databases in the cloud

Post on 18-Jan-2015

487 Views

Category:

Technology

1 Downloads

Preview:

Click to see full reader

DESCRIPTION

Quick overview of using master-slave replication and recap of NoSQL databases

TRANSCRIPT

Databases in the cloud

Master-Slave (RDBMS)

Replication

Master Slave

Client Client

Read/Write Read Only

OLTP, Real time interaction

Reporting, Adhoc, near real time

interaction

Master-Slave (RDBMS)

Replication

Master Slave

Client

Client

Read/Write

Read Only

OLTP, Real time interaction

Data Warehouse

ClientRead Only

Adhoc, Operations,

etc.

Master-Slave & NoSQL

Replication

Master Slave

Client

Client

Read/Write

Read Only

OLTP, Real time interaction

Data Warehouse

ClientRead Only

Adhoc, Operations,

etc.

NoSQL Flavors – Key Value Stores

For HUGE amounts of data, pointer references flattened rows

450-23

729-22

4AE-56

Cust23, Order123, 6-1-2013, 10:23:15, $56.75, store5, etc.

Cust45, Order323, 6-1-2013, 10:25:25, $36.75, store3, etc.

Cust11, Order221, 6-1-2013, 10:23:35, $50.03, store2, etc.

Key Value

Key Value Stores – Use Cases

• Massive amounts of writes (e.g. Tweets, distributed data transmission)

• Reads of large, static, structured historical data (orders, events, transactions, etc.)

Speed, ScalableNo Schema, Bad @ Complex data

NoSQL Flavors – Column Store

For HUGE amounts of data with relational database characteristics

450-23

729-22

4AE-56

Key

Customer23, Joe,34221

Order6-1-13, $56.75

LocationStore 5, Chain 2

Columns

Customer45, Jill, 34211

Order6-1-13, $36.75

LocationStore 3, Chain 2

Customer11, Sue 34441

Order6-1-13, $50.03

LocationStore 2, Chain 2

Column Store – Use Cases

• Massive amounts of data with dynamic columns

• Integrating data feeds from disparate sources

Speed, Scalableeasy to add columns

Poor at interconnected data

NoSQL Flavors – Document Store

For storing unstructured data where the values are contained in documents

450-23

Key

Documents

Document Store – Use Cases

• Centralized log file processing (many different formats)

• Metadata and asset management

• Product management (Catalog, Inventory Management, hierarchy)

Scalable,Flexible data model

Poor at interconnected data

NoSQL Flavors – Graph Database

For storing and managing interconnected relationships

http://saritsblog.blogspot.com/2011/09/graph-databases-hadoop.htmlSource:

Graph Database– Use Cases

• Social media analysis

• Visual representation of relationships

Source: Mashable.com

Great at graphing,Fast, scalable

Must traverse entire tree

top related