Apache Cassandra introduction

Cloud
Computing
Apache Cassandra
Omid mirabdolazimi
Fardin jamshidi
Professor : Dr Sadegh dorri nogoorani
1397/10/09

• Open source distributed database management
system for handling huge amounts of data across
many commodity systems
• Cassandra is a “NoSQL” or “Non-Relational”
database and can be described as a mix between a
“Key-value Store” and a “Column-Orientated”
database
What Is Apache Cassandra?

Where Did Cassandra Come From?
• Cassandra was initially created at Facebook
• Combination of Google Big Table and Amazon Dynamo
• It was created to power the “Inbox Search” feature
• Cassandra was released as open source in July of 2008
• It became an Apache Incubator project in February of
2009 and It became a full level project a year after that

Cassandra Architecture
• Built with the understanding that hardware & software
failures can happen
• Peer to Peer Architecture
• All nodes are the same
• Read/Write Anywhere
• Gossip Protocol
• Commit Log Captures All Activity
• Well suited for cloud deployments

Features of Cassandra
• Decentralized – No master & no single point of failure. Data is
distributed across the cluster
• Replication – Tailored for multiple-data center deployment
• Scalability – New machines can easily be added with no
downtime or interruption
• Fault Tolerance – Failed nodes can be replaced with no
downtime
• Cassandra Query Language (CQL) – An SQL-like alternative

More Advantages
• Always On Architecture – Continuous availability
with no downtime
• Faster linear-scale performance
• Operational Simplicity – Administration is
simplified
• Transaction Support
• No new equipment required - Very economical

CQL (Cassandra Query Language)
• CQL is very similar to SQL (Structured Query
Language) in terms of syntax and commands
• Statements directly change data and/or change the
way data is stored
• All statements end with a semi-colon
SELECT * FROM sampletable;

Who Uses Cassandra?
• Facebook
• WalmartLabs
• Constant Contact
• Digg
• AppScale
• Netflix
• Twitter
• Zoho
• IBM
• FormSpring
• Cisco WebEx
• Rackspace
• OpenX
• Adobe
• Comcast
• eBay

Apache Cassandra
Architecture & Data Model
Cloud
Computing

The Data Model
• Cassandra is sort of in its own data model class but
can be described as a hybrid of a “Key-value Store”
and a “Column-Orientated” database.
• Cassandra was modeled after Google’s “Big Table”
and Amazon’s “Dynamo”
• Cassandra even has similarities to a “Relational
Database” , but much more flexible.

Keyspaces
• A Keyspace is a “container” for your data
• Similar to a Database in an RDBMS
• Used to group Column Families together
• Typically, a Cluster has one Keyspace per application
• Replication is controlled on a per-keyspace basis

Column Family
• A “Column Family” is similar to a “Table” in a RDBMS
because it has columns and rows
• Relational database tables use a predefined, fixed
schema. Column families do not which makes them
very flexible
• Cassandra’s data model promotes “Denormalization”
which is the complete opposite of the relational
database

Replication Strategies
• Simple Strategy – Use this for a single data center. It
places the first replica on a node determined by the
partitioner. Does not consider topology.
• NetworkTopologyStrategy - If you plan to have your
cluster span across multiple data centers. Specifies
how many replicas you want in each data center.

Apache Cassandra
Relational Database Comparison
Cloud
Computing

Relational Databases
Relational databases have been around for 40+ years and
will always have a place in technology. There are many
situations where a relational database is the best choice
NoSQL databases are NOT going to replace relational
databases in all areas

Advantages of NoSQL
• Handles Big Data Much Better Than RDBMS
• Cheaper Hardware & Software
• Easier Scaling
• Much More Flexible (Relatively Schema-Free)
• Map and Reduce Capability
• Less Management

Advantages of RDBMS
• Better For Complex Data
• Better Support
• Better For “Relational” or “Object
Orientated” Data
• Been Around Longer
• Better Data Analytics

Scale-Up vs. Scale-Out
• Scale-Up Architecture (Relational) – Storage is confined to a
single form-factor which needs more resources to scale (CPU,
Memory, etc) to a single node.
• Scale-Out Architecture (NoSQL) - The total amount of disk
space can be easily expanded as needed. When a storage
array reaches it’s max, another will pick it up where the last
left off. This makes scaling significantly easier and less
expensive.

Things To Consider
When choosing a system/database for your projects or company,
there are many things to factor in…
Complexity
Type of Data
Budget
Amount of Data
Traffic
Programming Languages

Apache Cassandra
How It Works
Cloud
Computing

The “Write” Process
• Logs data to the commit log
• Writes data to the memtable
• Flushes data from the memtable
• Stores data on disk in SSTables
• Compaction

Deleting Data
• Cassandra does not delete as SSTables are
immutable
• Cassandra marks the data with a
“Tombstone” which is a marker in a row that
indicates that a column was deleted

The “Read” Process
• Reading data is done in parallel across all
nodes in a cluster
• If the node with the requested data is down
then the data will be read from the node
which holds a replica of the data

Apache Cassandra
Cassandra Software
Cloud
Computing

Compatible Operating Systems
• GNU/Linux
• Microsoft Windows
• Mac OSX

Prerequisites & Requirements
• Requires the most stable version of Java 7
• 4GB+ memory is recommended for
production environment

Cassandra Server
• Apache Cassandra Core Server
• Nodetool Admin Command Line Interface
• CQLSH and Cassandra-cli Development Shell
Apache provides binary tarballs and Debian packages -
http://cassandra.apache.org/download/

DataStax Community Distribution
DataStax Community Edition is a free software package that offers…
• Apache Cassandra Core Server
• Nodetool Admin Command Line Interface
• CQLSH and Cassandra-cli Development Shell
• Windows Installer & Mac OS X Binary
• OpsCenter Community Version
• Sample Database & Application
• CQL Utility

DataStax Enterprise Distribution
DataStax Enterprise Edition is a premium software package that offers…
• All that Community Offers
• OpsCenter Enterprise Version
• Apache Hadoop (With MapReduce, Hive & Pig)
• Enterprise Search
• Premium Support
• Advanced Security Features
• Workload Management Benefits

OpsCenter
• Simplified Data Management
• Easy To Use Visual Interface
• Centralized Dashboard
• Easy Installation
• Real-Time Analytics
• Multiple Data Center Support
• Automated Management
• Rebalance & Repair Clusters
• Alerts & Notifications

Available Client Drivers
There are client drivers available for the following languages…
Go
Node.js
Clojure
C++
Java
Python
Ruby
C# / .NET

Apache Cassandra
CQL Overview
Cloud
Computing

Defining Keyspaces in CQL
CREATE KEYSPACE people
WITH REPLICATION = { 'class' : 'SimpleStrategy',
'replication_factor' : 3 };
CREATE KEYSPACE people
WITH REPLICATION = {'class' : 'NetworkTopologyStrategy',
'dc1' : 3, 'dc2' : 2};

Compound Keys
A compound primary key includes the partition key, which
determines on which node data is stored, and one or more
additional columns that determine clustering.
You need to know which fields you want to be able to sort and
order by before you create the data model.

Creating Compound Keys
To create a compound primary key, use the keywords, PRIMARY KEY,
followed by the comma-separated list of column names enclosed in
parentheses.
CREATE TABLE emp (
empID int,
deptID int,
first_name varchar,
last_name varchar,
PRIMARY KEY (empID, deptID)
);

UUID’s
Universal Unique ID is a field type that is used to avoid
collisions in column names.
32 hex digits, 0-9 or a-f, which are case-insensitive, separated by dashes, -,
after the 8th, 12th, 16th, and 20th digits. For example: 01234567-0123-
0123-0123-0123456789ab

TIMEUUID’s
Time Universal Unique ID is a field type that is used to avoid
collisions in column names.
timeuuidUses the time in 100 nanosecond intervals since 00:00:00.00 UTC
(60 bits), a clock sequence number for prevention of duplicates (14 bits),
plus the IEEE 801 MAC address (48 bits) to generate a unique identifier.
For example: d2177dd0-eaa2-11de-a572-001b779c76e3
Some functions that you can use to insert timeUUIDs are “now()”,
“dateOf()” and “minTimeuuid()/maxTimeuuid”

Selecting Data
SELECT * FROM users;
SELECT COUNT(*) FROM users;
SELECT * FROM users LIMIT 25;
SELECT * FROM users WHERE city = ‘Boston’ (city must be defined as a primary
key or index)
CREATE TABLE users (
id uuid,
first_name varchar,
last_name varchar,
age int,
city varchar,
PRIMARY KEY (id, city, age)
);

Inserting Data
INSERT INTO users (id, first_name, last_name, age, city)
VALUES (now(), ‘John’, ‘Doe’, ‘33’, ‘Seattle’)
INSERT INTO users (id, first_name, last_name, emails)
VALUES('frodo', 'Frodo', 'Baggins', {'f@baggins.com',
'baggins@gmail.com'});
NOTES:
• If column exists, it is updated
• You can qualify table names by keyspace

Updating Data
UPDATE users SET age = 34 WHERE id = cfd66ccc-d857-
4e90-b1e5-df98a3d40cd6
UPDATE users SET age = 34, city = ‘Portland’ WHERE id
= cfd66ccc-d857-4e90-b1e5-df98a3d40cd6
UPDATE users
SET todo =
{ '2012-9-24' : 'enter mordor',
'2012-10-2 12:00' : 'throw ring into mount doom' }
WHERE user_id = 'frodo';

Deleting Data
DELETE email, phone
FROM users
WHERE user_name = 'jsmith';
DELETE todo ['2012-9-24'] FROM users WHERE id =
'frodo';

Altering Tables
# Change Type
ALTER TABLE users ALTER age TYPE int
# Add Column
ALTER TABLE users ADD state varchar;
# Drop Column
ALTER TABLE users DROP city

Apache Cassandra introduction

Recommended

More Related Content

What's hot (20)

Similar to Apache Cassandra introduction (20)

Recently uploaded (20)

Apache Cassandra introduction