K. Veeraraghavan, Dongyoon Lee, Benjamin Wester, Jessica Ouyang, Peter M. Chen, J. Flinn, S. Narayanasamy
Deterministic replay systems record and reproduce the execution of a hardware or software system. In contrast to replaying execution on uniprocessors, deterministic replay on multiprocessors is very challenging to implement efficiently because of the need to reproduce the order or values read by shared memory operations performed by multiple threads. In this paper, we present DoublePlay, a new way to efficiently guarantee replay on commodity multiprocessors. Our key insight is that one can use the simpler and faster mechanisms of single-processor record and replay, yet still achieve the scalability offered by multiple cores, by using an additional execution to parallelize the record and replay of an application. DoublePlay timeslices multiple threads on a single processor, then runs multiple time intervals (epochs) of the program concurrently on separate processors. This strategy, which we call uniparallelism, makes logging much easier because each epoch runs on a single processor (so threads in an epoch never simultaneously access the same memory) and different epochs operate on different copies of the memory. Thus, rather than logging the order of shared-memory accesses, we need only log the order in which threads in an epoch are timesliced on the processor. DoublePlay runs an additional execution of the program on multiple processors to generate checkpoints so that epochs run in parallel. We evaluate DoublePlay on a variety of client, server, and scientific parallel benchmarks; with spare cores, DoublePlay reduces logging overhead to an average of 15% with two worker threads and 28% with four threads.
{"title":"DoublePlay: parallelizing sequential logging and replay","authors":"K. Veeraraghavan, Dongyoon Lee, Benjamin Wester, Jessica Ouyang, Peter M. Chen, J. Flinn, S. Narayanasamy","doi":"10.1145/1950365.1950370","DOIUrl":"https://doi.org/10.1145/1950365.1950370","url":null,"abstract":"Deterministic replay systems record and reproduce the execution of a hardware or software system. In contrast to replaying execution on uniprocessors, deterministic replay on multiprocessors is very challenging to implement efficiently because of the need to reproduce the order or values read by shared memory operations performed by multiple threads. In this paper, we present DoublePlay, a new way to efficiently guarantee replay on commodity multiprocessors. Our key insight is that one can use the simpler and faster mechanisms of single-processor record and replay, yet still achieve the scalability offered by multiple cores, by using an additional execution to parallelize the record and replay of an application. DoublePlay timeslices multiple threads on a single processor, then runs multiple time intervals (epochs) of the program concurrently on separate processors. This strategy, which we call uniparallelism, makes logging much easier because each epoch runs on a single processor (so threads in an epoch never simultaneously access the same memory) and different epochs operate on different copies of the memory. Thus, rather than logging the order of shared-memory accesses, we need only log the order in which threads in an epoch are timesliced on the processor. DoublePlay runs an additional execution of the program on multiple processors to generate checkpoints so that epochs run in parallel. We evaluate DoublePlay on a variety of client, server, and scientific parallel benchmarks; with spare cores, DoublePlay reduces logging overhead to an average of 15% with two worker threads and 28% with four threads.","PeriodicalId":50918,"journal":{"name":"ACM Transactions on Computer Systems","volume":"1 1","pages":"15-26"},"PeriodicalIF":1.5,"publicationDate":"2011-03-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://sci-hub-pdf.com/10.1145/1950365.1950370","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"64122270","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
H. A. Lagar-Cavilla, J. Whitney, R. Bryant, Philip Patchin, M. Brudno, E. D. Lara, Stephen M. Rumble, M. Satyanarayanan, A. Scannell
A basic building block of cloud computing is virtualization. Virtual machines (VMs) encapsulate a user’s computing environment and efficiently isolate it from that of other users. VMs, however, are large entities, and no clear APIs exist yet to provide users with programatic, fine-grained control on short time scales. We present SnowFlock, a paradigm and system for cloud computing that introduces VM cloning as a first-class cloud abstraction. VM cloning exploits the well-understood and effective semantics of UNIX fork. We demonstrate multiple usage models of VM cloning: users can incorporate the primitive in their code, can wrap around existing toolchains via scripting, can encapsulate the API within a parallel programming framework, or can use it to load-balance and self-scale clustered servers. VM cloning needs to be efficient to be usable. It must efficiently transmit VM state in order to avoid cloud I/O bottlenecks. We demonstrate how the semantics of cloning aid us in realizing its efficiency: state is propagated in parallel to multiple VM clones, and is transmitted during runtime, allowing for optimizations that substantially reduce the I/O load. We show detailed microbenchmark results highlighting the efficiency of our optimizations, and macrobenchmark numbers demonstrating the effectiveness of the different usage models of SnowFlock.
{"title":"SnowFlock: Virtual Machine Cloning as a First-Class Cloud Primitive","authors":"H. A. Lagar-Cavilla, J. Whitney, R. Bryant, Philip Patchin, M. Brudno, E. D. Lara, Stephen M. Rumble, M. Satyanarayanan, A. Scannell","doi":"10.1145/1925109.1925111","DOIUrl":"https://doi.org/10.1145/1925109.1925111","url":null,"abstract":"A basic building block of cloud computing is virtualization. Virtual machines (VMs) encapsulate a user’s computing environment and efficiently isolate it from that of other users. VMs, however, are large entities, and no clear APIs exist yet to provide users with programatic, fine-grained control on short time scales.\u0000 We present SnowFlock, a paradigm and system for cloud computing that introduces VM cloning as a first-class cloud abstraction. VM cloning exploits the well-understood and effective semantics of UNIX fork. We demonstrate multiple usage models of VM cloning: users can incorporate the primitive in their code, can wrap around existing toolchains via scripting, can encapsulate the API within a parallel programming framework, or can use it to load-balance and self-scale clustered servers.\u0000 VM cloning needs to be efficient to be usable. It must efficiently transmit VM state in order to avoid cloud I/O bottlenecks. We demonstrate how the semantics of cloning aid us in realizing its efficiency: state is propagated in parallel to multiple VM clones, and is transmitted during runtime, allowing for optimizations that substantially reduce the I/O load. We show detailed microbenchmark results highlighting the efficiency of our optimizations, and macrobenchmark numbers demonstrating the effectiveness of the different usage models of SnowFlock.","PeriodicalId":50918,"journal":{"name":"ACM Transactions on Computer Systems","volume":"54 1","pages":"2:1-2:45"},"PeriodicalIF":1.5,"publicationDate":"2011-02-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"74044221","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Data center power consumption is growing to unprecedented levels: the EPA estimates U.S. data centers will consume 100 billion kilowatt hours annually by 2011. Much of this energy is wasted in idle systems: in typical deployments, server utilization is below 30%, but idle servers still consume 60% of their peak power draw. Typical idle periods---though frequent---last seconds or less, confounding simple energy-conservation approaches. In this article, we propose PowerNap, an energy-conservation approach where the entire system transitions rapidly between a high-performance active state and a near-zero-power idle state in response to instantaneous load. Rather than requiring fine-grained power-performance states and complex load-proportional operation from individual system components, PowerNap instead calls for minimizing idle power and transition time, which are simpler optimization goals. Based on the PowerNap concept, we develop requirements and outline mechanisms to eliminate idle power waste in enterprise blade servers. Because PowerNap operates in low-efficiency regions of current blade center power supplies, we introduce the Redundant Array for Inexpensive Load Sharing (RAILS), a power provisioning approach that provides high conversion efficiency across the entire range of PowerNap’s power demands. Using utilization traces collected from enterprise-scale commercial deployments, we demonstrate that, together, PowerNap and RAILS reduce average server power consumption by 74%.
{"title":"The PowerNap Server Architecture","authors":"David Meisner, Brian T. Gold, T. Wenisch","doi":"10.1145/1925109.1925112","DOIUrl":"https://doi.org/10.1145/1925109.1925112","url":null,"abstract":"Data center power consumption is growing to unprecedented levels: the EPA estimates U.S. data centers will consume 100 billion kilowatt hours annually by 2011. Much of this energy is wasted in idle systems: in typical deployments, server utilization is below 30%, but idle servers still consume 60% of their peak power draw. Typical idle periods---though frequent---last seconds or less, confounding simple energy-conservation approaches.\u0000 In this article, we propose PowerNap, an energy-conservation approach where the entire system transitions rapidly between a high-performance active state and a near-zero-power idle state in response to instantaneous load. Rather than requiring fine-grained power-performance states and complex load-proportional operation from individual system components, PowerNap instead calls for minimizing idle power and transition time, which are simpler optimization goals. Based on the PowerNap concept, we develop requirements and outline mechanisms to eliminate idle power waste in enterprise blade servers. Because PowerNap operates in low-efficiency regions of current blade center power supplies, we introduce the Redundant Array for Inexpensive Load Sharing (RAILS), a power provisioning approach that provides high conversion efficiency across the entire range of PowerNap’s power demands. Using utilization traces collected from enterprise-scale commercial deployments, we demonstrate that, together, PowerNap and RAILS reduce average server power consumption by 74%.","PeriodicalId":50918,"journal":{"name":"ACM Transactions on Computer Systems","volume":"3 1","pages":"3:1-3:24"},"PeriodicalIF":1.5,"publicationDate":"2011-02-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"80370291","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Distributed content-based publish/subscribe systems suffer from performance degradation and poor scalability caused by uneven load distributions typical in real-world applications. The reason for this shortcoming is the lack of a load balancing scheme. This article proposes a load balancing solution specifically tailored to the needs of content-based publish/subscribe systems that is distributed, dynamic, adaptive, transparent, and accommodates heterogeneity. The solution consists of three key contributions: a load balancing framework, a novel load estimation algorithm, and three offload strategies. A working prototype of our solution is built on an open-sourced content-based publish/subscribe system and evaluated on PlanetLab, a cluster testbed, and in simulations. Real-life experiment results show that the proposed load balancing solution is efficient with less than 0.2% overhead; effective in distributing and balancing load originating from a single server to all available servers in the network; and capable of preventing overloads to preserve system stability, availability, and quality of service.
{"title":"Load Balancing Content-Based Publish/Subscribe Systems","authors":"A. Cheung, H. Jacobsen","doi":"10.1145/1880018.1880020","DOIUrl":"https://doi.org/10.1145/1880018.1880020","url":null,"abstract":"Distributed content-based publish/subscribe systems suffer from performance degradation and poor scalability caused by uneven load distributions typical in real-world applications. The reason for this shortcoming is the lack of a load balancing scheme. This article proposes a load balancing solution specifically tailored to the needs of content-based publish/subscribe systems that is distributed, dynamic, adaptive, transparent, and accommodates heterogeneity. The solution consists of three key contributions: a load balancing framework, a novel load estimation algorithm, and three offload strategies. A working prototype of our solution is built on an open-sourced content-based publish/subscribe system and evaluated on PlanetLab, a cluster testbed, and in simulations. Real-life experiment results show that the proposed load balancing solution is efficient with less than 0.2% overhead; effective in distributing and balancing load originating from a single server to all available servers in the network; and capable of preventing overloads to preserve system stability, availability, and quality of service.","PeriodicalId":50918,"journal":{"name":"ACM Transactions on Computer Systems","volume":"82 1","pages":"9:1-9:55"},"PeriodicalIF":1.5,"publicationDate":"2010-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"82160633","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
S. Blagodurov, Sergey Zhuravlev, Alexandra Fedorova
Contention for shared resources on multicore processors remains an unsolved problem in existing systems despite significant research efforts dedicated to this problem in the past. Previous solutions focused primarily on hardware techniques and software page coloring to mitigate this problem. Our goal is to investigate how and to what extent contention for shared resource can be mitigated via thread scheduling. Scheduling is an attractive tool, because it does not require extra hardware and is relatively easy to integrate into the system. Our study is the first to provide a comprehensive analysis of contention-mitigating techniques that use only scheduling. The most difficult part of the problem is to find a classification scheme for threads, which would determine how they affect each other when competing for shared resources. We provide a comprehensive analysis of such classification schemes using a newly proposed methodology that enables to evaluate these schemes separately from the scheduling algorithm itself and to compare them to the optimal. As a result of this analysis we discovered a classification scheme that addresses not only contention for cache space, but contention for other shared resources, such as the memory controller, memory bus and prefetching hardware. To show the applicability of our analysis we design a new scheduling algorithm, which we prototype at user level, and demonstrate that it performs within 2% of the optimal. We also conclude that the highest impact of contention-aware scheduling techniques is not in improving performance of a workload as a whole but in improving quality of service or performance isolation for individual applications and in optimizing system energy consumption.
{"title":"Contention-Aware Scheduling on Multicore Systems","authors":"S. Blagodurov, Sergey Zhuravlev, Alexandra Fedorova","doi":"10.1145/1880018.1880019","DOIUrl":"https://doi.org/10.1145/1880018.1880019","url":null,"abstract":"Contention for shared resources on multicore processors remains an unsolved problem in existing systems despite significant research efforts dedicated to this problem in the past. Previous solutions focused primarily on hardware techniques and software page coloring to mitigate this problem. Our goal is to investigate how and to what extent contention for shared resource can be mitigated via thread scheduling. Scheduling is an attractive tool, because it does not require extra hardware and is relatively easy to integrate into the system. Our study is the first to provide a comprehensive analysis of contention-mitigating techniques that use only scheduling. The most difficult part of the problem is to find a classification scheme for threads, which would determine how they affect each other when competing for shared resources. We provide a comprehensive analysis of such classification schemes using a newly proposed methodology that enables to evaluate these schemes separately from the scheduling algorithm itself and to compare them to the optimal. As a result of this analysis we discovered a classification scheme that addresses not only contention for cache space, but contention for other shared resources, such as the memory controller, memory bus and prefetching hardware. To show the applicability of our analysis we design a new scheduling algorithm, which we prototype at user level, and demonstrate that it performs within 2% of the optimal. We also conclude that the highest impact of contention-aware scheduling techniques is not in improving performance of a workload as a whole but in improving quality of service or performance isolation for individual applications and in optimizing system energy consumption.","PeriodicalId":50918,"journal":{"name":"ACM Transactions on Computer Systems","volume":"28 2 1","pages":"8:1-8:45"},"PeriodicalIF":1.5,"publicationDate":"2010-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"77901349","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Prince Mahajan, Srinath T. V. Setty, Sangmin Lee, Allen Clement, L. Alvisi, M. Dahlin, Michael Walfish
This article describes the design, implementation, and evaluation of Depot, a cloud storage system that minimizes trust assumptions. Depot tolerates buggy or malicious behavior by any number of clients or servers, yet it provides safety and liveness guarantees to correct clients. Depot provides these guarantees using a two-layer architecture. First, Depot ensures that the updates observed by correct nodes are consistently ordered under Fork-Join-Causal consistency (FJC). FJC is a slight weakening of causal consistency that can be both safe and live despite faulty nodes. Second, Depot implements protocols that use this consistent ordering of updates to provide other desirable consistency, staleness, durability, and recovery properties. Our evaluation suggests that the costs of these guarantees are modest and that Depot can tolerate faults and maintain good availability, latency, overhead, and staleness even when significant faults occur.
{"title":"Depot: Cloud Storage with Minimal Trust","authors":"Prince Mahajan, Srinath T. V. Setty, Sangmin Lee, Allen Clement, L. Alvisi, M. Dahlin, Michael Walfish","doi":"10.1145/2063509.2063512","DOIUrl":"https://doi.org/10.1145/2063509.2063512","url":null,"abstract":"This article describes the design, implementation, and evaluation of Depot, a cloud storage system that minimizes trust assumptions. Depot tolerates buggy or malicious behavior by any number of clients or servers, yet it provides safety and liveness guarantees to correct clients. Depot provides these guarantees using a two-layer architecture. First, Depot ensures that the updates observed by correct nodes are consistently ordered under Fork-Join-Causal consistency (FJC). FJC is a slight weakening of causal consistency that can be both safe and live despite faulty nodes. Second, Depot implements protocols that use this consistent ordering of updates to provide other desirable consistency, staleness, durability, and recovery properties. Our evaluation suggests that the costs of these guarantees are modest and that Depot can tolerate faults and maintain good availability, latency, overhead, and staleness even when significant faults occur.","PeriodicalId":50918,"journal":{"name":"ACM Transactions on Computer Systems","volume":"46 1","pages":"12:1-12:38"},"PeriodicalIF":1.5,"publicationDate":"2010-10-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"87051399","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Y. Amir, C. Danilov, R. Musaloiu-Elefteri, N. Rivera
Wireless mesh networks extend the connectivity range of mobile devices by using multiple access points, some of them connected to the Internet, to create a mesh topology and forward packets over multiple wireless hops. However, the quality of service provided by the mesh is impaired by the delays and disconnections caused by handoffs, as clients move within the area covered by multiple access points. We present the architecture and protocols of SMesh, the first transparent wireless mesh system that offers seamless, fast handoff, supporting real-time applications such as interactive VoIP. The handoff and routing logic is done solely by the access points, and therefore connectivity is attainable by any 802.11 device. In SMesh, the entire mesh network is seen by the mobile clients as a single, omnipresent access point, giving the mobile clients the illusion that they are stationary. We use multicast for access points coordination and, during handoff transitions, we use more than one access point to handle the moving client. SMesh provides a hybrid routing protocol that optimizes routes over wireless and wired links in a multihomed environment. Experimental results on a fully deployed mesh network demonstrate the effectiveness of the SMesh architecture and its intra-domain and inter-domain handoff protocols.
{"title":"The SMesh wireless mesh network","authors":"Y. Amir, C. Danilov, R. Musaloiu-Elefteri, N. Rivera","doi":"10.1145/1841313.1841314","DOIUrl":"https://doi.org/10.1145/1841313.1841314","url":null,"abstract":"Wireless mesh networks extend the connectivity range of mobile devices by using multiple access points, some of them connected to the Internet, to create a mesh topology and forward packets over multiple wireless hops. However, the quality of service provided by the mesh is impaired by the delays and disconnections caused by handoffs, as clients move within the area covered by multiple access points. We present the architecture and protocols of SMesh, the first transparent wireless mesh system that offers seamless, fast handoff, supporting real-time applications such as interactive VoIP. The handoff and routing logic is done solely by the access points, and therefore connectivity is attainable by any 802.11 device. In SMesh, the entire mesh network is seen by the mobile clients as a single, omnipresent access point, giving the mobile clients the illusion that they are stationary. We use multicast for access points coordination and, during handoff transitions, we use more than one access point to handle the moving client. SMesh provides a hybrid routing protocol that optimizes routes over wireless and wired links in a multihomed environment. Experimental results on a fully deployed mesh network demonstrate the effectiveness of the SMesh architecture and its intra-domain and inter-domain handoff protocols.","PeriodicalId":50918,"journal":{"name":"ACM Transactions on Computer Systems","volume":"43 1","pages":"6:1-6:49"},"PeriodicalIF":1.5,"publicationDate":"2010-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"86292843","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Proactive obfuscation is a new method for creating server replicas that are likely to have fewer shared vulnerabilities. It uses semantics-preserving code transformations to generate diverse executables, periodically restarting servers with these fresh versions. The periodic restarts help bound the number of compromised replicas that a service ever concurrently runs, and therefore proactive obfuscation makes an adversary's job harder. Proactive obfuscation was used in implementing two prototypes: a distributed firewall based on state-machine replication and a distributed storage service based on quorum systems. Costs intrinsic to supporting proactive obfuscation in replicated systems were evaluated by measuring the performance of these prototypes. The results show that employing proactive obfuscation adds little to the cost of replica-management protocols.
{"title":"Proactive obfuscation","authors":"T. Roeder, F. Schneider","doi":"10.1145/1813654.1813655","DOIUrl":"https://doi.org/10.1145/1813654.1813655","url":null,"abstract":"Proactive obfuscation is a new method for creating server replicas that are likely to have fewer shared vulnerabilities. It uses semantics-preserving code transformations to generate diverse executables, periodically restarting servers with these fresh versions. The periodic restarts help bound the number of compromised replicas that a service ever concurrently runs, and therefore proactive obfuscation makes an adversary's job harder. Proactive obfuscation was used in implementing two prototypes: a distributed firewall based on state-machine replication and a distributed storage service based on quorum systems. Costs intrinsic to supporting proactive obfuscation in replicated systems were evaluated by measuring the performance of these prototypes. The results show that employing proactive obfuscation adds little to the cost of replica-management protocols.","PeriodicalId":50918,"journal":{"name":"ACM Transactions on Computer Systems","volume":"29 1","pages":"4:1-4:54"},"PeriodicalIF":1.5,"publicationDate":"2010-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"74897114","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
R. Guerraoui, Ron R. Levy, Bastian Pochon, Vivien Quéma
Total order broadcast is a fundamental communication primitive that plays a central role in bringing cheap software-based high availability to a wide range of services. This article studies the practical performance of such a primitive on a cluster of homogeneous machines. We present LCR, the first throughput optimal uniform total order broadcast protocol. LCR is based on a ring topology. It only relies on point-to-point inter-process communication and has a linear latency with respect to the number of processes. LCR is also fair in the sense that each process has an equal opportunity of having its messages delivered by all processes. We benchmark a C implementation of LCR against Spread and JGroups, two of the most widely used group communication packages. LCR provides higher throughput than the alternatives, over a large number of scenarios.
{"title":"Throughput optimal total order broadcast for cluster environments","authors":"R. Guerraoui, Ron R. Levy, Bastian Pochon, Vivien Quéma","doi":"10.1145/1813654.1813656","DOIUrl":"https://doi.org/10.1145/1813654.1813656","url":null,"abstract":"Total order broadcast is a fundamental communication primitive that plays a central role in bringing cheap software-based high availability to a wide range of services. This article studies the practical performance of such a primitive on a cluster of homogeneous machines.\u0000 We present LCR, the first throughput optimal uniform total order broadcast protocol. LCR is based on a ring topology. It only relies on point-to-point inter-process communication and has a linear latency with respect to the number of processes. LCR is also fair in the sense that each process has an equal opportunity of having its messages delivered by all processes.\u0000 We benchmark a C implementation of LCR against Spread and JGroups, two of the most widely used group communication packages. LCR provides higher throughput than the alternatives, over a large number of scenarios.","PeriodicalId":50918,"journal":{"name":"ACM Transactions on Computer Systems","volume":"1 1","pages":"5:1-5:32"},"PeriodicalIF":1.5,"publicationDate":"2010-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"87427660","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
M. Yabandeh, Nikola Knezevic, Dejan Kostic, Viktor Kuncak
We propose a new approach for developing and deploying distributed systems, in which nodes predict distributed consequences of their actions and use this information to detect and avoid errors. Each node continuously runs a state exploration algorithm on a recent consistent snapshot of its neighborhood and predicts possible future violations of specified safety properties. We describe a new state exploration algorithm, consequence prediction, which explores causally related chains of events that lead to property violation. This article describes the design and implementation of this approach, termed CrystalBall. We evaluate CrystalBall on RandTree, BulletPrime, Paxos, and Chord distributed system implementations. We identified new bugs in mature Mace implementations of three systems. Furthermore, we show that if the bug is not corrected during system development, CrystalBall is effective in steering the execution away from inconsistent states at runtime.
{"title":"Predicting and preventing inconsistencies in deployed distributed systems","authors":"M. Yabandeh, Nikola Knezevic, Dejan Kostic, Viktor Kuncak","doi":"10.1145/1731060.1731062","DOIUrl":"https://doi.org/10.1145/1731060.1731062","url":null,"abstract":"We propose a new approach for developing and deploying distributed systems, in which nodes predict distributed consequences of their actions and use this information to detect and avoid errors. Each node continuously runs a state exploration algorithm on a recent consistent snapshot of its neighborhood and predicts possible future violations of specified safety properties. We describe a new state exploration algorithm, consequence prediction, which explores causally related chains of events that lead to property violation.\u0000 This article describes the design and implementation of this approach, termed CrystalBall. We evaluate CrystalBall on RandTree, BulletPrime, Paxos, and Chord distributed system implementations. We identified new bugs in mature Mace implementations of three systems. Furthermore, we show that if the bug is not corrected during system development, CrystalBall is effective in steering the execution away from inconsistent states at runtime.","PeriodicalId":50918,"journal":{"name":"ACM Transactions on Computer Systems","volume":"28 1","pages":"2:1-2:49"},"PeriodicalIF":1.5,"publicationDate":"2010-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"90038447","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"计算机科学","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：