NoSQL offers the new strategy for data management which is designed to satisfy the velocity, volume and variety of data. This technology store and retrieve data which is not in the relational database. NoSQL is derived from “Not only SQL” to give more value like SQL query languages. The most popular Nosql Databases- MongoDB, Cassandra, HBase. The NoSQL databases will be covered in all Big Data Training in Chennai.
The following are the challenges of Relational databases
- It is not suitable for petabytes of data with different data types like videos, images, text etc.
- It has complex RDBMS model
- It is limited by CPU capabilities and memory
- It is limited by cache dependent read and write operations
- The process like Sharding will lead to operational problems.
- It limits the scalability in RDBMS.
NoSQL databases provide high performance and scalable, compared to relational databases. Relational database limitations are solved in NoSQL database. Hadoop Online Training is the best place to improve your knowledge in Hadoop.
With the concepts of BigTable, wide column store based on Apache Hadoop
HBase runs on top of the HDFS used to store value/key. HBase is split into tables and tables are further divided into column, column is declared in schema and finally a set of columns. Each and every value in HBase is called as cell and key includes column family, row-key, time-stamp, and column. HBase rows groups the value mappings which is identified by the row-key. Hadoop’s infrastructure is suitable for HBase and with the help of shelf servers it scales horizontally.
Based on BigTable and DynamoDB, wide column store is happening.
Cassandra is the most familiar NoSQL and DDBMS is managing more business today by providing high scalability, availability at the lower cost. Cassandra has decentralized architecture, here any operation can perform by any node also from the CAP theorem it offers AP. When compared to HBase reads, consistency is slower in Cassandra quorum reads. The eventual consistency is the great advantage in Cassandra. It will not support row based scans which is restricted in some use cases. Thus, Cassandra is good for single row or multiple rows based on column value.
MongoDB is the most familiar document stores. MongoDB – document based database. JSON/BSON is the format used in MongoDB for the data use. Master slave replications are used to make multiple copies of data. It maintains the databases effectively in expressive query language, strong consistency.
These were the three major NoSQL databases to be noticed in the recent year. Make use of this article for your enhancement. If you like to learn more and more about the NoSQL databases, take up Hadoop Training in Bangalore.