Azure Database Options - NoSql vs Sql

 The name “NoSQL” was in fact first used by Carlo Strozzi in 1998 as the name of file-
based database he was developing. It was a relational database.
 Neal Ford coined the term Polyglot Programming in 2006
 SABRE, a database still used in the airline industry, predates relational databases
(however, they began using a relational database is 2001)
 Ingres and System R were two early relational database prototypes in the 70s
 The term "relational database" was invented by E. F. Codd at IBM in 1970
 Some claim that “NoSQL” now means “Not Only SQL”, and that it isn’t anti-relational
 Rackspace used the term “NoSQL” for a conference in 2009.
 Terms: “Relational Winter” gives way to “Database Thaw” (ok, not a fact, but still fun)

Which Azure Database
Option Should I Choose? Anne Bougie
Senior Software Developer
Concurrency, Inc.
Twitter: @bougiefever
Understanding Database Options

Azure SQL
DocumentDb (Cosmos)
Azure Table Service (Cosmos)
Azure Redis Cache
Hbase
Gremlin (Cosmos)
SQL
NoSQL

SQL NoSQLRelational
Persistent
Transactional
Schema
Data Integrity
Schema-less
Clusters
Scalability
Fast Data Access
Partitions
Concurrency
Sharding

Querying Data
Transactions
Scalability
Configuration/Management
Schema
Speed
SQL NoSQL

Order
Player: Mojo Jojo
In-App Purchase
Order Id: 1234
Financial
Credit Card: 4012888888881881
Expiration: 5/2020
1 Gems 250 $5
2 Potions 10 $10
Order
Player Assets
Line Items
Credit Card
The object-relational impedance
mismatch is a set of conceptual and
technical difficulties that are often
encountered when a relational database
management system (RDBMS) is being
served by an application program (or
multiple application programs) written in
an object-oriented programming
language or style, particularly because
objects or class definitions must
be mapped to database tables defined by
relational schema.

Order
Player: Mojo Jojo
In-App Purchase
Order Id: 1234
Financial
Credit Card: 4012888888881881
Expiration: 5/2020
1 Gems 250 $5
2 Potions 10 $10
Order
Player Assets
Line Items
Credit Card

Order
Player: Mojo Jojo
In-App Purchase
Order Id: 1234
Financial
Credit Card: 4012888888881881
Expiration: 5/2020
1 Gems 250 $5
2 Potions 10 $10
{
orderId: 1234,
player: ‘Mojo Jojo’,
inAppPurchase {
date: 1/1/2017,
items: [
{
objectType: ‘Gem’,
quantity: 250,
priceEach: 0.02
},
{
objectType: ‘Potion’,
quantity: 10,
priceEach: 1.00
}
]
}
…

Document
Column Family
Graph
Key-Value
DocumentDb Azure Table Service
Azure Redis Cache
Azure Hbase
NoSQL
Gremlin

 Fast reads & writes
 Flat data structure
 Schema-less
 Partition key
 Row key
 Time stamp
 Service will scale out using partition
key

 In memory key-value store
 Very fast reads (faster than table
storage),
 Used as a database, cache and
message broker
 Transactions
 Expiration of items

 Stores data in json
 Schema-less
 Stores complex, hierarchical data
 Highly scalable

 Part of the Hadoop eco system
 Commonly augmented with Hive
 Can handle very large amounts of writes in a short period of time

 Nodes and relationships
 Data with many complex relationships
 Typically used to augment the system of record

Document
Column
Family
Key Value
Graph

Consistency
Availability
Partition
Tolerance
SQL
CosmosDB
CosmosDB
Hbase
CosmosDB
Azure Redis Cache
Never Gonna Get It

 What are the relationships
 How much data, and how fast is it coming in
 How the data will be accessed
 Data access performance requirements
 Consistency/Transactional requirements
 Entity complexity
 Programmer skill level

 Lots of data
 Simple data
structure
 Quickly perform
small read and
write operations
 Inexpensive,
fairly simple
 Need to add data
items willy nilly
 Lots of data
 Need high
performance
 More complex
data structure
 Need to add data
items willy nilly
 Lots of data
 Insanely huge
amounts of data
 Need high
performance
 No joins
 Lots of data
 Lots of
connections
between entities
 Quickly changing
relationships
between entities
Key Value Column Family Document Graph

 Azure Redis Cache
 Leaderboards, Shopping Carts
 Latest x items of anything
 Deletes and filters
 Cache
 Azure Table Storage
 Large amounts of data with a flat structure
 Fast querying using the partition and row keys
 DocumentDb
 Product catalogs, gaming, social networking
 Hbase
 Voting, Race, anything with huge amounts of data being generated in huge bursts, telemetry
 Gremlin
 Social networking relationships, anything with complex and changing relationships between
entities

Problem
 Lots of data
 Fairly simple data structure
 Lots of small reads and writes
 Need high performance
 Need high availability
 Need fast searching on columns other
than the key
Solution
 Use Azure Storage
 Augment with Redis Cache for
searching

Problem
 Really need high consistency
 Highly structured data
 Current data is not really large
 Historical data is huge
 And we need to report on historical
data
Solution
 Use Azure Sql
 Archive to DocumentDb

Just cause it’s old, doesn’t mean it’s
not cool anymore

 Slides on GitHub https://github.com/Bougiefever/AzureNoSqlDataPrimer
 A Newbie Guide to Databases https://blog.appdynamics.com/engineering/a-newbie-
guide-to-databases/
 That NoSQL Thing: Column (Family) Databases https://ayende.com/blog/4500/that-
no-sql-thing-column-family-databases
 NoSQL Distilled: A Brief Guide to the Emerging World of Polyglot Persistence
http://a.co/1YwtJ47
 On Sharding Graph Databases http://jimwebber.org/2011/02/on-sharding-graph-
databases/
 Azure Cosmos DB Documentation https://docs.microsoft.com/en-us/azure/cosmos-db/
Anne Bougie
anne.bougie@gmail.com
@bougiefever
http://www.bougiefever.com
https://github.com/Bougiefever

Azure Database Options - NoSql vs Sql

More Related Content

What's hot (20)

Similar to Azure Database Options - NoSql vs Sql (20)

Recently uploaded (20)

Azure Database Options - NoSql vs Sql

Editor's Notes