A population protocol describes a set of state change rules for a population of n indistinguishable finite-state agents (automata), undergoing random pairwise interactions. Within this very basic framework, it is possible to resolve a number of fundamental tasks in distributed computing, including: leader election, aggregate and threshold functions on the population, such as majority computation, and plurality consensus. For the first time, we show that solutions to all of these problems can be obtained quickly using finite-state protocols. For any input, the designed finite-state protocols converge under a fair random scheduler to an output which is correct with high probability in expected O(polylog n) parallel time. We also show protocols which always reach a valid solution, in expected parallel time O(n^ε), where the number of states depends only on the choice of ε>0. The stated time bounds hold for any semi-linear predicate computable in the population protocol framework. The key ingredient of our result is the decentralized design of a hierarchy of phase-clocks, which tick at different rates, with the rates of adjacent clocks separated by a factor of Θ(log n). The construction of this clock hierarchy relies on a new protocol composition technique, combined with an adapted analysis of a self-organizing process of oscillatory dynamics. This clock hierarchy is used to provide nested synchronization primitives, which allow us to view the population in a global manner and design protocols using a high-level imperative programming language with a (limited) capacity for loops and branching instructions.
{"title":"Brief Announcement: Population Protocols Are Fast","authors":"A. Kosowski, P. Uznański","doi":"10.1145/3212734.3212788","DOIUrl":"https://doi.org/10.1145/3212734.3212788","url":null,"abstract":"A population protocol describes a set of state change rules for a population of n indistinguishable finite-state agents (automata), undergoing random pairwise interactions. Within this very basic framework, it is possible to resolve a number of fundamental tasks in distributed computing, including: leader election, aggregate and threshold functions on the population, such as majority computation, and plurality consensus. For the first time, we show that solutions to all of these problems can be obtained quickly using finite-state protocols. For any input, the designed finite-state protocols converge under a fair random scheduler to an output which is correct with high probability in expected O(polylog n) parallel time. We also show protocols which always reach a valid solution, in expected parallel time O(n^ε), where the number of states depends only on the choice of ε>0. The stated time bounds hold for any semi-linear predicate computable in the population protocol framework. The key ingredient of our result is the decentralized design of a hierarchy of phase-clocks, which tick at different rates, with the rates of adjacent clocks separated by a factor of Θ(log n). The construction of this clock hierarchy relies on a new protocol composition technique, combined with an adapted analysis of a self-organizing process of oscillatory dynamics. This clock hierarchy is used to provide nested synchronization primitives, which allow us to view the population in a global manner and design protocols using a high-level imperative programming language with a (limited) capacity for loops and branching instructions.","PeriodicalId":198284,"journal":{"name":"Proceedings of the 2018 ACM Symposium on Principles of Distributed Computing","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2018-07-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"130087662","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}
Restricting the bandwidth in models of distributed graph computations naturally introduces challenges that arise due to communication bottlenecks. In this talk, I will survey techniques for proving lower bounds on the complexity of fundamental graph problems under limited bandwidth. For some problems, allowing relaxed solutions can significantly reduce the required amount of communication, enabling efficient computations. Two successful approaches for overcoming provable barriers, namely, approximations and testing, will be exemplified and discussed in the context of distributed computing.
{"title":"Barriers due to Congestion and Two Ways to Deal With Them","authors":"K. Censor-Hillel","doi":"10.1145/3212734.3212797","DOIUrl":"https://doi.org/10.1145/3212734.3212797","url":null,"abstract":"Restricting the bandwidth in models of distributed graph computations naturally introduces challenges that arise due to communication bottlenecks. In this talk, I will survey techniques for proving lower bounds on the complexity of fundamental graph problems under limited bandwidth. For some problems, allowing relaxed solutions can significantly reduce the required amount of communication, enabling efficient computations. Two successful approaches for overcoming provable barriers, namely, approximations and testing, will be exemplified and discussed in the context of distributed computing.","PeriodicalId":198284,"journal":{"name":"Proceedings of the 2018 ACM Symposium on Principles of Distributed Computing","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2018-07-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129473579","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}
We bound the performance guarantees that follow from Turán-like bounds for unweighted and weighted independent sets in bounded-degree graphs. In particular, a randomized approach of Boppana forms a simple 1-round distributed algorithm, as well as a streaming and preemptive online algorithm. We show it gives a tight (Δ+1)/2-approximation in unweighted graphs of maximum degree Δ, which is best possible for 1-round distributed algorithms. For weighted graphs, it gives only a (Δ+1)-approximation, but a simple modification results in an asymptotic expected 0.529(Δ+1)-approximation.
{"title":"Brief Announcement: Simple and Local Independent Set Approximation","authors":"R. Boppana, M. Halldórsson, Dror Rawitz","doi":"10.1145/3212734.3212793","DOIUrl":"https://doi.org/10.1145/3212734.3212793","url":null,"abstract":"We bound the performance guarantees that follow from Turán-like bounds for unweighted and weighted independent sets in bounded-degree graphs. In particular, a randomized approach of Boppana forms a simple 1-round distributed algorithm, as well as a streaming and preemptive online algorithm. We show it gives a tight (Δ+1)/2-approximation in unweighted graphs of maximum degree Δ, which is best possible for 1-round distributed algorithms. For weighted graphs, it gives only a (Δ+1)-approximation, but a simple modification results in an asymptotic expected 0.529(Δ+1)-approximation.","PeriodicalId":198284,"journal":{"name":"Proceedings of the 2018 ACM Symposium on Principles of Distributed Computing","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2018-07-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"120876515","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}
M. Pavlovic, Alex Kogan, Virendra J. Marathe, T. Harris
This brief announcement presents a fundamental concurrent primitive for persistent memory - a persistent atomic multi-word compare-and-swap (PMCAS).We present a novel algorithm carefully crafted to ensure that atomic updates to a multitude of words modified by the PMCAS are persisted correctly. Our algorithm leverages hardware transactional memory (HTM) for concurrency control, and has a total of 3 persist barriers in its critical path. We also overview variants based on just the compare-and-swap (CAS) instruction and a hybrid of CAS and HTM.
{"title":"Brief Announcement: Persistent Multi-Word Compare-and-Swap","authors":"M. Pavlovic, Alex Kogan, Virendra J. Marathe, T. Harris","doi":"10.1145/3212734.3212783","DOIUrl":"https://doi.org/10.1145/3212734.3212783","url":null,"abstract":"This brief announcement presents a fundamental concurrent primitive for persistent memory - a persistent atomic multi-word compare-and-swap (PMCAS).We present a novel algorithm carefully crafted to ensure that atomic updates to a multitude of words modified by the PMCAS are persisted correctly. Our algorithm leverages hardware transactional memory (HTM) for concurrency control, and has a total of 3 persist barriers in its critical path. We also overview variants based on just the compare-and-swap (CAS) instruction and a hybrid of CAS and HTM.","PeriodicalId":198284,"journal":{"name":"Proceedings of the 2018 ACM Symposium on Principles of Distributed Computing","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2018-07-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"122733573","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}
Interactive proof systems allow a resource-bounded verifier to decide an intractable language (or compute a hard function) by communicating with a powerful but untrusted prover. Such systems guarantee that the prover can only convince the verifier of true statements. In the context of centralized computation, a celebrated result shows that interactive proofs are extremely powerful, allowing polynomial-time verifiers to decide any language in PSPACE. In this work we initiate the study of interactive distributed proofs : a network of nodes interacts with a single untrusted prover, who sees the entire network graph, to decide whether the graph satisfies some property. We focus on the communication cost of the protocol --- the number of bits the nodes must exchange with the prover and each other. Our model can also be viewed as a generalization of the various models of "distributed NP'' (proof labeling schemes, etc.) which received significant attention recently: while these models only allow the prover to present each network node with a string of advice, our model allows for back-and-forth interaction. We prove both upper and lower bounds for the new model. We show that for some problems, interaction can exponentially decrease the communication cost compared to a non-interactive prover, but on the other hand, some problems retain non-trivial cost even with interaction.
{"title":"Interactive Distributed Proofs","authors":"Gillat Kol, R. Oshman, Raghuvansh R. Saxena","doi":"10.1145/3212734.3212771","DOIUrl":"https://doi.org/10.1145/3212734.3212771","url":null,"abstract":"Interactive proof systems allow a resource-bounded verifier to decide an intractable language (or compute a hard function) by communicating with a powerful but untrusted prover. Such systems guarantee that the prover can only convince the verifier of true statements. In the context of centralized computation, a celebrated result shows that interactive proofs are extremely powerful, allowing polynomial-time verifiers to decide any language in PSPACE. In this work we initiate the study of interactive distributed proofs : a network of nodes interacts with a single untrusted prover, who sees the entire network graph, to decide whether the graph satisfies some property. We focus on the communication cost of the protocol --- the number of bits the nodes must exchange with the prover and each other. Our model can also be viewed as a generalization of the various models of \"distributed NP'' (proof labeling schemes, etc.) which received significant attention recently: while these models only allow the prover to present each network node with a string of advice, our model allows for back-and-forth interaction. We prove both upper and lower bounds for the new model. We show that for some problems, interaction can exponentially decrease the communication cost compared to a non-interactive prover, but on the other hand, some problems retain non-trivial cost even with interaction.","PeriodicalId":198284,"journal":{"name":"Proceedings of the 2018 ACM Symposium on Principles of Distributed Computing","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2018-07-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"122798639","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}
Cryptocurrency and blockchain technologies are recently gaining wide adoption since the introduction of Bitcoin, being distributed, authority-free, and secure. Proof of Work (PoW) is at the heart of blockchain's security, asset generation, and maintenance. Although simple and secure, a hash-based PoW like Bitcoin's puzzle is often referred to as "useless'', and the used intensive computations are considered "waste'' of energy. A myriad of Proof of "something'' alternatives have been proposed to mitigate energy consumption; however, they either introduced new security threats and limitations, or the "work'' remained far from being really "useful''. In this work, we introduce Proof of eXercise (PoX): a sustainable alternative to PoW where an eXercise is a real world matrix-based scientific computation problem. We provide a novel study of the properties of Bitcoin's PoW, the challenges of a more "rational'' solution as PoX, and we suggest a comprehensive approach for PoX.
{"title":"Brief Announcement: Sustainable Blockchains through Proof of eXercise","authors":"Ali Shoker","doi":"10.1145/3212734.3212781","DOIUrl":"https://doi.org/10.1145/3212734.3212781","url":null,"abstract":"Cryptocurrency and blockchain technologies are recently gaining wide adoption since the introduction of Bitcoin, being distributed, authority-free, and secure. Proof of Work (PoW) is at the heart of blockchain's security, asset generation, and maintenance. Although simple and secure, a hash-based PoW like Bitcoin's puzzle is often referred to as \"useless'', and the used intensive computations are considered \"waste'' of energy. A myriad of Proof of \"something'' alternatives have been proposed to mitigate energy consumption; however, they either introduced new security threats and limitations, or the \"work'' remained far from being really \"useful''. In this work, we introduce Proof of eXercise (PoX): a sustainable alternative to PoW where an eXercise is a real world matrix-based scientific computation problem. We provide a novel study of the properties of Bitcoin's PoW, the challenges of a more \"rational'' solution as PoX, and we suggest a comprehensive approach for PoX.","PeriodicalId":198284,"journal":{"name":"Proceedings of the 2018 ACM Symposium on Principles of Distributed Computing","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2018-07-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127512437","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}
Distributed em routing is one of the most central and fundamental problems in the area of Distributed Graph Algorithms. It was extensively studied for almost thirty years. Nevertheless, the currently existing solutions for this problem require either prohibitively large construction (aka preprocessing) time, or prohibitively large memory usage either during the construction or during the routing phase, and suffer from suboptimal labels and tables' sizes. We devise a distributed routing scheme that enjoys the best of all worlds. Specifically, its construction time and memory requirements during the construction phase are near-optimal, and so is also the tradeoff between the sizes of routing tables and labels on the one hand, and the stretch on the other. On the way to this result, we also improve upon existing solutions for the distributed exact em tree routing problem. Previous solutions require Ω(√ ) memory, and provide tables and labels of size O(log n) and O(log^2 n), respectively. Our solution, on the other hand, requires just O(log n) memory, and has tables of size O(1), and labels of size O(log n). These bounds match the bounds of the best-known centralized solution.
{"title":"Near-Optimal Distributed Routing with Low Memory","authors":"Michael Elkin, Ofer Neiman","doi":"10.1145/3212734.3212761","DOIUrl":"https://doi.org/10.1145/3212734.3212761","url":null,"abstract":"Distributed em routing is one of the most central and fundamental problems in the area of Distributed Graph Algorithms. It was extensively studied for almost thirty years. Nevertheless, the currently existing solutions for this problem require either prohibitively large construction (aka preprocessing) time, or prohibitively large memory usage either during the construction or during the routing phase, and suffer from suboptimal labels and tables' sizes. We devise a distributed routing scheme that enjoys the best of all worlds. Specifically, its construction time and memory requirements during the construction phase are near-optimal, and so is also the tradeoff between the sizes of routing tables and labels on the one hand, and the stretch on the other. On the way to this result, we also improve upon existing solutions for the distributed exact em tree routing problem. Previous solutions require Ω(√ ) memory, and provide tables and labels of size O(log n) and O(log^2 n), respectively. Our solution, on the other hand, requires just O(log n) memory, and has tables of size O(1), and labels of size O(log n). These bounds match the bounds of the best-known centralized solution.","PeriodicalId":198284,"journal":{"name":"Proceedings of the 2018 ACM Symposium on Principles of Distributed Computing","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2018-07-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"124044105","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 winner of the 2018 Principles of Distributed Computing Doctoral Dissertation Award is Dr. Rati Gelashvili, for his dissertation titled "On the Complexity of Synchronization," written under the supervision of Prof. Nir Shavit at the Massachusetts Institute of Technology. The field of distributed algorithms revolves around efficiently solving synchronization tasks, such as leader election and consensus in different models. Gelashvili's thesis provides an extraordinary study of the complexity of solving synchronization tasks, which is both deep and broad. It makes significant contributions towards understanding the complexity of solving synchronization tasks in various models. In particular, it pushes the boundary of our understanding of consensus, the algorithmic process by which asynchronous computation threads coordinate with each other, which has been the subject of extensive research for over 30 years. In one part of his thesis, Gelashvili challenges the underpinnings of Herlihy's consensus-based computability hierarchy, which has been the theoretical basis for classifying the computational power of concurrent data structures and synchronization primitives in multiprocessors and multicore machines for two and a half decades. He observes that Herlihy's classical hierarchy treats synchronization instructions as distinct objects, an approach that is far from the real-world, where multiprocessors do let processes apply supported atomic instructions to arbitrary memory locations. Gelashvili shows that, contrary to common belief, solving consensus does not require multicore architectures to support "strong" synchronization instructions such as compare-and-swap. Rather, combinations of "weaker" instructions such as decrement and multiply suffice. He goes on to propose an alternative complexity-based hierarchy for concurrent objects. The dissertation further opens a new line of research by proving a linear-space bound for the anonymous case of randomized consensus, the first major progress on this problem in 15 years, which won the Best Paper Award at DISC 2015, and for which Gelashvili developed novel lower bound techniques. Apart from their great importance, these results are also technically complex and mathematically beautiful.
{"title":"2018 Doctoral Dissertation Award","authors":"L. Alvisi, I. Keidar, A. Richa, A. Schwarzmann","doi":"10.1145/3212734.3232541","DOIUrl":"https://doi.org/10.1145/3212734.3232541","url":null,"abstract":"The winner of the 2018 Principles of Distributed Computing Doctoral Dissertation Award is Dr. Rati Gelashvili, for his dissertation titled \"On the Complexity of Synchronization,\" written under the supervision of Prof. Nir Shavit at the Massachusetts Institute of Technology. The field of distributed algorithms revolves around efficiently solving synchronization tasks, such as leader election and consensus in different models. Gelashvili's thesis provides an extraordinary study of the complexity of solving synchronization tasks, which is both deep and broad. It makes significant contributions towards understanding the complexity of solving synchronization tasks in various models. In particular, it pushes the boundary of our understanding of consensus, the algorithmic process by which asynchronous computation threads coordinate with each other, which has been the subject of extensive research for over 30 years. In one part of his thesis, Gelashvili challenges the underpinnings of Herlihy's consensus-based computability hierarchy, which has been the theoretical basis for classifying the computational power of concurrent data structures and synchronization primitives in multiprocessors and multicore machines for two and a half decades. He observes that Herlihy's classical hierarchy treats synchronization instructions as distinct objects, an approach that is far from the real-world, where multiprocessors do let processes apply supported atomic instructions to arbitrary memory locations. Gelashvili shows that, contrary to common belief, solving consensus does not require multicore architectures to support \"strong\" synchronization instructions such as compare-and-swap. Rather, combinations of \"weaker\" instructions such as decrement and multiply suffice. He goes on to propose an alternative complexity-based hierarchy for concurrent objects. The dissertation further opens a new line of research by proving a linear-space bound for the anonymous case of randomized consensus, the first major progress on this problem in 15 years, which won the Best Paper Award at DISC 2015, and for which Gelashvili developed novel lower bound techniques. Apart from their great importance, these results are also technically complex and mathematically beautiful.","PeriodicalId":198284,"journal":{"name":"Proceedings of the 2018 ACM Symposium on Principles of Distributed Computing","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2018-07-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"130113263","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}
We provide three different approaches to the minimum cut problem in the congested clique model of distributed computing. In this model, n nodes of the graph, each of which knows its own edges, can communicate in synchronous rounds; per round each node can send B-bits to each other node, where typically B=O(log n). At the end, each node should know its own part of the output, e.g., which side of the cut it is on. Our first algorithm is an O(1) round algorithm that finds a 1+o(1) approximation of the minimum cut. If the min-cut size is O(n^1/3 ), the algorithm finds an exact min-cut. This algorithm combines Karger's random sampling and his contraction algorithm; Nagamochi--Ibaraki--Nishizeki--Poljak's k--connectivity certificates; and Ahn--Guha--McGregor's algorithm for finding those certificates in the streaming model. To get an efficient implementation, we provide an algorithm that can solve simultaneously polynomially many instances of the MST problem in O(1) rounds. Our second algorithm is an O(log^3 n) round exact algorithm, based on the Karger-Stein approach. Its time complexity improves when larger messages are allowed. To implement this algorithm we present a general method to perform divide and conquer algorithms in the congested clique model. Our third algorithm is an O(log^2 n) round exact algorithm based on Karger's state of the art sequential exact min-cut algorithm, which works via tree-packing.
{"title":"Congested Clique Algorithms for the Minimum Cut Problem","authors":"M. Ghaffari, Krzysztof Nowicki","doi":"10.1145/3212734.3212750","DOIUrl":"https://doi.org/10.1145/3212734.3212750","url":null,"abstract":"We provide three different approaches to the minimum cut problem in the congested clique model of distributed computing. In this model, n nodes of the graph, each of which knows its own edges, can communicate in synchronous rounds; per round each node can send B-bits to each other node, where typically B=O(log n). At the end, each node should know its own part of the output, e.g., which side of the cut it is on. Our first algorithm is an O(1) round algorithm that finds a 1+o(1) approximation of the minimum cut. If the min-cut size is O(n^1/3 ), the algorithm finds an exact min-cut. This algorithm combines Karger's random sampling and his contraction algorithm; Nagamochi--Ibaraki--Nishizeki--Poljak's k--connectivity certificates; and Ahn--Guha--McGregor's algorithm for finding those certificates in the streaming model. To get an efficient implementation, we provide an algorithm that can solve simultaneously polynomially many instances of the MST problem in O(1) rounds. Our second algorithm is an O(log^3 n) round exact algorithm, based on the Karger-Stein approach. Its time complexity improves when larger messages are allowed. To implement this algorithm we present a general method to perform divide and conquer algorithms in the congested clique model. Our third algorithm is an O(log^2 n) round exact algorithm based on Karger's state of the art sequential exact min-cut algorithm, which works via tree-packing.","PeriodicalId":198284,"journal":{"name":"Proceedings of the 2018 ACM Symposium on Principles of Distributed Computing","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2018-07-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127835422","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}
We present a deterministic abortable mutual exclusion algorithm for a cache-coherent (CC) model with read, write, Fetch-And-Add (F&A), and CAS primitives, whose RMR complexity is O(log_W N) , where W is the size of the F&A registers. Under the standard assumption of W=Θ(log N), our algorithm's RMR complexity is Olog N/log log N); if W=Θ(N^ε), for 0 < ε < 1 (as is the case in real multiprocessor machines), the RMR complexity is O(1). Our algorithm is adaptive to the number of processes that abort. In particular, if no process aborts during a passage, its RMR cost is O(1).
{"title":"Deterministic Abortable Mutual Exclusion with Sublogarithmic Adaptive RMR Complexity","authors":"A. Alon, Adam Morrison","doi":"10.1145/3212734.3212759","DOIUrl":"https://doi.org/10.1145/3212734.3212759","url":null,"abstract":"We present a deterministic abortable mutual exclusion algorithm for a cache-coherent (CC) model with read, write, Fetch-And-Add (F&A), and CAS primitives, whose RMR complexity is O(log_W N) , where W is the size of the F&A registers. Under the standard assumption of W=Θ(log N), our algorithm's RMR complexity is Olog N/log log N); if W=Θ(N^ε), for 0 < ε < 1 (as is the case in real multiprocessor machines), the RMR complexity is O(1). Our algorithm is adaptive to the number of processes that abort. In particular, if no process aborts during a passage, its RMR cost is O(1).","PeriodicalId":198284,"journal":{"name":"Proceedings of the 2018 ACM Symposium on Principles of Distributed Computing","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2018-07-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127861603","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：