Taming the Contention in Consensus-Based Distributed Systems

被引:1
作者
Arun, Balaji [1 ]
Peluso, Sebastiano [2 ]
Palmieri, Roberto [3 ]
Losa, Giuliano [4 ]
Ravindran, Binoy [1 ]
机构
[1] Virginia Tech, Bradley Dept Elect & Comp Engn, Blacksburg, VA 24061 USA
[2] Facebook Inc, Seattle, WA 98109 USA
[3] Lehigh Univ, Bethlehem, PA 18015 USA
[4] Galois Inc, Portland, OR 97204 USA
基金
美国国家科学基金会;
关键词
Consensus protocol; Delays; Computer crashes; Switches; Runtime; Fault tolerance; Distributed systems; fault tolerance; consensus; leaderless consensus; contention-agnostic consensus;
D O I
10.1109/TDSC.2020.2970186
中图分类号
TP3 [计算技术、计算机技术];
学科分类号
0812 ;
摘要
Contention plays a crucial role in the design of consensus protocols. State-of-the-art solutions optimize their performance for either very low or high contention situations. We propose Caesar, a novel multi-leader Generalized Consensus protocol, most suitable for geographical replication, that is optimized for low-to-moderate contention. With an evaluation study, we show that Caesar outperforms other multi-leader (e.g., EPaxos) and single-leader (e.g., Multi-Paxos) competitors by up to 1.7x and 3.5x, respectively, in the presence of 30 percent conflicting requests, in a geo-replicated setting. Furthermore, we acknowledge that there is no one-size-fits- all consensus solution, especially for all levels of contentious workloads. Thus, we also propose Spectrum, a consensus framework that is able to switch consensus protocols at runtime to enable a dynamic reaction to changes in the workload and deployment characteristics. We show empirically that Spectrum can guarantee high availability even during periods of transition between consensus protocols.
引用
收藏
页码:2907 / 2925
页数:19
相关论文
共 32 条
[1]  
Arun B., 2019, SPECTRUM FRAMEWORK A
[2]  
Arun B., 2017, SPEEDING CONSENSUS C
[3]   Speeding up Consensus by Chasing Fast Decisions [J].
Arun, Balaji ;
Peluso, Sebastiano ;
Palmieri, Roberto ;
Losa, Giuliano ;
Ravindran, Binoy .
2017 47TH ANNUAL IEEE/IFIP INTERNATIONAL CONFERENCE ON DEPENDABLE SYSTEMS AND NETWORKS (DSN), 2017, :49-60
[4]   The Nearest Replica Can Be Farther Than You Think [J].
Bogdanov, Kirill ;
Peon-Quiros, Miguel ;
Maguire, Gerald Q., Jr. ;
Kostic, Dejan .
ACM SOCC'15: PROCEEDINGS OF THE SIXTH ACM SYMPOSIUM ON CLOUD COMPUTING, 2015, :16-29
[5]   Uniform consensus is harder than consensus [J].
Charron-Bost, B ;
Schiper, A .
JOURNAL OF ALGORITHMS-COGNITION INFORMATICS AND LOGIC, 2004, 51 (01) :15-37
[6]   Spanner: Google's Globally Distributed Database [J].
Corbett, James C. ;
Dean, Jeffrey ;
Epstein, Michael ;
Fikes, Andrew ;
Frost, Christopher ;
Furman, J. J. ;
Ghemawat, Sanjay ;
Gubarev, Andrey ;
Heiser, Christopher ;
Hochschild, Peter ;
Hsieh, Wilson ;
Kanthak, Sebastian ;
Kogan, Eugene ;
Li, Hongyi ;
Lloyd, Alexander ;
Melnik, Sergey ;
Mwaura, David ;
Nagle, David ;
Quinlan, Sean ;
Rao, Rajesh ;
Rolig, Lindsay ;
Saito, Yasushi ;
Szymaniak, Michal ;
Taylor, Christopher ;
Wang, Ruth ;
Woodford, Dale .
ACM TRANSACTIONS ON COMPUTER SYSTEMS, 2013, 31 (03)
[7]  
Couceiro M, 2013, I C DEPEND SYS NETWO
[8]   Transactional Auto Scaler: Elastic Scaling of Replicated In-Memory Transactional Data Grids [J].
Didona, Diego ;
Romano, Paolo ;
Peluso, Sebastiano ;
Quaglia, Francesco .
ACM TRANSACTIONS ON AUTONOMOUS AND ADAPTIVE SYSTEMS, 2014, 9 (02)
[9]   Clock-RSM: Low-Latency Inter-Datacenter State Machine Replication Using Loosely Synchronized Physical Clocks [J].
Du, Jiaqing ;
Sciascia, Daniele ;
Elnikety, Sameh ;
Zwaenepoel, Willy ;
Pedone, Fernando .
2014 44TH ANNUAL IEEE/IFIP INTERNATIONAL CONFERENCE ON DEPENDABLE SYSTEMS AND NETWORKS (DSN), 2014, :343-354
[10]   IMPOSSIBILITY OF DISTRIBUTED CONSENSUS WITH ONE FAULTY PROCESS [J].
FISCHER, MJ ;
LYNCH, NA ;
PATERSON, MS .
JOURNAL OF THE ACM, 1985, 32 (02) :374-382