Transaction commit protocols for mobile environment

The proposed techniques for wireless environments during the last decade have limited support for dynamically changing environments. Due to its nature, the mobile computing environment is extremely dynamic and subject to rapid and unpredictable changes. Similarly, the characteristics of mobile applications affect their transactional requirements. The challenge is to reflect on solutions offering more flexibility and adaptability. In this article, the contribution was focused mainly on the problem of atomic commit that ensures the atomicity property. The trail of adapting mobile transaction commit protocols to context changes has been explored. This has led to the formalization of a flexible transaction model CATSM that supports adaptable properties and a commit protocol CA-TCP that enables adaptation to application requirements and mobile context in terms of transactional properties and execution cost. An architecture based on the concept of adaptation policy has also been designed for the implementation of the proposed solution.

Download Full-text

A Survey on Efficient Concurrency Control Algorithm in Distributed Database Systems

International Journal of Scientific Research in Computer Science Engineering and Information Technology ◽

10.32628/cseit1952228 ◽

2019 ◽

pp. 1176-1185

Author(s):

Rebecca Nyasuguta Arika ◽

W. Cheruiyot

Keyword(s):

Concurrency Control ◽

Control Algorithm ◽

Distributed Databases ◽

Database Systems ◽

Distributed Database ◽

Communication Overhead ◽

Distributed Database Systems ◽

Commit Protocols ◽

Message Exchange ◽

Transaction Commit

Transaction commit protocols help in reaching an agreement among the participating nodes when a transaction has to be committed or aborted. To initiate an agreement each participating node is asked to vote its decision on the operations on its transactional fragment. The participating nodes can decide to either commit or abort an ongoing transaction. In case of a node failure, the active participants take essential steps such as running the termination protocol to preserve database correctness. This paper sought to investigate the current distributed databases commit protocols such as 2PC and 3PC in order to pin-point their shortcomings. For instance, 2PC suffers from blocking of participant site in case of coordinator failure and increased latency due to forced writes of logs. On its part, 3PC suffers more communication overhead due to extra pre-commit phase. Based on these setbacks, an efficient protocol is suggested towards the end of this paper that it believed to address some of the challenges such as blocking and extra message exchange between communicating nodes.

Download Full-text