Facebook Messenger Case Study with Apache Hbase

Let’s discuss about the Facebook Messenger Case Study with Apache Hbase.

Introduction

We all know that Facebook Messenger is an instant messaging service. Originally developed by Facebook in 2008.

Here users can send messages, photos, videos, stickers, audio, and files, as well as react to other users’ messages and interact with bots. The service also supports voice and video calling. The Messenger service supports 300 million users to send 120 billion messages per month.

Earlier during 2008 Facebook was using open-source database called Cassandra, it is failed to handle huge data sets efficiently generated by Facebook Messenger, as the indexes and data sets grew large, the performance decreases. Facebook Messenger started to face some of the major problems like,

  1. Storage problem.
  2. Fails to perform fast processing.
  3. Low performance.
  4. Fails to maintain consistency with storage and performance.

Hbase came up with the solution for the above problems faced by Facebook Messenger.

Hbase started to provides some of the features like,

  1. Auto load balancing and failover of huge data set.
  2. Compression support to store huge data set in hdfs.
  3. Multiple shards ( It means many nodes in the cluster) per server.

HDFS is the underlying file system used by Hbase. It provides several features like end-to-end checksums, replication and automatic load re-balancing.

Solution

Hbase provides complete solution for the Facebook messenger probem is as shown in the below diagram.

Figure: HBase as a solution to Facebook messenger

Hbase provides solutions to Facebook messenger like,

  1. It started to store all the streaming data generated from various Fcaebook services like chat, message, email and sms etc., using HDFS as underlying file system.
  2. Hbase is a column oriented database, perform fast processing of data.
  3. It is mainly designed and implemented for real time data streaming for faster data retrieving mechanism using random access technique it achieve high performance.
  4. Hbase is fault tolerance and auto failover in nature.
  5. It will analyze a large set of semi-structured, unstructured data and structured data.

“That’s all about the Facebook Messenger Case Study with Apache Hbase”