Oracle OpenWorld 2016 Review - Focus on Data, BigData, Streaming Data, Machine Learning & Database Development

The AMIS Team
Oracle OpenWorld 2016, Nieuwegein, 13th October 2016
Oracle OpenWorld 2016 Review
Data - Database Development, BigData, BI

2
Data
• ACID is expensive
• OLTP is a niche
• Moving data takes long – bring processing to data
• SQL is omnipresent – expose all data in an SQL friendly way
– Including NoSQL and data on Hadoop
• Data from the past should be able to help us predict the future
– Bring on machine learning (aka AI aka predictive analytics)
– Aided by the citizen data scientist in (Big) Data Discovery
• Fast Data (big data at high velocity) should be handled in real time
– Enter: Streaming Analytics & Apache Kafka
• Oh and ehm …. a next major release of Oracle Database is available
– 12cR2 – only on the cloud for now
– Highlights: Sharding, More PDB (“virtual database”),
Approximate Query Processing, Leverage In-Memory
even more, JSON document generation and faster
JSON processing, Analytic Views

3
Learning
• How do we learn?
– Try something (else) => get feedback => learn
• Eventually:
– We get it (understanding) so we can predict the outcome
of a certain action in a new situation
– Or we have experienced enough situations to predict
the outcome in most situations with high confidence
• Through interpolation, extrapolation, etc.
– We remain clueless

4
Machine Learning
• Analyze Historical Data (input and result – training set) to discover
Patterns & Models
• Iteratively apply Models to [additional] Input (test set) and compare model
outcome with known actual result to improve the model
• Use Model to predict outcome for entirely new data

5
Data Discovery
A B C D E F G
1104534 ZTR 0.1 anijs 2 36 T
631148 ESE 132 rivier 0 21 S
-3 WGN 71 appel 0 1 -
1262300 ZTR 56 zes 2 41 T
315529 HVN 1290 hamer 0 11 -
788914 ASM 676 zwaluw 0 26 T
157762 HVN 9482 wie 0 6 -
946681 DHG 42 rond 1 31 T
-31539 WGN 2423 bruin 0 0 -
47338 HVN 54 hamer 0 16 P

6
Scatter Plot
Attribute F (Y-axis)vs Attribute A
0
5
10
15
20
25
30
35
40
45
-500000 0 500000 10000001500000
Y-Values
Y-Values

7
Scatter Plot
Attribute F (Y-axis)vs Attribute A
0
5
10
15
20
25
30
35
40
45
1960 1980 2000 2020
Age of Lucas Jellema vs Year
Y-Values

8
Data Discovery
Time City - - #Kids Age Level of
Education
1104534 ZTR 0.1 anijs 2 36 T
631148 ESE 132 rivier 0 21 S
-3 WGN 71 appel 0 1 -
1262300 ZTR 56 zes 2 41 T
315529 HVN 1290 hamer 0 11 -
788914 ASM 676 zwaluw 0 26 T
157762 HVN 9482 wie 0 6 -
946681 DHG 42 rond 1 31 T
-31539 WGN 2423 bruin 0 0 -
47338 HVN 54 hamer 0 16 P

9
Machine Learning, Data Mining
& Predictive Analytics

10
Recent developments
• More compute capacity, smarter algorithms and better analytical tooling
– Evolving Machine Learning
– Smart text and speech analysis (NLP, ESA)
– Real time predictions become a reality
– Streaming (event) Analytics
– Visualization
– Citizen Data Scientist
– SQL against Big Data
• More data available & accessible
(IoT, Social, Media, IT operations,
business processes,…)
• Better/larger/cheaper/faster
data storage capabilities

11
Many cloud services around
Big Data & Analytics

12
Big Data Integration
Reference Architecture
IngestPrepare
Transform,
Merge, Enrich
Save
GovernGovern
Serve
Analyze & Act
Present,
Leverage
& ‘Action’
Extract
Explore
Purge

14
Mapping Oracle portfolio to
Reference Architecture
Big Data Discovery
Data Visualization
BI CS
IT Analytics
Security Analytics
Log Analytics

15
Tip: OEMM - Oracle Enterprise
Metadata Management

16
Citizen Data Scientist
• Data Visualization CS
• Big Data Preparation CS
• Big Data Discovery CS
• Machine Learning CS

17
Oracle Machine Learning
Cloud Service

20
https://www.zeppelinhub.com/

22
Traditional approach
• All enterprise data is in the Oracle [relational] Database
– Except very unstructured documents - and sometimes even those

23
Center of the Data Universe is
shifting

24
Variety in data – Data Tiering
• How long relevant (hot vs cold vs dead)?
• How fine grained and how accurate?
• What is it used for?
– By whom, where, in what way, using which tools
• What format is it in/should it be in?
• How to be processed?
• What confidentiality & integrity is required?
• How much of it?

25
Trends around data storage
and data processing
• Take processing to data [to reduce data movement]
– Exadata SmartScan in Storage Cells (SQL & R processing)
– Hadoop MapReduce/Spark
– Coherence Processors
– Streaming Analytics
– Microservices, stand alone data domains
• Distributed data partitions – for scalability and parallelization [and fault
tolerance when also replicated]:
– Shards (Oracle Database 12cR2) and Partitioned External Tables
– TimesTen Velocity Scale – distributed In-Memory OLTP
– Hadoop HDFS, Apache Kafka
• New paradigms regarding transactional data
– CQRS (for example Oracle Database In Memory (read) / In Flash/On disk
(read/write), Write behind cache)
– Event Sourcing, Transaction Log

26
Oracle Database
• How much of your data
– Arrives through (business) transactions that require true ACID?
– Is involved in current business operations?
– Will ever be updated [again]?
– Plays a direct role in integrity [of other records]?
– Is actively accessed [on a regular basis] ?
– Really has to be in the OLTP engine?
• How much of the data currently in your OLTP engine could be off-loaded
– If that data remains accessible through SQL (even from within the OLTP engine,
without altering existing queries) with reasonable response times
• What if such off-loading
– Improves performance of the OLTP engine for transactions
– Shortens batch jobs [by engaging distributed, scale out processing options]
– Opens up possibilities for advanced analytics
– Potentially lowers the cost [licenses & specialized hardware] for the OLTP engine
– Introduces some change and complexity

27
Oracle Big Data SQL
• Big Data SQL: A ‘franchised query engine,’ enables scalable, integrated
access in situ to the entire Big Data Management System (BDMS)
– Meta data, Query execution, Workload Management, Data Optimization
– Primary role for Oracle Database – foundation for BDMS
See Statement of Direction: http://www.oracle.com/technetwork/database/bigdata-appliance/overview/sod-bdms-2015-04-final-2516729.pdf

Oracle Database Development
28

29
Oracle Forms
• Release 12.2.1.1 is available
– 12.2.2 (or 12.2.1.2) is planned for late 2016
– Support for Forms 12c: Premier Oct 2020, Extended 2023 – (and moving)
• Forms usually runs in browser – using the Java JRE plugin for Applets
– Modern browsers have stopped or will stop supporting the Java plugin
– Forms will either have to run on outdated browsers (IE, old versions of Firefox or
Chrome) or run outside the browser
– The main changes around Forms are around running Forms outside the browser –
as standalone Java Web Start (jnlp) application
– Also: Forms Helper – script for customizing post-install environment (simplified
WLST)
• On Reports:
– Reports 12c exists – it is the terminal release
– From now on, reporting should be done using BI Publisher
– BI Publisher has become part of the Developer Suite and will be included in the
WebLogic Suite

31
APEX – 5.1
• Interactive Grid – A new rich, client-side region type that allows editing multiple rows of
data in a dynamic, JSON-enabled grid, and supports multiple grids on a single page.
• Master / Detail / Detail – Provide a wizard interface to define declarative
master/detail/detail regions.
• New Charting Engine – Include a new JavaScript (Oracle JET) based charting engine
developed by Oracle which produces responsive and accessible HTML5 charts.
– AnyChart is on the way out
• Ability to have multiple tabs open to the same APEX application and isolate session
state
• Improved Wizards - fewer steps and more attributes set by default.
• Declarative RTL Support –declarative methods to control user interface direction-
support for both Left-to-Right and Right-To-Left languages.
• Packaged Applications – Improved framework and enhancements to the packaged
applications.
• Status: EA 2 is available (hosted) as of September 2016
– APEX 5.1 Production – early 2017?

32
New in SQL in 12cR2
• Listagg improvements
• Error handling for CAST function & new Validate_Conversion function
• Materialized Views
– Real Time Materialized Views (stale plus logs)
– Statement Level Refresh
• AL32UTF8 As the Default Database Character Set
• New capabilities for generating JSON documents directly from SQL
queries, improved JSON support in In Memory processing
• Beyond 12cR2
• Approximate Query Processing (using HyperLogLog)
• Analytical Views
• Band Join- better performance for non-equijoins
• Temporary, cached in memory tables for duration of cursor
• Partitioned External Tables

33
New in PL/SQL in 12cR2
• Deprecated procedures and functions
• Accessible by at procedure or function level
• JSON support: generation of JSON documents using PL/SQL API and
Oracle supplied Object Types (somewhat akin to XMLType)
– JSON SQL functions available in PL/SQL expressions
• Supplied package dbms_plsql_code_coverage to identify code units not
touched in specific
[test] scenarios
• PL/Scope enhancements –
more fine grained reporting
• Edition Based Redefinition
does ‘garbage collection’ –
editioned objects no
longer in use
are cleaned up

34
Other Database Development
News
• JS Stored Procedures
• SQL Developer GUI Debugger
– One session can have another start debugging
– At breakpoint: execute SQL to inspect run context – including PL/SQL state
• SQLcl
• ORDS – Oracle REST Data Services

36
Summary of
Oracle OpenWorld 2016
• (5 days filled to brim with 1800+ sessions, 12 keynotes, 150+ demo
booths, hundreds of vendors and quite a few rumors & hallway tales)
• Infrastructure [as a Service]
– Generation 2 Data Center
– Network & IOPS (storage, NVMe, Flash)
– Exadata SL
• Abdication of the single, central, enterprise Oracle RDBMS – and the
limelight for data
– PDBs
– Sharding
– Hadoop & Spark (& SQL & R)
– Machine Learning
• Adoption of open source projects, industry trends & community darlings
– Node.js, Docker, Microservices, Git(Hub), Python, Slack, …

37
Summary of
Oracle OpenWorld 2016 (2)
• Cloud [First] strategy
– Migrate & Extend i/o [bidirectional] Lift & Shift
– Cloud@Customer
– Status and future of On Premises software (and yet Engineered Systems)
– Ops in Oracle Data Centers
– Subscription Models, Suites (i/o a la carte)
– How fast can Oracle move [without spreading itself too thin]?
• SaaS and [Unlimited] Applications
– SaaS portfolio quite extensive
– UX is important asset of the SaaS applications
– Real cloud elements are improving (APIs, extensibility)
– Traditional Apps are still evolving [as promised] – and seem to benefit from SaaS and
technological advances across the board
• Oracle Public Cloud consistency, architecture and the Dogfood Doctrine
– Fabric and foundational components
– Designated capabilities and mutual integration

38
Oracle OpenWorld 2016 Tag Cloud

Oracle OpenWorld 2016 Review - Focus on Data, BigData, Streaming Data, Machine Learning & Database Development

Oracle OpenWorld 2016 Review - Focus on Data, BigData, Streaming Data, Machine Learning & Database Development

More Related Content

What's hot (20)

Viewers also liked (20)

Similar to Oracle OpenWorld 2016 Review - Focus on Data, BigData, Streaming Data, Machine Learning & Database Development (20)

More from Lucas Jellema (20)

Recently uploaded (20)

Oracle OpenWorld 2016 Review - Focus on Data, BigData, Streaming Data, Machine Learning & Database Development

Editor's Notes