Nosql Introduction, Basics
3 likes415 views
This document discusses NoSQL databases as alternatives to SQL databases. It defines NoSQL as databases that don't use the relational data model or have a fixed schema. The document outlines several types of NoSQL databases and discusses considerations for evaluating them, such as data model, query model, consistency model, and commercial support. It provides examples of how companies like Twitter, Instagram, Pinterest, and LinkedIn use NoSQL databases for applications like caching, logging, and graph storage.
Nosql Introduction, Basics
- 1. NosqL
- 2. SQL has Ruledfor two decades Store persistent data Application Integration Concurrency Control Mostly Standard Reporting
- 4. don’t use the relational data model, tend to be designed to run on a cluster don’t have a fixed schema, allow tostore any data in any record seeks to solve the scalabilityandbig data performance issues
- 5. 5 Considerations When EvaluatingNoSQL Data Model Querymodel Consistencymodel API Commercialsupport
- 6. Types of NoSQL databases Document: CouchDB, Couchbase, MongoDB Key-value:Riak,MemcacheDB, Redis Graph: Neo4J Column: Cassandra, HBase
- 7. Document Based { _id:”5466b7c22”, company:{ _id:”5466b7c22”, company: “google”, symbol : “GOOG”, stock_price: “858”, stock_volume:10M }}
- 11. Query Model ThebiggestdifferencebetweenNoSQLsystemsliesintheability toquerydata efficiently. Document databases provide the richest query Key-value stores providea singlemeansofaccessingdata:by key.
- 12. Consistency Model CONSISTENT SYSTEMS EVENTUALLY CONSISTENT SYSTEMS
- 13. Reasons to use NoSQL databases Scale (horizontal) Simpler datamodel (less joins) Redundancy/Reliability Schema less (no modelling or prototyping) Rapid Development/coder friendly Flexibility semi-structured/ unstructured/Structured Cheaperthanrelational/ commodity Creatinga cachingLayer Environment data/logs Graphs/relationships Distributedstorage Realtime analysis
- 14. Challenges against NoSQL ACID Maturity Analytics/ BI / Reporting Ecosystem/tools/addons search Security Data loss duplicatedata(Normalisation) Administration Expertiseavailability
- 15. Relationalis not dead Thisleads us to a world of Polyglot Persistence
- 17. How Twitter UsesNoSQL is heavily dependent on MySQL users generate 12terrabytes of data a day - about four petabytes peryear Twitter uses Scribe to write logs to Hadoop Hbase is built on top of Hadoop using Cassandra for atomic counting.
- 18. Does Stack Overflow use caching? Redis for site cache and global cache around 1,300,000 keys in our Redis cache a few hundred read/writes a second to Redis CPU usage on their dedicated Redis machine is0 – 1 percent Memory usage is around 6G
- 19. Instagram loves Redis for caching complex objects Pinterest uses Redis to store the graph of followers and caching, HBase LinkedIn uses Hadoop, Voldemort and Espresso.
- 20. Can we use NoSql in our project?
Editor's Notes
- #4: Not Only Sql There is no standard definition of what NoSQL means. The term began with a workshop organized in 2009,
- #18: Twitter uses Pig, a high-level language running on top of Hadoop. Yahoo created Pig for rapid Hadoop development, Hbase is built on top of Hadoop and is designed for low-latency and data mutability. Twitter uses it to power its people search. FlockDB is a real-time, distributed database. As mentioned above, it was created and open-sourced by Twitter. The company uses it for social graph analysis. It's still MySQL underneath Cassandra, an open-source NoSQL database created by Facebook.