Presentation is loading. Please wait.

Presentation is loading. Please wait.

relational thoughts on NoSql

Similar presentations


Presentation on theme: "relational thoughts on NoSql"— Presentation transcript:

1 relational thoughts on NoSql
change image relational thoughts on NoSql Eric Peterson

2 why you're here curious about NoSql hype justified?
pros/cons of SQL vs NoSql document db don’t know what to ask

3 what is nosql an unstructured data repository where you can store any json you want and retrieve it by id created by developers who didn't want the database to get in the way of storing data

4 SQL VS NOSQL what sql vs nosql why dis/advantages types
document db sql vs nosql dis/advantages decision factors demo

5 what is nosql? no relational structure enforced no sql language *
“database” (junkyard) still evolving analogy: junkyard - store whole cars or store parts by type

6 why nosql? scale number of users amount of data commodity hardware
no animation

7 Types table storage (key/value) (redis)
column store (hbase / hdinsight) graph databases (grid db) document db (mongo / Document DB)

8 document db Mongo DB only 1 rule; must have _id database
collections ..“tables”? documents ..“rows”? find = select

9 document db lightweight - web / mobile highly scalable
data stored in json format JavaScript Object Notation

10 json definition commas and semicolons delimited strings
{} denotes an object set of key/value pairs can be accessed with dot notation [] denotes an array useful for iterating

11 json example 1 example: simple object {"document":"file2.csv"}

12 json example 2 example: array of objects [{"document":"file2.csv"}
,{"document":"file2.html"}]

13 json example 3 example: compound document { "document": "file2.csv",
"bytes": 2479, "characters:"2479", "tags": ["original", "audited"] }

14 sql vs nosql SQL NoSql business logic in engine none language TSQL
Javascript data structure defined up-front not defined / fluid data types many few joins in code or none scalability vertical horizontal animate to bring in one row at a time MapReduce Aggregation Pipeline

15 sql advantages inherent transaction support corporate support (MS)
one-stop shop (MS) excellent documentation and examples

16 sql advantages programming talent (yeah!) business logic in db
data consistency

17 sql disadvantages Windows-only * optimized for MS products
painful structure changes

18 nosql advantages elastic scaling (sharding) big data no licensing fees
no DBAs (huh?)

19 nosql advantages flexible data models faster iteration
data stored in front-end format DML familiar to developers

20 nosql disadvantages more engineering effort what works with what?
options overwhelming immature product support & documentation no data validation

21 sql vs nosql no animation

22 sql vs nosql no animation

23 MS Document DB SQL MS Doc. DB NoSql business logic in engine none
language TSQL TSQL/javascript Javascript data structure defined up-front not defined / fluid data types many few joins in code or none scalability vertical horizontal animate to bring in one row at a time MapReduce Aggregation Pipeline

24 decision factors corporate strategy familiarity transactions
scalability support

25 demo create customer collection & document find all records
find one record find complex record Joins (manual, BBRef, $ref, $id, $db) Quick ref?

26 more info http://bit.ly/psrelvsnon http://bit.ly/relvsnon

27 thanks! contact info Eric Peterson


Download ppt "relational thoughts on NoSql"

Similar presentations


Ads by Google