Download as KEY, PPTX
Uploaded bySean Collins
Database Throwdown Introduction
This document provides an overview of key database terminology including ACID properties, CAP theorem, and common data types. ACID properties (atomicity, consistency, isolation, durability) describe promises that databases make regarding transactions. The CAP theorem states that a distributed system can only guarantee two of three properties: consistency, availability, or partition tolerance. Common database types include relational, key-value, column-oriented, and document-oriented.
Database Throwdown Introduction
- 1. Database Terminology in15 minutes or your money back!
- 2.
- 3. Currently Works With.... 9AM - 5PM (SQL) Oracle 11gR2 5-???? (NoSQL) Redis SimpleDB
- 4. Concepts ACID CAP All of theseterms describe the promises that a database system will abide by. A contract for your data
- 5.
- 6. Atomicity All operations willoccur, or none will occur Simple example: INSERT INTO Names (firstname,lastname) VALUES (Sean, Collins)
- 7. Consistency All operations willleave the database in a known good state Keys, Constraints, triggers, etc.. Cannot have operations that violate the rules for your data Customer doesn’t exist Primary key that already exists
- 8. Isolation Operations cannot interferewith one- another Handle multiple, concurrent operations (One write, many reads) in a defined manner.
- 9. Durability Completed operations arebullet-proof Yank out the power cable Crash the database server process When the database server comes back up - database is in a consistent state - with your data intact.
- 10. CAP CAP Theorem -Eric Brewer (Conjecture) Seth Gilbert and Nancy Lynch (Theorem) Consistency Availability Partition Tolerance
- 11. Consistency All nodes ina distributed system see the same data - at that exact moment Simple example: Update item B on node #1, query node #2 about B - get back the updated data
- 12. Availability System will processrequests, despite failures in individual nodes Not a guarantee that the operation will succeed Just a guarantee that you will get a response back No guarantee is made for WHEN you will get a response back
- 13. Partition Tolerance System willcontinue to operate, even when arbitrarily many messages between nodes are lost
- 14. PICK TWO CAP Theoremasserts that of the three properties, a system can only have two.
- 15. Data Types Relational Database Tables Columns Rows
- 16. Common NoSQL Datatypes Key/Value Column Store Document etc....
- 17. ACID vs. BASE Someoneclever... har har “Basic Availability” “Soft State” “Eventually Consistent” Much prefer CAP theorem - exposes the tradeoffs that you have to choose between