HAcid: A lightweight transaction system for HBase

No Thumbnail Available

URL

Journal Title

Journal ISSN

Volume Title

School of Science | Master's thesis
Checking the digitized thesis and permission for publishing
Instructions for the author

Authors

Date

2012

Department

Tietotekniikan laitos

Major/Subject

Tietojenkäsittelyteoria

Mcode

T-79

Degree programme

Language

en

Pages

77

Series

Abstract

The scalability of a database is an important issue for applications that deal with large amounts of data, such as web services. The presence of rapidly increasing high-volume data sets is a phenomenon commonly known as Big Data. As an alternative to traditional relational databases, the so-called NoSQL distributed databases have proved to be robust in Big Data applications. Most NoSQL databases, such as Bigtable and HBase, do not depend on high-end hardware, but are designed to easily scale by distributing the workload to a set of servers with conventional hardware. Cloud Computing infrastructures are suitable for these databases. HBase focuses on offering scalability and thus does not provide transactions with ACID (atomicity, consistency, isolation, durability) properties. Recently, however, there have been ~any attempts towards supporting ACID transactions in these databases. One important application of this feature is the support for incremental updates to a data repository, such as a web search index. Most of the existing transactional systems for HBase are built on top of HBase itself, with transactional metadata in the database and algorithms in the client-side. We have built HAcid, a new open-source transactional system for HBase. As most similar existing systems, it is a client library that keeps transactional metadata in HBase to avoid introducing new server-side software. The novelty of HAcid is its lightweight characteristics: it uses minimal bookkeeping resources and is straightforward to install. The purpose is not to minimize transaction latency, but to provide an easy approach to ACID transactions in HBase.

Description

Supervisor

Heljanko, Keijo

Thesis advisor

Schumacher, André

Keywords

distributed database, transaction, cloud computing, HBase

Other note

Citation

Permanent link to this item

https://urn.fi/URN:NBN:fi:aalto-2020122859139

Collections

[dipl] Perustieteiden korkeakoulu / SCI