NoSQL, Growing up at Oracle

NoSQL, growing up at Oracle
Robert Greene, product management

Webcast Agenda
 Modern data management workload
 Lessons learned and applied
 System architecture is key
 Feature convergence on essentials
 A lasting NoSQL technology

modern data management workload
1. Write Intensive
2. Concurrent
3. Available

Modern Workloads
Amazon constructs each custom user web page by sending requests to 150 different
services and data stores. Over 500 individual key-value lookups per page server.
Write & Read intensive

Modern workloads
Concurrent

Modern Workloads
 Built-in, always on
Available

system architecture is key
1. Linear scaling
2. Replicated
3. Asynchronous
4. Data local

Architecture
 Elastic Auto Sharding
(split, add, contract)
Store
Shard
M
Shard
M
R
R R
Application
NoSQL Driver
R
 Writes to elected
node on ACID
 Reads from any
node in system
Shard
M
R
R
Expand and Rebalance
Shard
M
R
R
Linear scaling & replicated

Architecture
Asynchronous

Architecture
Data local
Pkey A Value A
F1 F2 F3 V1 V2 V3
Pkey B Value B
F1 F2 F3 V1 V2 V3
Skey A Pkey A
V2 V3 F1 F2 F3
Skey B Pkey B
V2 V3 F1 F2 F3
Skey A Pkey A
V1 F1 F2 F3
Skey B Pkey B
V1 F1 F2 F3
Primary Records
Secondary Indexes
Shard 1
Writes
Reads
Shard 2
Shard N
Index on (V1)
Index on (V2, V3)
Value V1
V1a V2b V3c

lessons learned and applied
1. Non-web retail
2. Integrated
3. Use cases unknown

Lessons learned
Non-Web Retail

Lessons learned
Authorization
DW Rules
Publisher
Rule
Manager
Customer
Rules
Fraud
Alerts
Rules Storage/Retrieval
RiskService
Ext Service Ext Streams
Integrated

Lessons learned
 New application: Retrieve all data for a TV
series (seasons, episodes, actors, and so on)
 Perfect for a NoSQL database
– One simple retrieval returns all information
– Much simpler than relational database; no joins
Future use cases unknown
 And then, a new feature request: Retrieve all data
for a given actor
 Incompatible with original NoSQL document model
 ‘Actors’ cannot be retrieved separately
 Simple feature to add with relational database
 Just another SQL query on the same schema

feature convergence on essentials
1. Transactions
2. Standards : SQL, Indexes
3. Secure : Auth, Roles, Audit
4. HA : Data Centers

Feature Convergence
 Ease of use
– The 5% case
– Development simplification
– Process improvement
– Prevent index divergence
– Data consistency
Transactions

Feature Convergence
 Ease of use
– Data Modeling
– Integration
– SQL Query
 Late Binding (soft-schema)
 Indexes
 Easy JSON representation
Tables

Feature Convergence
Secure & Highly Available

A lasting NoSQL technology
1. Simple key : value access
2. Concurrent capable
3. Integrated
4. Manageable

advanced key-value database
1. Don’t care what the value is
Tables: Primary key – Shard key
Data locality and indexing
BASE & Transactions
2. JSON API
2. JSON API
3. Table API
Oracle NoSQL Database

Join NoSQL Database Community
Twitter
https://twitter.com/#!/OracleNoSQL
LinkedIn
http://www.linkedin.com/groups?gid=4147754
Oracle’s NoSQL DB blog
https://blogs.oracle.com/nosql
Oracle Technology Network
http://bit.ly/1f0d8wU
Developer Webcast Series
http://bit.ly/1doV2jl
Oracle.com/BigData

NoSQL, Growing up at Oracle

More Related Content

What's hot

Viewers also liked

Similar to NoSQL, Growing up at Oracle

More from DATAVERSITY

Recently uploaded

NoSQL, Growing up at Oracle