nosql and mongodb: ¯\_(ツ)_/¯

An overview of nosql, distributed systems and mongodb for the Kuali Days developer conference.

Transcript

1. nosql ¯\_(ツ)_/¯
2. ALERT ALERT ALERT
4. nosql
5. what even is nosql "A NoSQL database provides a mechanism for storage and retrieval of data that is modeled in means other than the tabular relations used in relational databases." - wikipedia
6. tradeoffs (with exceptions) heterogeneous data vs defined schemas write speed vs consistency many machines vs one or few machines OLAP vs OLTP
7. then why? they *might* model your data better they can be faster they can be used as secondary stores they *might* support larger scale
8. nosql taxonomy Document Key-Value Other MongoDb RethinkDb CouchDb Riak DynamoDb neo4j Redis HBase
9. someone else explains this well
10. distributed systems detour
11. CAP Theorem consistency - a read sees all previously completed writes availability - reads and writes always succeed partition tolerance - system works despite failures in parts
12. CAP Theorem PICK TWO AND YOU HAVE TO PICK PARTITION TOLERANCE AND IF YOU MESS UP YOU MIGHT GET NONE OF THEM
13. end distributed systems detour
14. Riak conflict resolution "With all of our options available, we want to resolve this conflict. Since the way of resolving this conflict is largely use-case specific, our application can decide how to proceed." - Little Riak Book
15. WELP
16. Use Riak 70 million players 7.5 million concurrent users 11,000 messages per second League Of Legends Chat http://highscalability.com/blog/2014/10/13/how-league-of- legends-scaled-chat-to-70-million-players-it-t.html
17. mongodb ¯\_(ツ)_/¯ schema-less document store
18. sql gibbering madness mongodb
19. database == database collection == table document == row
20. demo
21. more stuff flexible query language multiple indicies decent admin features
22. demo
23. no joins
24. denormalization vs in-app joins
25. demo
26. weaknesses no transactions tooling is poor collection-level locking ops can be hard
27. why? dip in to nosql working with json unstructured data large community working set in memory == SUPER FAST
28. why? PROTOTYPING
29. resources • http://www.slideshare.net/arangodb/introduction-to- column-oriented-databases • http://en.wikipedia.org/wiki/NoSQL • http://littleriakbook.com/ • http://vimeo.com/95066828 • http://aphyr.com/tags/jepsen • http://en.wikipedia.org/wiki/CAP_theorem • http://codahale.com/you-cant-sacrifice-partition- tolerance/ • http://docs.mongodb.org/manual/