Publications

MRV*: A data type library for concurrent applications

INForum: National Symposium of Informatics, 2023

Authors

Abstract

The existence of items that are frequently accessed and modified is a significant obstacle to achieving high performance in data management systems. It is therefore becoming increasingly important to find methods of dealing with the simultaneous execution of operations that manipulate data efficiently, providing applications with lower response times. However, existing approaches do not allow invariants to be guaranteed over the data, have strict assumptions about concurrency, or are valid only for numerical values. This work aims to build a library of data structures that tolerate high concurrency through partitioning and randomization, as proposed for MRVs and suitable for distributed systems, but offering a more diverse range of structures to support different applications. This proposal is evaluated with an implementation in SQL database systems.

BibTeX

@article{mrv-library,
	title     = {MRV*: A data type library for concurrent applications},
	author    = {Maior, Adriano and Faria, Nuno and Pereira, José},
	journal   = {Proceedings of the 2023 National Symposium of Informatics},
	volume    = {14},
	year      = {2023},
}

Resources