Emergence of NoSQL Data Stores

Relational Database
1. Not good at handling big data
Relational and OOP databases don't work well together
OOP Basics
1. Class
  1. Object
2. Encapsulation, Polymorphism and Inheritance
Object Relational Impedance Mismatch
ORM (Object relational Mapping)
Enterprise Architecture
1. Integration Approach
2. Application Database
Scalability
NoSQL Technologies
1. neo4J
2. CouchDB
3. MongoDB
4. Cassandra
Why use NoSQL
1. Realtime interaction with users
2. Agility, Consistency, performance, scalability
3. Open Source
aggregate vs aggregate ingnorant
1. Relational Models are aggregate ignorant
  1. relational for same data requires many tables
2. NoSQL sees aggregate as clusters
  1. collection of data related to a unit
Graph database
1. neo4J
  1. Nodes form network
    1. Social networking connecting friends
Distribution models
1. High read
2. High write
Replication and Sharding
1. Sharding
  1. Different data across different nodes
2. Replication
  1. same data across different nodes
Replication & Master
1. Traffic only through master for write
  1. inappropriate for high write
Replication Peer to peer
1. Traffic written to any node
2. greater resilience
Combining Master/Slave and Sharding
1. All write through the server
  1. Different nodes have different data
    1. If master fails must know which node to write too
Consistency Locks
1. Pessimistic
  1. Lock table for editing at beginning of workflow
2. Optimistic
  1. Verify with server before making changes at end of workflow

Zusammenfassung der Ressource

	Erstellt von Jody Mc Laughlin vor fast 11 Jahre