On the Growth of the Prime Numbers Based Encoded Vector Clock
The vector clock is a fundamental tool for tracking causality in parallel and distributed applications. Unfortunately, it does not scale well to large systems because each process needs to maintain a vector of size n, where n is the total number of processes in the system. To address this problem, the encoded vector clock (EVC) was recently proposed. The EVC is based on the encoding of the vector clock using prime numbers and uses a single number to represent vector time. The EVC has all the properties of the vector clock and yet uses a single number to represent global time. However, the single number EVC tends to grow fast and may soon exceed the size of the traditional vector clock. In this paper, we evaluate the growth rate of the size of the EVC using a simulation model. The simulations show that the EVC grows relatively fast, and the growth rate depends on the mix of internal events and communication events. To overcome this drawback, the EVC can be used in conjunction with several scalability techniques that can allow the use of the EVC in practical applications. We then present a case study of detecting memory consistency errors in MPI one-sided applications using EVC.
KeywordsCausality Vector clock Prime numbers Encoding Happened-before relation Scalability Performance
We thank Rahul Sathe for his help with the simulations.
- 1.Ahuja, M., Kshemkalyani, A.D., Carlson, T.: A basic unit of computation in distributed systems. In: 10th International Conference on Distributed Computing Systems (ICDCS 1990), 28 May–1 June 1990, Paris, France, pp. 12–19 (1990)Google Scholar
- 3.Chen, Z., Dinan, J., Tang, Z., Balaji, P., Zhong, H., Wei, J., Huang, T., Qin, F.: MC-Checker: detecting memory consistency errors in MPI one-sided applications. In: International Conference for High Performance Computing, Networking, Storage and Analysis, SC 2014, New Orleans, LA, USA, 16–21 November 2014, pp. 499–510 (2014)Google Scholar
- 4.Diep, T.D., Fürlinger, K., Thoai, N.: MC-CChecker: a clock-based approach to detect memory consistency errors in MPI one-sided applications. In: Proceedings of the 25th European MPI Users’ Group Meeting, EuroMPI 2018, pp. 9:1–9:11 (2018)Google Scholar
- 6.Flanagan, C., Freund, S.N.: FastTrack: efficient and precise dynamic race detection. In: Proceedings of the 2009 ACM SIGPLAN Conference on Programming Language Design and Implementation, PLDI 2009, Dublin, Ireland, 15–21 June 2009, pp. 121–133 (2009)Google Scholar
- 9.Kshemkalyani, A.D., Khokhar, A.A., Shen, M.: Encoded vector clock: using primes to characterize causality in distributed systems. In: Proceedings of the 19th International Conference on Distributed Computing and Networking, ICDCN 2018, Varanasi, India, 4–7 January 2018, pp. 12:1–12:8 (2018)Google Scholar
- 12.Kulkarni, S.S., Vaidya, N.H.: Effectiveness of delaying timestamp computation. In: Proceedings of the ACM Symposium on Principles of Distributed Computing, PODC 2017, Washington, DC, USA, 25–27 July 2017, pp. 263–272 (2017)Google Scholar
- 14.Mattern, F.: Virtual time and global states of distributed systems. In: Proceedings of the Parallel and Distributed Algorithms Conference, pp. 215–226 (1988)Google Scholar
- 15.Meldal, S., Sankar, S., Vera, J.: Exploiting locality in maintaining potential causality. In: Proceedings of the Tenth Annual ACM Symposium on Principles of Distributed Computing, PODC 1991, pp. 231–239 (1991)Google Scholar
- 18.Shen, M., Kshemkalyani, A.D., Khokhar, A.A.: Detecting unstable conjunctive locality-aware predicates in large-scale systems. In: IEEE 12th International Symposium on Parallel and Distributed Computing, ISPDC 2013, Bucharest, Romania, 27–30 June 2013, pp. 127–134 (2013)Google Scholar
- 21.Ward, P.A.S., Taylor, D.J.: A hierarchical cluster algorithm for dynamic, centralized timestamps. In: Proceedings of the 21st International Conference on Distributed Computing Systems (ICDCS 2001), Phoenix, Arizona, USA, 16–19 April 2001, pp. 585–593 (2001)Google Scholar