MapReduce Performance in MongoDB Sharded Collections

Ajdari, Jaumin; Kasami, Brilant

doi:10.14569/ijacsa.2018.090617

Cited by 1 publication

(1 citation statement)

References 11 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…MongoDB uses some kinds of aggregation; for instance, the aggregation pipeline is a data aggregation system based on the concept of data processing pipelines. Documents enter a multi-stage pipeline that transforms the documents into aggregated results [19] [20].…”

Section: Graph Databasesmentioning

confidence: 99%

Improving The Performance of Big Data Databases

Arif

Ali

2019

KJAR

View full text Add to dashboard Cite

Real-time monitoring systems utilize two types of database, they are relational databases such as MySQL and non-relational databases such as MongoDB. A relational database management system (RDBMS) stores data in a structured format using rows and columns. It is relational because the values of the tables are connected. A non-relational database is a database that does not adopt the relational structure given by traditional. In recent years, this class of databases has also been referred to as Not only SQL (NoSQL). This paper discusses many comparisons that have been conducted on the execution time performance of types of databases (SQL and NoSQL). In SQL (Structured Query Language) databases different algorithms are used for inserting and updating data, such as indexing, bulk insert and multiple updating. However, in NoSQL different algorithms are used for inserting and updating operations such as default-indexing, batch insert, multiple updating and pipeline aggregation. As a result, firstly compared with related papers, this paper shows that the performance of both SQL and NoSQL can be improved. Secondly, performance can be dramatically improved for inserting and updating operations in the NoSQL database compared to the SQL database. To demonstrate the performance of the different algorithms for entering and updating data in SQL and NoSQL, this paper focuses on a different number of data sets and different performance results. The SQL part of the paper is conducted on 50,000 records to 3,000,000 records, while the NoSQL part of the paper is conducted on 50,000 to 16,000,000 documents (2GB) for NoSQL. In SQL, three million records are inserted within 606.53 seconds, while in NoSQL this number of documents is inserted within 67.87 seconds. For updating data, in SQL 300,000 records are updated within 271.17 seconds, while for NoSQL this number of documents is updated within just 46.02 seconds.

show abstract

Section: Graph Databasesmentioning

confidence: 99%