Oracle Autonomous Database

Cristian Necula

#thinkautonomous
Copyright © 2018, Oracle and/or its affiliates. All rights reserved. |
Safe Harbor Statement
The following is intended to outline our general product direction. It is intended for
information purposes only, and may not be incorporated into any contract. It is not a
commitment to deliver any material, code, or functionality, and should not be relied upon
in making purchasing decisions. The development, release, and timing of any features or
functionality described for Oracle’s products remains at the sole discretion of Oracle.

Copyright © 2018, Oracle and/or its affiliates. All rights reserved. | 2

“By 2018, 75% of enterprise and ISV development will include
“By 2018, 75% of enterprise and ISV development will include
Cognitive/AI or Machine
Cognitive/AI Learning
or Machine Learningfunctionality inleast
functionality in at at least
one one
application, including
application, all business
including analytics
all business analytics tools.”
tools.”
Source: IDC, “ IDC Future Scape: Worldwide Analytics, Cognitive/AI, and Big Data 2017 Predictions “, DOC # US41866016, Nov2016

“By 2020, AI technologies will be a top-five

investment priority for more than 30% of CIOs. ”
Source: Gartner, “ Applying Artificial Intelligence to Drive Business Transformation: A Gartner Trend Insight Report “ , Whit Andrews, 29 August
2017.

Copyright © 2018, Oracle and/or its affiliates. All rights reserved. |

What is Data Science?
ARTIFICIAL INTELLIGENCE
• Data science is an interdisciplinary field that A program that can sense, reason, act, and adapt
“intelligence demonstrated by machines”
uses scientific methods, processes, algorithms
and systems to extract knowledge and
insights from data. MACHINE LEARNING
Algorithms whose performance improves as
• Data scientist’s responsibilities include they are exposed to more data over time
“give computer systems ability to learn”
preparing data, exploring and visualizing data,
and building models using programming
languages like Python or R. DATA SCIENCE DEEP
LEARNING
• Data science includes methods and
Subset of ML in which multi-
techniques such as Artificial Intelligence (AI), layered neural networks
learn from vast amounts of
Machine Learning (ML), and Deep Learning. data

Copyright
Copyright
©©2018,
2017,
Oracle
Oracle
and/or
and/or
itsits
affiliates.
affiliates.
AllAll
rights
rights
reserved.
reserved.| 4
Faster, better decision making across all business processes and all industry sectors
AI is going to impact almost every industry and business process.

Finance Marketing Operations Healthcare

Sales Logistics
Consumer Goods Education
Service AI Finance

Media HR IT Transportation

Manufacturing Energy
Copyright © 2018, Oracle and/or its affiliates. All rights reserved. | 5
Automatic or Autonomous ?
Automatic Autonomous
Examples:
No need to use the steering wheel or
• Cruise control
brake.
• Emergency stopping
Simply tell the car where you are going.
Autonomous Car • Warnings for lane changes

Examples:
• Automatic storage management All features automatically implemented
• Automatic workload repository Simply tell the database your goals
Autonomous Database • SQL Plan Management

Copyright © 2018, Oracle and/or its affiliates. All rights reserved. | Confidential
Oracle Autonomous Database Revolutionizes Data Management

• Transform from database builder and maintainer to

user of autonomous database services
– Automates everything: provisioning, securing, repairing,
tuning, driving, and more ….
– Easy to create new databases and transform existing ones

• Enables you to:

– Innovate faster, spend less, and ensure data safety
– Transform your I.T. to modern cloud model

Copyright © 2018, Oracle and/or its affiliates. All rights reserved. | 7

Journey to Autonomous Database
Oracle has invested thousands of engineer years automating and optimizing database

• Autonomous Health Framework 18c

• Automatic Diagnostic Framework
• Automatic Refresh of Clones
• Automatic Columnar Flash
• Automatic IM population
• Automatic Memory Management 12c • Automatic Application
• Automatic Segment Space Mgmt Continuity
• Automatic Statistics Gathering
• Automatic Storage Management
• Automatic Workload Repository
• Automatic Diagnostic Monitor
11g
• Automatic Query Rewrite • Automatic SQL Tuning
• Automatic Undo Management • Automatic Workload Capture/Replay
10g • Automatic SQL Plan Management
9i • Automatic Capture of SQL Monitor
• Automatic Data Optimization

Copyright © 2018, Oracle and/or its affiliates. All rights reserved. | 8

 Journey to Autonomous Database
And thousands of engineer years automating and optimizing database infrastructure

• Exadata Cloud Service

• In-Memory Columnar in Flash 2018
• Smart Fusion Block Transfer
• Direct-to-wire Protocol
• JSON and XML offload
• Instant failure detection
• Network Resource Mgmt
• Prioritized File Recovery
• IO Priorities
• Data Mining Offload
• Offload Decryption
• Database Aware Flash Cache
• Storage Indexes
• Hybrid Columnar Data
• Smart Scan
• Infiniband Scale-Out

2008

Copyright © 2018, Oracle and/or its affiliates. All rights reserved. | 9

Autonomous Completes the Journey
Brings Full Automation to Entire Database Lifecycle

ORACLE Oracle
AUTONOMOUS Cloud
DATABASE

Complete Complete Automated

Infrastructure Database Data Center Operations
Automation Automation and Machine Learning

World’s First Autonomous Database

Copyright © 2018, Oracle and/or its affiliates. All rights reserved. | 10

 ORACLE
AUTONOMOUS
DATABASE

How It Works

Copyright © 2018, Oracle and/or its affiliates. All rights reserved. | 11

Autonomous Database Machine Learning
Diagnostics, recovery and optimizations for each layer of the deployment stack

Database Infrastructure Database Operations Workload Optimizations

Detection and recovery of Hang Management Query Optimizer

failed/sick server, storage or Anomaly Detection Real-time statistics
switch/link Maintenance Slot Identification Automatic Indexing
Bug Identification and Prioritization

Copyright © 2018, Oracle and/or its affiliates. All rights reserved. | 12

Continuous Optimization – Enabled by Machine Learning
• SQL Plans are like driving directions
– Should adapt as data volume (traffic) changes

• Indexes are like roads and bridges

– Should adapt if nature of the workload evolves

• Changes in data volume and SQL workload

are continuously captured
• Machine Learning algorithm processes
changes to find new optimal plans and
indexes
– Improved driving directions, roads, bridges

Copyright © 2018, Oracle and/or its affiliates. All rights reserved. | 13

Autonomous Optimization – Machine Learning meets Mission Critical

Identify • Machine Learning continuously monitors workload

Candidate and identifies potential new SQL plans and indexes
Improvements
– Plans are like driving directions, indexes are like new roads
• However, SQL plan and index changes are risky
Test
Improvement to – Changes that improve 999 cases but hurt 1 are unacceptable
Ensure Benefit
• Goal of Autonomous is to avoid ALL slowdowns
• Proposed changes tested outside application workload
Validate Change • Then benefit is validated on first execute of every
for Each SQL
changed SQL
– If performance regresses then old plan is restored
Copyright © 2018, Oracle and/or its affiliates. All rights reserved. | 14
Full Database Lifecycle Automation

Provision Secure Manage

Rapidly and easily creates Protects data from all Automates all infrastructure
mission critical databases external and internal threats and database maintenance

Creates Exadata+ Continuous threat detection, Patches all software online+,

Cloud Infrastructure, Applies security updates tunes settings, performs
Real Application Clusters+ online+, prevents admin all OS and SYSDBA
scale-out database snooping, encrypts all data operations
+ Unique to Oracle

Copyright © 2018, Oracle and/or its affiliates. All rights reserved. | 15

Full Database Lifecycle Automation

1 2 3 4 5

Protect Scale Optimize

Recovers from any failure Scales online for highest Optimally runs workloads
without downtime performance and lowest cost without human direction

Automates backup, restore, Instant online elasticity+ Automatically optimizes

application transparent+ of serverless data formats, parallelism+,
cluster failover, diagnoses compute and storage memory, and plans for each
and repairs errors+ enables true pay-per-use+ workload

+ Unique to Oracle

Copyright © 2018, Oracle and/or its affiliates. All rights reserved. | 16

One Autonomous Database - Optimized by Workload

ORACLE
AUTONOMOUS
DATABASE
Available March 2018 Available August 2018

Autonomous Autonomous
Data Warehouse (ADW) Transaction Processing (ATP)
Best for all Analytic Workloads: Best for TP and Mixed Workloads:
• Data Warehouse, Data Mart • Transactions, Batch, Reporting, IoT
• Data Lake, Machine Learning • Application Dev, Machine Learning

Copyright © 2018, Oracle and/or its affiliates. All rights reserved. | 17

Autonomous Optimizations - Specialized by Workload

ADW ATP
Primary Goal Fast Complex Analytics Fast Transactions Processing
Data Formats Columnar Row
Data Access Acceleration Creates Data Summaries RDMA for messaging and IO
Memory Usage Parallel Joins and Aggregations Data Caching to Avoid IO
Statistics Automatically manages optimizer statistics as data changes

Copyright © 2018, Oracle and/or its affiliates. All rights reserved. | 18

Automated Tuning
• “Load and go”
– Define tables, load data, run queries
• No tuning required
• No special database expertise required
• No need to worry about tablespaces, partitioning,
compression, in-memory, indexes, parallel
execution
– Fast performance out of the box with zero
tuning
– Simple web-based monitoring console
– Built-in resource-management plans

Copyright © 2018, Oracle and/or its affiliates. All rights reserved. |

Autonomous Data Warehouse:
Architecture
Developer Tools Autonomous
Data Warehouse Data Warehouse
Services
(EDWs, DW, departmental marts and sandboxes)
Service Management Built-in Access Tools

Oracle Exadata Cloud Service

SQL Developer
Oracle Database Cloud Service
Service Console Oracle ML
Express Cloud Service
Data Integration
Services
Oracle Data Integration
Platform Cloud
DW Database
3rd Party DI on
Oracle Cloud Compute

3rd Party DI
On-premises

Copyright © 2018, Oracle and/or its affiliates. All rights reserved. | Confidential
Autonomous Data Warehouse | Key Use Cases

Data Marts / Sandboxes for Data

Warehouses Data Scientists Lakes

Business Analytics Machine Learning Query All Data

Copyright © 2018, Oracle and/or its affiliates. All rights reserved. | 21

OracleML: Built-in notebook

• Collaborative UI for data scientists

– Easy access to shared notebooks,
templates, permissions, scheduler, etc.
• Based on Apache Zeppelin
• Roadmap: Common UI for data
scientists across multiple services

Copyright © 2018, Oracle and/or its affiliates. All rights reserved. | 22

Third Party Business Intelligence/Integration Platforms
Supporting new architecture

Copyright © 2018, Oracle and/or its affiliates. All rights reserved. | Confidential – Oracle Internal/Restricted/Highly Restricted 23
ATP In Action On Netsuite Workload
• Ran a complex Netsuite workload, and compared ATP to existing expert tuning
• 17,542 SQL statements, 1,852 tables, 8,151 indexes - years of tuning to create these indexes
– Before running on ATP, all indexes and statistics were dropped

1172 1155 4663

Elapsed Indexes
Expert Autonomous Expert
Time Used 1733
(sec)
Autonomous

ATP achieved identical performance to expert manual tuning

ATP stays tuned as workload changes
Copyright © 2018, Oracle and/or its affiliates. All rights reserved. | 24
Automated Tuning
Performance Comparison: Autonomous DW vs Hand-Tuned

ADW Customer Net

• ADW-tuned database compared to Workload
(Load-and-Go) (Hand-Tuned) Result
customer-tuned database Customer #1 39 34 …as good
– Customer-tuned systems used indexes,
Customer #2 52 53 …as good
partitioning, compression etc as defined by
customer Customer #3 1032 911 …as good
– ADW automatically tuned Customer #4 2138 3051 Faster
– Identical hardware
Customer #5 132 196 Faster
ADW consistently exceeds or Customer #6 1478 1831 Faster
matches hand-tuned performance Customer #7 2551 5137 Faster

MSTR 1163 1504 Faster

SSB 232 209 …as good

Copyright © 2018, Oracle and/or its affiliates. All rights reserved. | Confidential
Autonomous Transaction Processing | Key Use Cases

Departmental / Mission Mixed Application

Critical Applications Workloads Development

Transactions Analytics

Copyright © 2018, Oracle and/or its affiliates. All rights reserved. | 26

Dedicated Exadata Cloud Infrastructure
• Customers can choose serverless ADW or ATP databases or deploy
on Dedicated Exadata Cloud Infrastructure for highest isolation

• Complete stack is isolated from other tenants:

– Compute, Storage, Cluster Interconnect
– Virtual Cloud Network is Hardware Enforced

• Unique Fully Isolated Cloud+ in Public Cloud

– Allows control over consolidation, updates
– Guaranteed 99.995% availability - under 2.5 min of downtime per month
+ Unique to Oracle Copyright © 2018, Oracle and/or its affiliates. All rights reserved. | 27
Autonomous Database on Cloud At Customer
• Autonomous Database is Cloud Only since it automates the entire stack:
– Servers, storage, network, interconnect, OS, VM, GI, DB

• Some customers cannot move to public cloud

– Regulations, corporate policies, network latency

• Autonomous Database on Exadata Cloud at Customer +

– Autonomous database in customer data center
– Same cloud hardware, software, interfaces, APIs
– Same pay-per-use subscription model
+ Unique to Oracle

Copyright © 2018, Oracle and/or its affiliates. All rights reserved. | 28

Conclusion: Revolutionize Your Data Management

• Transform from expensive and error prone building and maintaining of

databases to using autonomous database services on modern cloud

• Innovate more for less

– Develop and optimize new applications faster
– Cut runtime costs up to 90% ORACLE
– Eliminate full-stack administration costs
AUTONOMOUS
DATABASE
• Ensure data safety
– Protect against cyber-attack and data-theft vulnerabilities
– Guarantee 99.995% uptime – including maintenance
Copyright © 2018, Oracle and/or its affiliates. All rights reserved. | 29
Try It Now for Free

2 TB Autonomous Database FREE for 3,300 Hours

https://cloud.oracle.com/tryit

#thinkautonomous
Copyright © 2018, Oracle and/or its affiliates. All rights reserved. | 30