NoSQL Database- cassandra column Base DB

1. +
NoSQL – Part 2
CAP Theorem & Column Oriented
Mohammad Sadegh Salehi
Dr.Baraani
Winter2015 Sheikh Bahaie
University

2. +
NoSQL (part 2) - CAP Theorem & Column Oriented
33
2
Winter 2015
Agenda
—Review NoSQL
—Dynamo and BigTable
—NoSQL Classification
—Key-value Stores
—Column Oriented
—Casandra
—Why Casandra
—Question

3. +
NoSQL (part 2) - CAP Theorem & Column Oriented
33
3
Winter 2015
What is NoSQL
review
 Stands for Not Only SQL
 Class of non-relational data storage systems
 Usually do not require a fixed table schema nor do
they use the concept of joins
 All NoSQL offerings relax one or more of the ACID
properties (will talk about the CAP theorem)

4. +
NoSQL (part 2) - CAP Theorem & Column Oriented
33
4
Winter 2015
Dynamo and BigTable
 Three major papers were the seeds of the NoSQL
movement
• BigTable (Google)
• Dynamo (Amazon)
—Gossip protocol (discovery and error detection)
—Distributed key-value data store
—Eventual consistency
• CAP Theorem (discuss in a sec ..)

5. +
NoSQL (part 2) - CAP Theorem & Column Oriented
33
5
Winter 2015

6. +
NoSQL (part 2) - CAP Theorem & Column Oriented
33
7
Winter 2015
What kinds of NoSQL
Review
 NoSQL solutions fall into two major areas:
• Key/Value or ‘the big hash table’.
—Amazon S3 (Dynamo)
—Voldemort
—Scalaris
• Schema-less which comes in multiple flavors, column-
based, document-based or graph-based.
—Cassandra (column-based)
—CouchDB (document-based)
—Neo4J (graph-based)
—HBase (column-based)

7. +
NoSQL (part 2) - CAP Theorem & Column Oriented
33
8
Winter 2015
Key-Value Stores
 Extremely simple interface
• Data model: (key, value) pairs
• Operations:
—Insert(key,value),
—Fetch(key),
—Update(key),
—Delete(key).
 Implementation: efficiency, scalability, fault-
tolerance
• Records distributed to nodes based on key
• Replication
• Single-record transactions,“eventual consistency”

8. +
NoSQL (part 2) - CAP Theorem & Column Oriented
33
9
Winter 2015
Key-Value Data Stores
 Storing Session Information
 User Profiles, Preferences: Almost every user has
a unique userID as well as preferences such as
language, color, timezone, which products the
user has access to , and so on.
Suitable Use Cases

9. +
NoSQL (part 2) - CAP Theorem & Column Oriented
33
10
Winter 2015
Key-Value Data Stores
 As we want the shopping carts to be available
all the time, across browsers, machines, and
sessions, all the shopping information can be put
into value where the key is the userID
Shopping Cart Data

10. +
NoSQL (part 2) - CAP Theorem & Column Oriented
33
11
Winter 2015
Key-Value Data Stores
 Relationships among data
 Multi-operation Transactions
 Query by Data
 Operations by Sets
Not to Use

11. +
NoSQL (part 2) - CAP Theorem & Column Oriented
33
12
Winter 2015
Column-oriented
 Store data in column order
 Allow key-value pairs to be stored (and retrieved
on key) in a massively parallel system,
• Data model: families of attributes defined in a schema,
new attributes can be added,
• Storing principle: big hashed distributed tables,
• Properties: partitioning (horizontally and/or vertically),
high availability etc. completely transparent to
application,
Intro

12. +
NoSQL (part 2) - CAP Theorem & Column Oriented
33
13
Winter 2015

13. +
NoSQL (part 2) - CAP Theorem & Column Oriented
33
14
Winter 2015
Cassandra
 Apache Cassandra™ is a free
 Distributed…
 High performance…
 Extremely scalable…
 Fault tolerant (i.e. no single point of failure)…
 Post-relational database solution.
 Cassandra can serve as both real-time datastore and as a
read-intensive database.
 Compiles to: C++, Java, PHP, Ruby, Erlang, Perl, ...
Thrift

14. +
NoSQL (part 2) - CAP Theorem & Column Oriented
33
15
Winter 2015
Cassandra
Infographic

15. +
NoSQL (part 2) - CAP Theorem & Column Oriented
33
16
Winter 2015
Cassandra
 Originally developed at Facebook
 Follows the BigTable data model: column-oriented
 Uses the Dynamo Eventual Consistency model
 Written in Java
 Open-sourced and exists within the Apache family
 Uses Apache Thrift as it’s API
 Some of its myriad users:

16. +
NoSQL (part 2) - CAP Theorem & Column Oriented
33
17
Winter 2015
Cassandra
 keyspace: Usually the name of the application; e.g.,
'Twitter', 'Wordpress‘.
 column family: structure containing an unlimited
number of rows
• Simple
• Super (nested Column Families)
 column: a tuple with name, value and time stamp
• Each Column has
— Name
— Value
— Timestamp
 key: name of record
 super column: contains more columns
Data Model

17. +
NoSQL (part 2) - CAP Theorem & Column Oriented
33
18
Winter 2015
Cassandra – Data Model
keyspace
settings
column family
settings
column
name value timestamp

18. +
NoSQL (part 2) - CAP Theorem & Column Oriented
33
19
Winter 2015
Cassandra
Column Family & Super Column Family

19. +
NoSQL (part 2) - CAP Theorem & Column Oriented
33
20
Winter 2015
Cassandra
 Cassandra was designed with the understanding that
system/hardware failures can and do occur
 Peer-to-peer, distributed system
 All nodes the same
 Data partitioned among all nodes
in the cluster
 Custom data replication to ensure
fault tolerance
 Read/Write-anywhere design
Architecture Overview

20. +
NoSQL (part 2) - CAP Theorem & Column Oriented
33
21
Winter 2015
Cassandra
 Each node communicates with each other through the
Gossip protocol, which exchanges information across
the cluster every second,
 A commit log is used on each node to capture write
activity. Data durability is assured,
 Data also written to an in-memory
structure (memtable) and then to
disk once the memory structure is
full (an SStable).
Architecture Overview

21. +
NoSQL (part 2) - CAP Theorem & Column Oriented
33
22
Winter 2015
Why Cassandra?
 Gigabyte to Petabyte scalability
 Linear performance gains through adding nodes
 No single point of failure
 Easy replication / data distribution
 Multi-data center and Cloud capable
 No need for separate caching layer
 Tunable data consistency
 Flexible schema design
 Data Compression
 CQL language (like SQL)
 Support for key languages and platforms
 No need for special hardware or software

22. +
NoSQL (part 2) - CAP Theorem & Column Oriented
33
23
Winter 2015
Why Cassandra?
 Capable of comfortably scaling to petabytes
 New nodes = Linear performance increases
 Add new nodes online
Big Data Scalability
1
2
Double Throughput
Capabilities
1
2
3
4

23. +
NoSQL (part 2) - CAP Theorem & Column Oriented
33
24
Winter 2015
Why Cassandra?
 All nodes the same
 Customized replication affords tunable data redundancy
 Read/write from any node
 Can replicate data among different physical data center
racks
No Single Point of Failure

24. +
NoSQL (part 2) - CAP Theorem & Column Oriented
33
25
Winter 2015
Why Cassandra?
 Peer-to-peer architecture removes need for special
caching layer and the programming that goes with it
 The database cluster uses the memory from all
participating nodes to cache the data assigned to each
node
 No irregularities between a memory cache and database
are encountered
No Need for Caching Software
Database Server
Memcached Servers
Application Servers
Writes
Reads

25. +
NoSQL (part 2) - CAP Theorem & Column Oriented
33
26
Winter 2015
Why Cassandra?
 Uses Google’s Snappy data compression algorithm
 Compresses data on a per column family level
 Internal tests at DataStax show up to 80%+ compression
of raw data
 No performance penalty (and some increases in overall
performance due to less physical I/O)!
Data Compression
Portfolio Keyspace
Customer Column Family

26. +
NoSQL (part 2) - CAP Theorem & Column Oriented
33
27
Winter 2015
Why Cassandra?
 Very similar to RDBMS SQL syntax
 Create objects via DDL (e.g. CREATE…)
 Core DML commands supported: INSERT, UPDATE,
DELETE
 Query data with SELECT
CQL Language
Portfolio Keyspace
1
2
3
4
5
6
SELECT *
FROM USERS
WHERE STATE = ‘TX’;

27. +
NoSQL (part 2) - CAP Theorem & Column Oriented
33
28
Winter 2015
Comparison with MySQL
 MySQL > 50 GB Data
Writes Average : ~300 ms
Reads Average : ~350 ms
 Stats provided by Authors using facebook data.
 Cassandra > 50 GB Data
Writes Average : 0.12 ms
Reads Average : 15 ms

28. +
NoSQL (part 2) - CAP Theorem & Column Oriented
33
29
Winter 2015
Cassandra Tools
........DesktopnoSqlCassandra-sadeghnoSqlCassandra-sadegh.mp4

29. +
NoSQL (part 2) - CAP Theorem & Column Oriented
33
30
Winter 2015
Where to get Cassandra?
 Go to www.datastax.com
 DataStax makes free smart start installers available for
Cassandra that include:
• The most up-to-date Cassandra version that is production quality
• A version of DataStax OpsCenter, which is a visual, browser-
based management tool for managing and monitoring
Cassandra
• Drivers and connectors for popular development languages
• Same database and application
• Automatic configuration assistance for ensuring optimal
performance and setup for either stand-alone or cluster
implementations
• Getting Started Guide

30. +
NoSQL (part 2) - CAP Theorem & Column Oriented
33
31
Winter 2015
Where Can I Learn More?
www.datastax.com
 Free Online Documentation
 User/Customer Cas Studies
 Technical White Papers
 Software downloads
 Technical Articles
 User Forums
 Videos
 Tutorials
 FAQ’s
 Blogs

31. +
NoSQL (part 2) - CAP Theorem & Column Oriented
33
32
Winter 2015
Resources
Sites
 Cassandra
• http://cassandra.apache.org
 NoSQL News websites
• http://nosql.mypopescu.com
• http://www.nosqldatabases.com
 “a practical guide to noSQL”, Posted by Denise Miura on
March 17, 2011 at
• http://blogs.marklogic.com/2011/03/17/a-practical-
guide-to-nosql/

32. +
NoSQL (part 2) - CAP Theorem & Column Oriented
33
33
Winter 2015
Resources
Books
 “Cassandra The Definition Guide”, O'Reilly Media, nov2013
 “Cassandra Essential Toturial”, DataStax 2014
 “Professional NoSQL”, Wrox, 2011
 “NoSQL Distilled”, Martin Fowler, 2013

33. +
NoSQL (part 2) - CAP Theorem & Column Oriented
33
34
Winter 2015
Questions

34. +
Mohammad Sadegh Salehi
3adegh.ce@gmail.com
Thank You