Benchmarking Top NoSQL Databases: Apache Cassandra, Apache HBase and MongoDB

Referent
Einrichtung Titel des Vortrages 1
WP-Benchmarking Top NoSQL
Databases
Apache Cassandra, Apache HBase and MongoDB
Presented By
Athiq Ahamed
Supriya

Referent
Introduction
 Enormous amount of data-BigData
 Scalabilty issue in RDBMS
 Rise of NoSQL databases
 Amazon Dynamo
 Big table
 CAP Theorem
 BASE system

Referent
CAP Theorem
 Consistency
 Availability
 Partition tolerance
CAP theorem states that only two of the properties can be
achieved at a time.

Referent
RDBMS NoSQL
Supports powerful query
language
Supports very simple query
language
It has a fixed schema No fixed schema
Follows ACID (Atomicity,
Consistency, Isolation and
Durability)
It is only eventually consistent
Supports transactions Does not support transactions
RDBMS vs NoSQL
Content:tutorialspoint.com

Referent
 Basically available: System guarantees availability, in
terms of the CAP theorem
 Soft state: State of the system may change over time,
because of eventual consistency model
 Eventual consistency: System will become consistent over
time
BASE
Content:www.edureka.in

Referent
 Fast Performance is the key.
 POC processes include right benchmarks:
 Configurations
 Parameters
 Workloads
Making the right choice!
Selection of NoSQL

Referent
 Yahoo Cloud Serving Benchmark (YCSB)
 Top 3 NoSQL databases-Apache Cassandra, Apache
Hbase and MongoDB.
 Amazon Web Services EC2 instances for hosting the tests
 Test performed 3 times on 3 different days
Benchmark configuration

Referent
 The tests ran on large size instances (15GB RAM and 4
CPU cores)
 Instances used customized Ubuntu with Oracle Java 1.6
installed as a base.
 A customized script written to drive the benchmark
processes
Benchmark configuration

Referent
 Each NoSQL system performs differently, not alike.
 Components and Internal working.
 Apache Cassandra: Columnar database model
 Apache HBase: Columnar database model
 MongoDB: Document storage database model
Understanding NoSQL Databases

Referent
Apache Cassandra
 Cassandra is scalable, fault-tolerant, and consistent. All
nodes are equal.
 Its distribution design is based on Amazon’s Dynamo and
its data model on Google’s Bigtable.
 Key components: Node, Cluster, Commit log, Mem-table,
SSTable and Bloom filter
Content:http://www.tutorialspoint.com/cassandra/cassandra_architecture.htm

Referent
 Ring structure, peer to peer architecture
 All nodes are equal
 This improves general database availablity
 Scaling up and scaling down is easier
 Cassandra has key-value, column oriented database
Apache Cassandra

Referent
Apache Cassandra
Content:http://demoiselle.sourceforge.net/component/demoiselle-
cassandra/1.0.0/images/datamodel1.png

Referent
 Cassandra has an internal keyspace called system, stores
metadata about the cluster.
 Metadata:
 The node‘s token
 The cluster name
 Keyspace n schema definitions (dynamic loading)
 Whether or not the node is bootstrapped
Apache Cassandra
Content:https://www.edureka.co/blog/category/apache-cassandra/

Referent
 Commit log: Crash recovery mechanism. Every write
operation is written to commit log
 Mem-Table: A memory resident data structure.
 SSTable: It is a disk file to which the data is flushed from
the mem-table
Apache Cassandra

Referent
 Bloom filters are used as a performance booster
 Bloom filter are very fast, quick algorithms for testing a
member in the set.
 Bloom filters serves as a special kind of cache – quick
lookups/search as they reside in memory
Apache Cassandra

Referent
 Gossip protocol: Communiction between nodes, co-
ordination and failure check
 Anti-Entropy protocol: Replica sync mechanism enusing
data on different nodes are updated (Merkle trees)
 Snitches ensures host proximity
Apache Cassandra

Referent
Apache Cassandra- Read/Write operation

Referent
 Sparse, distributed, sorted map and multidimensional and
consistent.
 Hbase is a Key/value store
 Consists Row key, Column family, columns and timestamp.
Apache HBase

Referent
Apache HBase
Content:http://zhangjunhd.github.io/assets/2013-02-25-apache-hbase/rowkey-

Referent
 Region: Contiguous rows form a region
 Region server(RS): Serves one or more regions.
 Master server: Daemon responsible for managing Hbase
cluster
 HDFS: Distributed, open source file system containing
HBase‘s data
 Zookeeper: Distributed, open source co-ordinated service
for co-ordination of master and region servers.
Apache HBase Components
Content: https://www.mapr.com/blog/in-depth-look-hbase-architecture

Referent
Apache Hbase Architecture

Referent
 Client obtains meta table RS from Zookeeper
 Client gets RS which holds the corresponding rowkey
 Client receives the row from the respective Region server
 Client caches this information along with the location of
meta table server.
First Read/Write to HBase

Referent
 WAL: Write Ahead Log is a file on the distributed file
system. It is used to store new data
 Block Cache: It is the read cache. It stores frequently
read data in memory
 Mem Store: Write cache that stores new data which is not
written to disk yet.
 Hfiles stores the rows as sorted key values on disk
HBase RS Components

Referent
 Client writes the data to the WAL file stored on disk
 WAL is used to recover not yet persisted data in case a
server crashes.
 Once data is written to WAL, it is placed in Mem Store
Hbase Write steps (1)

Referent
 All write/read are to/from the primary node.
 HDFS replicates WAL and Hfile blocks. Replication
happens automatically.
 When data is written in HDFS, one copy is written locally
and then it is replicated to a secondary node and later to
tertiary node.
HDFS Write steps (2)

Referent
 Cassandra usecase: Availability and Partition tolerant
requirements.
Consistency is tunable by setting it high in the option
 Hbase usecase: Consistency and Scalability. However, at
less number of nodes/threads, availability is achieved high
Cassandra and Hbase

Referent
 Document-oriented database
 High performance and automatic scaling
 High consistency and partition tolerant
 Replication and failover for high availability
 Low latency
 Flexible indexing
MongoDB

Referent
 Document is the basic unit for MongoDB(row)
 Collection is similar to a table
 A single instance has multiple independent databases
 Every document has a special key, “_id”
 Powerful JavaScript shell for administration
 Configdb contains metadata of clusters
MongoDB Concepts

Referent
MongoDB Simple Architecture

Referent
 A mongo receives queries from applications
 Uses metadata from config server for the data
 Mangos directs write operations to a particular shard
 Mongos uses the cluster metadata from the config
database
Read/Write MongoDB

Referent
 Scalability
 Availability
 Partition Tolerant
 Consistency
MOST IMPORTANT PERFORMANCE
Yahoo Cloud Serving Benchmark (YCSB)
Recap Importance of Benchmark and Factors

Referent
Results: Load Process

Referent
Results: Read/Write Mix Workload

Referent
Results: Read/Scan Mix Workload

Referent
Results: Read Latency across all workloads

Referent
Results: Insert Latency across all workloads

Referent
Lets MIGRATE from traditional data base !!!!
Live Demo

Referent
 Identify data model for the application
 Corresponding data sets have to be known
 Whether the application requires replication
 Identify the performance requirements
 Prototype the application
 Test the performance of the prototype
Discussion

Referent
Conclusion
 NoSQL replaced tradition relational databases
 Performance is the key feature
 Importance of benchmarks
 Top three NoSQL data base’s performance tested
 Cassandra outperforms all the other NoSQL data bases
 Decide based on application

Referent

Benchmarking Top NoSQL Databases: Apache Cassandra, Apache HBase and MongoDB

More Related Content

What's hot (20)

Viewers also liked (6)

Similar to Benchmarking Top NoSQL Databases: Apache Cassandra, Apache HBase and MongoDB (20)

Recently uploaded (20)

Benchmarking Top NoSQL Databases: Apache Cassandra, Apache HBase and MongoDB

Editor's Notes