference on Distributed Computing Systems","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2008-06-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131265963","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}
A multi-object operation incurs communication or synchronization overhead when the requested objects are distributed over different nodes. The object pair correlations (the probability for a pair of objects to be requested together in an operation) are often highly skewed and yet stable over time for real-world distributed applications. Thus, placing strongly correlated objects on the same node (subject to node space constraint) tends to reduce communication overhead for multi-object operations. This paper studies the optimization of correlation-aware data placement. First, we formalize a restricted form of the problem as a variant of the classic Quadratic Assignment problem and we show that it is NP-hard. Based on a linear programming relaxation, we then propose a polynomial-time approximation algorithm that finds an object placement with communication overhead at most two times that of the optimal placement. We further show that the computation cost can be reduced by limiting the optimization scope to a relatively small number of most important objects. We quantitatively evaluate our approach on keyword index placement for full-text search engines using real traces of 3.7 million web pages and 6.8 million search queries. Compared to the correlation-oblivious random object placement, our approach achieves 37-86% communication overhead reduction on a range of optimization scopes and system sizes. The communication reduction is 30-78% compared to a correlation-aware greedy approach.
{"title":"Correlation-Aware Object Placement for Multi-Object Operations","authors":"Ming Zhong, Kai Shen, J. Seiferas","doi":"10.1109/ICDCS.2008.60","DOIUrl":"https://doi.org/10.1109/ICDCS.2008.60","url":null,"abstract":"A multi-object operation incurs communication or synchronization overhead when the requested objects are distributed over different nodes. The object pair correlations (the probability for a pair of objects to be requested together in an operation) are often highly skewed and yet stable over time for real-world distributed applications. Thus, placing strongly correlated objects on the same node (subject to node space constraint) tends to reduce communication overhead for multi-object operations. This paper studies the optimization of correlation-aware data placement. First, we formalize a restricted form of the problem as a variant of the classic Quadratic Assignment problem and we show that it is NP-hard. Based on a linear programming relaxation, we then propose a polynomial-time approximation algorithm that finds an object placement with communication overhead at most two times that of the optimal placement. We further show that the computation cost can be reduced by limiting the optimization scope to a relatively small number of most important objects. We quantitatively evaluate our approach on keyword index placement for full-text search engines using real traces of 3.7 million web pages and 6.8 million search queries. Compared to the correlation-oblivious random object placement, our approach achieves 37-86% communication overhead reduction on a range of optimization scopes and system sizes. The communication reduction is 30-78% compared to a correlation-aware greedy approach.","PeriodicalId":240205,"journal":{"name":"2008 The 28th International Conference on Distributed Computing Systems","volume":"31 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2008-06-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"114707018","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}
Advances in networking, sensors, medical devices and smart phones have made it feasible to monitor and provide medical and other assistance to people either in their homes or outside. Aging populations will benefit from reduced costs and improved healthcare through assisted living based on these technologies. However, these systems challenge current state-of-the-art techniques for usability, reliability, and security. In this paper we present the PAS open architecture for assisted living, which allows independently developed third party components to collaborate. Furthermore, we incorporate cell phones in PAS as the local intelligence in order to enhance the robustness and ubiquity. We discuss key technological issues in assisted living systems, such as software architecture layout, power preserving, security and privacy; and results from our pilot study in a real assisted living facility are presented.
{"title":"PAS: A Wireless-Enabled, Cell-Phone-Incorporated Personal Assistant System for Independent and Assisted Living","authors":"Zheng Zeng, Sammy Yu, Wook Shin, J. Hou","doi":"10.1109/ICDCS.2008.99","DOIUrl":"https://doi.org/10.1109/ICDCS.2008.99","url":null,"abstract":"Advances in networking, sensors, medical devices and smart phones have made it feasible to monitor and provide medical and other assistance to people either in their homes or outside. Aging populations will benefit from reduced costs and improved healthcare through assisted living based on these technologies. However, these systems challenge current state-of-the-art techniques for usability, reliability, and security. In this paper we present the PAS open architecture for assisted living, which allows independently developed third party components to collaborate. Furthermore, we incorporate cell phones in PAS as the local intelligence in order to enhance the robustness and ubiquity. We discuss key technological issues in assisted living systems, such as software architecture layout, power preserving, security and privacy; and results from our pilot study in a real assisted living facility are presented.","PeriodicalId":240205,"journal":{"name":"2008 The 28th International Conference on Distributed Computing Systems","volume":"35 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2008-06-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"117269687","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}
Internet coordinate systems have emerged as an efficient method to estimate the latency between pairs of nodes without any communication between them. However, most coordinate systems have been evaluated solely on data sets built by their authors from measurements gathered over large periods of time. Although they show good prediction results, it is unclear whether the accuracy is the result of the system design properties or is more connected to the characteristics of the data sets. In this paper, we revisit a simple question: how do the features of the embedding space and the inherent attributes of the data sets interact in producing good embeddings? We adapt the Vivaldi algorithm to use Hyperbolic space for embedding and evaluate both Euclidean and Hyperbolic Vivaldi on seven sets of real-world latencies. Our results show that node filtering and latency distributions can significantly influence the accuracy of the predictions. For example, although Euclidean Vivaldi performs well on data sets that were chosen, constructed and filtered by the designers of the algorithm, its performance and robustness decrease considerably when run on third party data sets that were not filtered a priori. Our results offer important insight into designing and building coordinate systems that are both robust and accurate in Internet-like environments.
{"title":"Measurement Manipulation and Space Selection in Network Coordinates","authors":"C. Lumezanu, N. Spring","doi":"10.1109/ICDCS.2008.27","DOIUrl":"https://doi.org/10.1109/ICDCS.2008.27","url":null,"abstract":"Internet coordinate systems have emerged as an efficient method to estimate the latency between pairs of nodes without any communication between them. However, most coordinate systems have been evaluated solely on data sets built by their authors from measurements gathered over large periods of time. Although they show good prediction results, it is unclear whether the accuracy is the result of the system design properties or is more connected to the characteristics of the data sets. In this paper, we revisit a simple question: how do the features of the embedding space and the inherent attributes of the data sets interact in producing good embeddings? We adapt the Vivaldi algorithm to use Hyperbolic space for embedding and evaluate both Euclidean and Hyperbolic Vivaldi on seven sets of real-world latencies. Our results show that node filtering and latency distributions can significantly influence the accuracy of the predictions. For example, although Euclidean Vivaldi performs well on data sets that were chosen, constructed and filtered by the designers of the algorithm, its performance and robustness decrease considerably when run on third party data sets that were not filtered a priori. Our results offer important insight into designing and building coordinate systems that are both robust and accurate in Internet-like environments.","PeriodicalId":240205,"journal":{"name":"2008 The 28th International Conference on Distributed Computing Systems","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2008-06-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"132597790","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}
Despite decades of research, there have not been developed concrete defense solutions for most of current attacks to Internet services, let alone new attack types. An essential problem to overcome is that malicious traffic can be similar to legitimate ones. Thus a more fundamental model which should be based on the overall performance of servers/subnets without inspecting each traffic must be remedied. Based on this observation, we propose a novel system framework, called detection of malicious users (DMU) which attempts to solve various attack types. Motivated by DMU, we introduce a new theoretical model, called size constraint group testing (SCGT). Several algorithms based on SCGT for various networking scenarios are proposed. We also provide several fundamental results on SCGT, revealing some necessary conditions to obtain an O(1) detection time algorithm.
{"title":"On Detection of Malicious Users Using Group Testing Techniques","authors":"M. Thai, Ying Xuan, Incheol Shin, T. Znati","doi":"10.1109/ICDCS.2008.75","DOIUrl":"https://doi.org/10.1109/ICDCS.2008.75","url":null,"abstract":"Despite decades of research, there have not been developed concrete defense solutions for most of current attacks to Internet services, let alone new attack types. An essential problem to overcome is that malicious traffic can be similar to legitimate ones. Thus a more fundamental model which should be based on the overall performance of servers/subnets without inspecting each traffic must be remedied. Based on this observation, we propose a novel system framework, called detection of malicious users (DMU) which attempts to solve various attack types. Motivated by DMU, we introduce a new theoretical model, called size constraint group testing (SCGT). Several algorithms based on SCGT for various networking scenarios are proposed. We also provide several fundamental results on SCGT, revealing some necessary conditions to obtain an O(1) detection time algorithm.","PeriodicalId":240205,"journal":{"name":"2008 The 28th International Conference on Distributed Computing Systems","volume":"13 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2008-06-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"128374354","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}
Xiaomeng Huang, Fengyuan Ren, Guangwen Yang, Yongwei Wu, W. Zhen, Chuang Lin
Since TCP congestion control is ill-suited for high speed networks, designing a replacement for TCP has become a challenge. To address this problem, we extend the population ecology theory to design a novel congestion control algorithm. We treat the network flows as the species in nature, the throughput of the flows as the population number, and the bottleneck bandwidth as the food resources. Then we use the key idea of constructing population ecology models to develop a novel congestion control model, and implement the corresponding end-to-end transport protocol through measurement, which called Population Ecology TCP (PE-TCP). The theoretical analysis and simulation results validate that PE-TCP achieves high utilization, fast convergence, fair bandwidth allocation, and near-zero packet drops. These qualities are desirable for high speed networks.
{"title":"End-to-End Congestion Control for High Speed Networks Based on Population Ecology Models","authors":"Xiaomeng Huang, Fengyuan Ren, Guangwen Yang, Yongwei Wu, W. Zhen, Chuang Lin","doi":"10.1109/ICDCS.2008.26","DOIUrl":"https://doi.org/10.1109/ICDCS.2008.26","url":null,"abstract":"Since TCP congestion control is ill-suited for high speed networks, designing a replacement for TCP has become a challenge. To address this problem, we extend the population ecology theory to design a novel congestion control algorithm. We treat the network flows as the species in nature, the throughput of the flows as the population number, and the bottleneck bandwidth as the food resources. Then we use the key idea of constructing population ecology models to develop a novel congestion control model, and implement the corresponding end-to-end transport protocol through measurement, which called Population Ecology TCP (PE-TCP). The theoretical analysis and simulation results validate that PE-TCP achieves high utilization, fast convergence, fair bandwidth allocation, and near-zero packet drops. These qualities are desirable for high speed networks.","PeriodicalId":240205,"journal":{"name":"2008 The 28th International Conference on Distributed Computing Systems","volume":"65 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2008-06-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129634483","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}
Ao-Jan Su, D. Choffnes, F. Bustamante, A. Kuzmanovic
Many large-scale distributed systems can benefit from a service that allows them to select among alternative nodes based on their relative network positions. A variety of approaches propose new measurement infrastructures that attempt to scale this service to large numbers of nodes by reducing the amount of direct measurements to end hosts. In this paper, we introduce a new approach to relative network positioning that eliminates direct probing by leveraging pre-existing infrastructure. Specifically, we exploit the dynamic association of nodes with replica servers from large content distribution networks (CDNs) to determine relative position information - we call this approach CDN-based relative network positioning (CRP). We demonstrate how CRP can support two common examples of location information used by distributed applications: server selection and dynamic node clustering. After describing CRP in detail, we present results from an extensive wide-area evaluation that demonstrates its effectiveness.
{"title":"Relative Network Positioning via CDN Redirections","authors":"Ao-Jan Su, D. Choffnes, F. Bustamante, A. Kuzmanovic","doi":"10.1109/ICDCS.2008.54","DOIUrl":"https://doi.org/10.1109/ICDCS.2008.54","url":null,"abstract":"Many large-scale distributed systems can benefit from a service that allows them to select among alternative nodes based on their relative network positions. A variety of approaches propose new measurement infrastructures that attempt to scale this service to large numbers of nodes by reducing the amount of direct measurements to end hosts. In this paper, we introduce a new approach to relative network positioning that eliminates direct probing by leveraging pre-existing infrastructure. Specifically, we exploit the dynamic association of nodes with replica servers from large content distribution networks (CDNs) to determine relative position information - we call this approach CDN-based relative network positioning (CRP). We demonstrate how CRP can support two common examples of location information used by distributed applications: server selection and dynamic node clustering. After describing CRP in detail, we present results from an extensive wide-area evaluation that demonstrates its effectiveness.","PeriodicalId":240205,"journal":{"name":"2008 The 28th International Conference on Distributed Computing Systems","volume":"30 2","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2008-06-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"114043370","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}
The multi-level storage architecture has been widely adopted in servers and data centers. However, while prefetching has been shown as a crucial technique to exploit the sequentiality in accesses common for such systems and hide the increasing relative cost of disk I/O, existing multi-level storage studies have focused mostly on cache replacement strategies. In this paper, we show that prefetching algorithms designed for single-level systems may have their limitations magnified when applied to multi-level systems. Overly conservative prefetching will not be able to effectively use the lower-level cache space, while overly aggressive prefetching will be compounded across levels and generate large amounts of wasted prefetch. We take an innovative approach to this problem: rather than designing a new, multi-level prefetching algorithm, we developed prefetching-coordinator (PFC), a hierarchy-aware optimization applicable to any existing prefetching algorithms. PFC does not require any application hints, a priori knowledge on the application access pattern or the native prefetching algorithm, or modification to the I/O interface. Instead, it monitors the upper-level access patterns as well as the lower-level cache status, and dynamically adjusts the aggressiveness of the lower-level prefetching activities. We evaluated PFC with extensive simulation study using a verified multi-level storage simulator, an accurate disk simulator, and access traces with different access patterns. Our results indicate that PFC dynamically controls lower-level prefetching in reaction to multiple system and workload parameters, improving the overall system performance in all 96 test cases. Working with four well-known existing prefetching algorithms adopted in real systems, PFC obtains an improvement of up to 35% to the average request response time, with an average improvement of 14.6% over all cases.
{"title":"PFC: Transparent Optimization of Existing Prefetching Strategies for Multi-Level Storage Systems","authors":"Zhe Zhang, Kyuhyung Lee, Xiaosong Ma, Yuanyuan Zhou","doi":"10.1109/ICDCS.2008.89","DOIUrl":"https://doi.org/10.1109/ICDCS.2008.89","url":null,"abstract":"The multi-level storage architecture has been widely adopted in servers and data centers. However, while prefetching has been shown as a crucial technique to exploit the sequentiality in accesses common for such systems and hide the increasing relative cost of disk I/O, existing multi-level storage studies have focused mostly on cache replacement strategies. In this paper, we show that prefetching algorithms designed for single-level systems may have their limitations magnified when applied to multi-level systems. Overly conservative prefetching will not be able to effectively use the lower-level cache space, while overly aggressive prefetching will be compounded across levels and generate large amounts of wasted prefetch. We take an innovative approach to this problem: rather than designing a new, multi-level prefetching algorithm, we developed prefetching-coordinator (PFC), a hierarchy-aware optimization applicable to any existing prefetching algorithms. PFC does not require any application hints, a priori knowledge on the application access pattern or the native prefetching algorithm, or modification to the I/O interface. Instead, it monitors the upper-level access patterns as well as the lower-level cache status, and dynamically adjusts the aggressiveness of the lower-level prefetching activities. We evaluated PFC with extensive simulation study using a verified multi-level storage simulator, an accurate disk simulator, and access traces with different access patterns. Our results indicate that PFC dynamically controls lower-level prefetching in reaction to multiple system and workload parameters, improving the overall system performance in all 96 test cases. Working with four well-known existing prefetching algorithms adopted in real systems, PFC obtains an improvement of up to 35% to the average request response time, with an average improvement of 14.6% over all cases.","PeriodicalId":240205,"journal":{"name":"2008 The 28th International Conference on Distributed Computing Systems","volume":"3 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2008-06-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"133224615","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：