Comparison of hashing methods for supporting consistency in distributed databases

Abstract

The object of research is the methods of ensuring consistency in distributed systems. Distributed systems allow you to use any service regardless of the user's geolocation while maintaining high performance. With the increase in the volume and speed of information exchange in modern systems, it is more difficult to maintain consistency between different nodes. This can lead to the limited performance or malfunction when applied by the end users. In any case, the storage of outdated data can create conflict situations at the user level. The criticality of this depends on the industry in which the system operates and how important data consistency is. For example, domain DNS updates are done once a day and in most cases this is sufficient, but in the case of military systems this is unacceptable, as it can lead to human loss and financial loss due to the use of obsolete information. This work is a continuation of past work, which describes the PH-1 algorithm. The main goal was to show that it can be put on a pair with existing algorithms and to show the feasibility of use in distributed systems. The article compares the division-based algorithm and cryptographic hash functions MD5 and SHA-1. Based on the results of this analysis we can conclude that PH-1 is resistant to collisions in cases where the abovementioned algorithms form collisions, but at the same time impose certain restrictions for usage. This algorithm produces a hash value proportional to the length of the input data array, but certainly smaller than it. This means that the use of PH-1 makes almost no sense for the large data. One possible use may be to verify critical data, which should be as consistent as possible at different nodes of the distributed system. Ref. 4, pic. 2, tabl. 1.

Keywords

• distributed databases
• distributed systems
• hashing
• hash functions
• consistency
• collision resistance