Relational vs. Non-Relational
6 likes9,572 views
This document compares relational and non-relational databases. It discusses how in 2003 the main databases were relational, but by 2010 non-relational databases grew popular in the "NoSQL movement". However, the document argues that there are no truly new database designs and that relational and non-relational databases can be combined. It advises to choose a database based on the specific problem and features needed rather than general classifications. The document provides examples of which types of databases fit certain data and access needs.
Relational vs. Non-Relational
- 1. Relational vs. Non-Relational Josh Berkus PostgreSQL Experts Inc. SCALE 2010
- 2. 2003: â—Ź MySQL â—Ź PostgreSQL â—Ź FireBird â—Ź BerkeleyDB â—Ź Derby â—Ź HSQLDB â—Ź SQLite
- 3. 2003: â—Ź MySQL â—Ź PostgreSQL â—Ź FireBird â—Ź BerkeleyDB â—Ź Derby â—Ź HSQLDB â—Ź SQLite
- 4. 2010 â—Ź MySQL â—Ź PostgreSQL â—Ź FireBird â—Ź BerkeleyDB â—Ź Derby â—Ź HSQLDB â—Ź SQLite
- 8. G ra ph D o c um ent Neo4J C ouchD B HyperGraphDB B erkeleyD B -X M L Jena S olr K ey-va lue D is tributed M emcached C assandra Tokyo C abinet Hypertable db4o M yS Q L ND B R IAK H iera rc hic a l M ongoD B
- 10. Are not the same
- 11. E m bedded O LT P S QLite P ostgreS Q L Firebird M yS Q L HS Q L O racle S Q L S erver M PP S trea m ing TeraD ata Greenplum S treambase Aster Truviso C -S to re LucidD B M onetD B
- 12. NoFins
- 15. Mythbust #2
- 17. There
- 18. are
- 19. no
- 20. new
- 21. database
- 22. designs
- 23. There are only new implementations and combinations
- 24. “A database storing application- friendly formatted objects, each containing collections of attributes which can be searched through a document ID, or the creation of ad-hoc indexes as needed by the application.”
- 25. CouchDB, 2007 Pick, 1965
- 26. CouchDB, 2007 embeddable Pick JSON storage REST API map/reduce
- 29. “renaissance of non-relational databases”
- 30. Mythbust #3
- 31. “non-relational databases are toys”
- 32. Google Bigtable
- 33. Amazon Dynamo
- 35. US Vetrans' Administration Pick, Caché
- 36. Mythbust #4
- 37. “Relational databases will become obsolete”
- 38. “Three decades past, the relational empire conquered the hierarchical hegemony. Today, an upstart challenges the relational empire's dominance ...” XML Databases 2001 --Philip Wadler, Keynote VLDB, Rome, September 2001
- 39. Anyone remember XML databases?
- 40. No?
- 41. What happened?
- 42. established relational and non-relational databases hybridized XML
- 43. Oracle XML PostgreSQL XML2 BerkeleyDB XML DB2
- 44. Mythbust #5
- 45. “Relational databases are for when you need ACID transactions.”
- 46. Transactions ≠Relational
- 47. Robust Transactions without Relationality: BerkeleyDB Amazon Dynamo SQL Without Transactions: MySQL MyISAM MS Access
- 48. Mythbust #6
- 50. You
- 51. do
- 52. not
- 53. have
- 54. to choose
- 55. one database.
- 56. Choose the database system which f ts your i current application goals. or ...
- 57. Use more than one together MySQL + Memcached PostgreSQL + CouchDB or ...
- 58. Use a Hybrid MySQL NDB PostgreSQL Hstore HadoopDB
- 59. But what about relational vs non-relational?
- 60. Relational OLTP Databases* ● Transactions: more mature support – including multi-statement ● Constraints: enforce data rules absolutely ● Consistency: enforce structure 100% ● Complex reporting: keep management happy! ● Vertical scaling (but not horizontal) * mature ones, anyway
- 61. SQL vs. Not SQL SQL promotes: â—Ź portability â—Ź managed changes over time â—Ź multi-application access â—Ź many mature tools
- 62. SQL vs. Not SQL But … SQL is a full programming language, and you have to learn it to use it.
- 63. SQL vs. Not SQL No-SQL allows: â—Ź programmers as DBAs â—Ź no impedance â—Ź fast interfaces â—Ź fast development
- 64. The main reason to use SQL-RDBMSs
- 65. “Immortal Data” your data has a life independent of this specif c i application implementation
- 66. How do I choose?
- 67. Def ne the problem i you are trying to solve
- 68. ● “I need a database for my blog” ● “I need to add 1000's of objects per second on a low- end device.” ● “I need my database to enforce business rules across several applications.” ● “I want my application to be location-sensitive.” ● “I need to cache data an access it 100K times per second.” ● “I need to produce summary reports across 2TB of data.” ● “I have a few hundred government documents I need to serve on the web and mine” ● “I need to know who-knows-who-knows-who.” ● “I need to data mine 1000's of 30K bug reports per minute.”
- 69. Def ne the features i you actually need â—Ź many connections â—Ź multi-server scalability â—Ź complex query logic â—Ź APIs â—Ź redundancy â—Ź data integrity â—Ź schema/schemaless â—Ź data mining
- 70. i f t the database to the task
- 71. “I need a database for my blog”
- 72. Use anything! â—Ź MySQL â—Ź PostgreSQL â—Ź MongoDB â—Ź SQLite â—Ź CouchDB â—Ź Flatf les i â—Ź DBaseIII â—Ź Something you wrote yourself
- 73. “I need my database to unify several applications and keep them consistent.”
- 74. PostgreSQL “OLTP SQL-Relational Database” “It's not just a database: it's a development platform”
- 75. Postgres 9 alpha out now!
- 76. “I need my application to be location-aware.”
- 77. PostGIS “Geographic Relational Database”
- 78. PostGIS ● Queries across “contains” “near” “closest” ● Complex geometric map objects – polygons – lines (roads, etc) ● Make your own Google Maps! – Sunday, 4:30pm, Concourse A
- 79. “I need to store 1000's of event objects per second on embedded hardware.”
- 80. db4object “Embedded Key-Value Store”
- 81. db4object â—Ź German Train System â—Ź Insert 1000's of objects per second â—Ź Low-end embedded console computer â—Ź Simple access in native programming language (Java, .NET) â—Ź compromise: embedded SQL database: SQLite
- 82. “I need to access 100K objects per second over 1000's of connections.”
- 83. memcached “Distributed In-Memory Key-Value Store”
- 84. memcached â—Ź Use: public website â—Ź Used for caching 1000's of serialized objects per second â—Ź Available for 100000's of requests per second across 1000's of connections â—Ź Cache each object only once per site â—Ź Supplements a relational database Alternatives: Redis, TokyoTyrant
- 85. “I need to produce complex summary reports over 2TB of data.”
- 87. LucidDB â—Ź For reporting and analysis â—Ź Large quantities of data (TB) â—Ź Complex OLAP & analytics queries â—Ź Business intelligence â—Ź compliments a transactional database
- 88. “I have 100's of government documents I need to serve on the web and mine for data.”
- 90. CouchDB 1.CividDB Project 2.Storing lots and lots of government documents 3.Don't know what's in them 4.Don't know how they are structured 5.Store them, f gure out structure later by data i mining. It's also good for web sites!
- 91. “I have a social application and I need to know who-knows- who-knows- who-knows- who-knows-who.”
- 93. Neo4j ● Social Network Website ● 6 degrees of separation ● “you may also like” ● type and degrees of relationship
- 94. “I get 1000's of 30K bug reports per minute and I need to mine them for trends.”
- 95. Hadoop “Massively Parallel Data Mine”
- 96. Hadoop â—Ź Massive bug report data feed â—Ź 1000's of bug reports per minute â—Ź Each bug report 2-45K of data â—Ź Need to extract trends and correlate inexact data â—Ź Summarize in daily & weekly reports
- 97. Conclusion ● Different database systems do better at different tasks. – every database feature is a tradeoff – no database can do all things well ● Relational vs. non-relational doesn't matter – pick the database(s) for the project or the task
- 98. Questions? ● PostgreSQL Project ● PostgreSQL Booth – www.postgresql.org trade show f oor l – josh@postgresql.org ● PostgreSQL 9.0 ● PostgreSQL Experts BOF, Tonight 7pm – www.pgexperts.com ● www.pgexperts.com/documents.html ● Open Source Database Survey – Selena Deckleman – Open Source Database Survey: www.ossdbsurvey.org Copyright 2010 Josh Berkus, distributable under the creative commons attribution license, except for 3rd-party images which are property of their respective owners. Special thanks to Selena Deckelman for the bunnies image and for the Open Source Database survey.