Hadoop (BDA) and Oracle Technologies on BI Projects Mark Rittman, CTO, Rittman Mead Dutch Oracle Users Group, Jan 14th 2015

Transcription

1 Hadoop (BDA) and Oracle Technologies on BI Projects Mark Rittman, CTO, Rittman Mead Dutch Oracle Users Group, Jan 14th 2015

2 About the Speaker Mark Rittman, Co-Founder of Rittman Mead Oracle ACE Director, specialising in Oracle BI&DW 14 Years Experience with Oracle Technology Regular columnist for Oracle Magazine Author of two Oracle Press Oracle BI books Oracle Business Intelligence Developers Guide Oracle Exalytics Revealed Writer for Rittman Mead Blog : mark.rittman@rittmanmead.com Twitter

3 About Rittman Mead Oracle BI and DW Gold partner Winner of five UKOUG Partner of the Year awards in including BI World leading specialist partner for technical excellence, solutions delivery and innovation in Oracle BI Approximately 80 consultants worldwide All expert in Oracle BI and DW Offices in US (Atlanta), Europe, Australia and India Skills in broad range of supporting Oracle tools: OBIEE, OBIA ODIEE Essbase, Oracle OLAP GoldenGate Endeca

4 Agenda Part 1 : The Hadoop (BDA) technical stack for Oracle BI/DW projects Why are Oracle BI/DW customers adopting Hadoop (BDA) technologies? What are the Oracle and Cloudera products being used? New Oracle products on the roadmap - Big Data Discovery, Big Data SQL futures Where does OBIEE, ODI etc fit in with these new products Rittman Mead s development platform Part 2 : Rittman Mead Hadoop (BDA) + Oracle BI Project Experiences What is Cloudera CDH, and the BDA, like to work with? How do we approach projects and PoCs? What architecture and approach do we actually take, now? How well do OBIEE and ODI work with Hadoop and BDA? What are the emerging techs, products and architectures we see for 2015+?

5 Part 1 : The Hadoop (BDA) technical stack for Oracle BI/DW projects or How did we get here?

6 15+ Years in Oracle BI and Data Warehousing Started back in 1997 on a bank Oracle DW project Our tools were Oracle 7.3.4, SQL*Plus, PL/SQL and shell scripts Went on to use Oracle Developer/2000 and Designer/2000 Our initial users queried the DW using SQL*Plus And later on, we rolled-out Discoverer/2000 to everyone else And life was fun

7 The Oracle-Centric DW Architecture Over time, this data warehouse architecture developed Added Oracle Warehouse Builder to automate and model the DW build Oracle 9i Application Server (yay!) to deliver reports and web portals Data Mining and OLAP in the database Oracle 9i for in-database ETL (and RAC) Data was typically loaded from Oracle RBDMS and EBS It was turtles Oracle all the way down

8 The State of the Art for BI & DW Was This.. Oracle Discoverer Drake - Combining Relational and OLAP Analysis for Oracle RDBMS Oracle Portal, part of Oracle 9iAS Oracle Warehouse Builder 9iAS / Paris

9 Then Came Siebel Analytics and OBIEE

10 The Oracle BI & DW World Changed Siebel Analytics replaced Oracle DIscoverer Oracle Data Integrator replaced Oracle Warehouse Builder Hyperion Essbase Replaced Oracle OLAP You were as likely to be loading from SQL Server as from Oracle They made us do things we didn t like to do Add a mid-tier virtual DW engine on top of the database Export data out of Oracle into an OLAP server Improve query performance using tools outside of the Oracle data warehouse It was all a bit scary Not to mention that WebLogic stuff

11 Introducing - The Oracle Reference DW Architecture Recognizing the difference between long-term storage of DW data (the foundation layer) And organizing the data for queries and easy navigation (the access + performance layer ) Also recognising where OBIEE had been game-changing - federated queries Things are good again

12 and now this happened

13

14 Today s Oracle Information Management Architecture Actionable Events Actionable Insights Actionable Information Structured Enterprise Data Input Events Event Engine Data Reservoir Data Factory Enterprise Information Store Reporting Other Data Execution Innovation Events & Data Discovery Lab Discovery Output

15 Virtualization & Query Federation Today s Layered Data Warehouse Architecture Data Sources Enterprise Performance Management Data Ingestion Access & Performance Layer Data Engines & Poly-structured sources Structured Data Sources Operational Data COTS Data Master & Ref. Data Streaming & BAM Foundation Data Layer Past, current and future interpretation of enterprise data. Structured to support agile access & navigation Immutable modelled data. Business Process Neutral form. Abstracted from business process changes Pre-built & Ad-hoc BI Assets Content Docs SMS Web & Social Media Raw Data Reservoir Immutable raw data reservoir Raw data at rest is not interpreted Information Interpretation Information Services Discovery Lab Sandboxes Project based data stores to support specific discovery objectives Rapid Development Sandboxes Project based data stored to facilitate rapid content / presentation delivery Data Science

16 The Oracle Data Warehousing Platform

17 Introducing The Data Reservoir? A reservoir is a lake than also can process and refine (your data) Wide-ranging source of low-density, lower-value data to complement the DW

18 Oracle s Big Data Products Oracle Big Data Appliance Optimized hardware for Hadoop processing Cloudera Distribution incl. Hadoop Oracle Big Data Connectors, ODI etc Oracle Big Data Connectors Oracle Big Data SQL Oracle NoSQL Database Oracle Data Integrator Oracle R Distribution OBIEE, BI Publisher and Endeca Info Discovery

19 Just Released - Oracle Big Data SQL Part of Oracle Big Data 4.0 (BDA-only) Also requires Oracle Database 12c, Oracle Exadata Database Machine More on this later SQL Queries Exadata Database Server SmartScan SmartScan Exadata Storage Servers Hadoop Cluster Oracle Big Data SQL

20 Coming Soon : Oracle Big Data Discovery Combining of Endeca Server search, analysis and visualisation capabilities with Apache Spark data munging and transformation Analyse, parse, explore and wrangle data using graphical tools and a Spark-based transformation engine Create a catalog of the data on your Hadoop cluster, then search that catalog using Endeca Server Create recommendations of other datasets, based on what you re looking at now Visualize your datasets, discover new insights

21 Coming Soon : Oracle Data Enrichment Cloud Service Cloud-based service for loading, enriching, cleansing and supplementing Hadoop data Part of the Oracle Data Integration product family Used up-stream from Big Data Discovery Aims to solve the data quality problem for Hadoop

22 Combining Oracle RDBMS with Hadoop + NoSQL High-value, high-density data goes into Oracle RDBMS Better support for fast queries, summaries, referential integrity etc Lower-value, lower-density data goes into Hadoop + NoSQL Also provides flexible schema, more agile development Successful next-generation BI+DW projects combine both - neither on their own is sufficient

23 Productising the Next-Generation IM Architecture

24 Still a Key Role for Data Integration, and BI Tools Fast, scaleable low-cost / flexible-schema data capture using Hadoop + NoSQL (BDA) Long-term storage of the most important downstream data - Oracle RBDMS (Exadata) Fast analysis + business-friendly interface : OBIEE, Endeca (Exalytics), RTD etc

25 OBIEE for Enterprise Analysis Across all Data Sources Dashboards, analyses, OLAP analytics, scorecards, published reporting, mobile Presented as an integrated business semantic model Optional mid-tier query acceleration using Oracle Exalytics In-Memory Machine Access data from RBDMS, applications, Hadoop, OLAP, ADF BCs etc Business Presentation Layer (Reports, Dashboards) Enterprise Semantic Business Model In-Memory Caching Layer Application Sources Hadoop / NoSQL Sources DW / OLAP Sources

26 Bringing it All Together : Oracle Data Integrator 12c ODI provides an excellent framework for running Hadoop ETL jobs ELT approach pushes transformations down to Hadoop - leveraging power of cluster Hive, HBase, Sqoop and OLH/ODCH KMs provide native Hadoop loading / transformation Whilst still preserving RDBMS push-down Extensible to cover Pig, Spark etc Process orchestration Data quality / error handling Metadata and model-driven

27 Oracle s Product Strategy

28 Rittman Mead Hadoop (BDA) + Oracle BI Project Experiences Working with (Cloudera) Hadoop, + Hive, NoSQL, etc Working with the Oracle Big Data Appliance Typical Hadoop + BI Use-Cases How Rittman Mead approaches Hadoop + Oracle BI projects Hadoop things that keep the CIO awake at night ODI and Hadoop OBIEE and Hadoop Oracle Big Data SQL Futures - Apache Spark, Next-Generation Hive, Big Data Discovery

29 Why is Hadoop of Interest to Us? Gives us an ability to store more data, at more detail, for longer Provides a cost-effective way to analyse vast amounts of data Hadoop & NoSQL technologies can give us schema-on-read capabilities There s vast amounts of innovation in this area we can harness And it s very complementary to Oracle BI & DW

30 Oracle & Hadoop Use-Cases Use Hadoop as a low-cost, horizontally-scalable DW archive Use Hadoop, Hive and MapReduce for low-cost ETL staging Support standalone-hadoop analysis with Oracle reference data Extend the DW with new data sources, datatypes, detail-level data

31 The Killer, Tech-Focused Use Case : Data Reservoir A reservoir is a lake than also can process and refine (your data) Wide-ranging source of low-density, lower-value data to complement the DW

32 Typical Business Use Case : 360 Degree View of Cust / Process OLTP transactional tells us what happened (in the past), but not why Common customer requirement now is to get a 360 degree view of their activity Understand what s being said about them External drivers for interest, activity Understand more about customer intent, opinions One example is to add details of social media mentions, likes, tweets and retweets etc to the transactional dataset Correlate twitter activity with sales increases, drops Measure impact of social media strategy Gather and include textual, sentiment, contextual data from surveys, media etc

33 Initial PoC over 4-6 Weeks Focus on high-productivity data analyst tools to identify key data, insights Typically performed using R, CDH on VMs, lots of scripting, lots of client interaction Focus on the discovery phase Governance, dashboards, productionizing can come later

34 Discovery vs. Exploitation Project Phases Discovery and monetising steps in Big Data projects have different requirements Discovery phase Unbounded discovery Self-Service sandbox Wide toolset Promotion to Exploitation Commercial exploitation Narrower toolset Integration to operations Non-functional requirements Code standardisation & governance

35 Rittman Mead Development Lab 4 x 64GB VM Servers 256GB RAM across cluster 36TB Storage VMWare ESXi VCenter Additional iscsi 6TB storage Synology DS414 NAS Demo / Free Software Installs Cloudera CDH5 Express - and BDP2.2 for Tez Oracle RBDMS, OBIEE, ODI etc Oracle Big Data Connectors Oracle EM 12cR4 vmhost2 vmhost3 vmhost4 vmhost5 BDP 2.2 Cluster 5 x nodes Oracle RDBMS LDAP OBIEE 11g VCenter 64GB RAM, 6TB Disk Core i7 4x3.6GHz VMWare ESXi 5.5 CDH 5.3 Cluster 5 x nodes Kerberos-Secured KDC ODI12c BI Apps 11g 64GB RAM, 6TB Disk Core i7 4x3.6GHz VMWare ESXi 5.5 Vigor 2830n Router VPN, DHCP etc CDH 5.2 Cluster 6 x nodes (16-32GB RAM / node) 64GB RAM, 6TB Disk Core i7 4x3.6GHz VMWare ESXi 5.5 iscsi LUN shared VMFS cluster filesystem Synology DS414 NAS, 6TB (For testing VMWare VMotion failover, large HDFS datasets etc) 64GB RAM, 6TB Disk Core i7 4x3.6GHz VMWare ESXi 5.5 Mac Mini Server OS X Server DNS etc EM 12c R4 16GB RAM, 1TB Disk Core i7 2x3.6GHz

36 Cluster Management VMWare VSphere 5 + VCenter Server Oracle Enterprise Manager 12cR4 Cloud Control OSX Server Yosemite

37 So how well does it work?

38 Part 1 : Rittman Mead Hadoop (BDA) + Oracle BI Project Experiences

39 Typical RM Project BDA Topology Starter BDA rack, or full rack Kerberos-secured using included KDC server Integration with corporate LDAP for Cloudera Manager, Hue etc Developer access through Hue, Beeline, R Studio End-user access through OBIEE, Endeca and other tools With final datasets usually exported to Exadata or Exalytics

40 Oracle Big Data Appliance Engineered system for big data processing and analysis Optimized for enterprise Hadoop workloads 288 Intel Xeon E5 Processors 1152 GB total memory 648TB total raw storage capacity Cloudera Distribution of Hadoop Cloudera Manager Open-source R Oracle NoSQL Database Community Edition Oracle Enterprise Linux + Oracle JVM New - Oracle Big Data SQL

41 Working with Oracle Big Data Appliance Don t underestimate the value of pre-integrated - massive time-saver for client No need to integrate Big Data Connectors, ODI Agent etc with HDFS, Hive etc etc Single support route - raise SR with Oracle, they will route to Cloudera if needed Single patch process for whole cluster - OS, CDH etc etc Full access to Cloudera Enterprise features Otherwise just another CDH cluster in terms of SSH access etc We like it ;-)

42 Cloudera Distribution including Hadoop (CDH) Like Linux, you can set up your Hadoop system manually, or use a distribution Key Hadoop distributions include Cloudera CDH, Hortonworks HDP, MapR etc Cloudera CDH is the distribution Oracle use on Big Data Appliance Provides HDFS and Hadoop framework for BDA Includes Pig, Hive, Sqoop, Oozie, HBase Cloudera Impala for real-time SQL access Cloudera Manager & Hue

43 Cloudera Manager and Hue Web-based tools provided with Cloudera CDH Cloudera Manager used for cluster admin, maintenance (like Enterprise Manager Commercial tool developed by Cloudera Not enabled by default in BigDataLite VM Hue is a developer / analyst tool for working with Pig, Hive, Sqoop, HDFS etc Open source project included in CDH

44 Working with Cloudera Hadoop (CDH) - Observations Very good product stack, enterprise-friendly, big community, can do lots with free edition Cloudera have their favoured Hadoop technologies - Spark, Kafka Also makes use of Cloudera-specific tools - Impala, Cloudera Manager etc But ignores some tools that have value - Apache Tez for example Easy for an Oracle developer to get productive with the CDH stack But beware of some immature technologies / products Hive!= Oracle SQL Spark is very much an alpha product Limitations in things like LDAP integration, end-to-end security Lots of products in stack = lots of places to go to diagnose issues

45 CDH : Things That Work Well HDFS as a low-cost, flexible data store / reservoir; Hive for SQL access to structured + semi-structured HDFS data Pig, Spark, Python, R for data analysis and munging Cloudera Manager and Hue for web-based admin + dev access Hive Metastore / HCatalog HDFS Cluster Filesystem RDBMS Imports Real-Time Logs / Events File / Unstructured Imports

46 Oracle Big Data Connectors Oracle-licensed utilities to connect Hadoop to Oracle RBDMS Bulk-extract data from Hadoop to Oracle, or expose HDFS / Hive data as external tables Run R analysis and processing on Hadoop Leverage Hadoop compute resources to offload ETL and other work from Oracle RBDMS Enable Oracle SQL to access and load Hadoop data

47 Working with the Oracle Big Data Connectors Oracle Loader for Hadoop, Oracle SQL Connector for HDFS - rarely used Sqoop works both way (Oracle>Hadoop, Hadoop>Oracle) and is good enough OSCH replaced by Oracle Big Data SQL for direct Oracle>Hive access Oracle Advanced Analytics for Hadoop has been very useful though Run MapReduce jobs from R Run R functions across Hive tables

48 Oracle R Advanced Analytics for Hadoop Key Features Run R functions on Hive Dataframes Write MapReduce functions in R

49 Initial Data Scoping & Discovery using R R is typically used at start of a big data project to get a high-level understanding of the data Can be run as R standalone, or using Oracle R Advanced Analytics for Hadoop Do basic scan of incoming dataset, get counts, determine delimiters etc Distribution of values for columns Basic graphs and data discovery Use findings to drive design of parsing logic, Hive data structures, need for data scrubbing / correcting etc

50 Design Pattern : Discovery Lab Actionable Events Actionable Insights Actionable Information Structured Enterprise Data Input Events Event Engine Data Reservoir Data Factory Enterprise Information Store Reporting Other Data Execution Innovation Events & Data Discovery Lab Discovery Output

51 Design Pattern : Discovery Lab Specific focus on identifying commercial value for exploitation Small group of highly skilled individuals (aka Data Scientists) Iterative development approach data oriented NOT development oriented Wide range of tools and techniques applied Searching and discovering unstructured data Finding correlations and clusters Filtering, aggregating, deriving and enhancing data Data provisioned through Data Factory or own ETL Typically separate infrastructure but could also be unified Reservoir if resource managed effectively

52 For the Future - Oracle Big Data Discovery

53 Interactive Analysis & Exploration of Hadoop Data

54 Share and Collaborate on Big Data Discovery Projects

55 Typical RM Big Data Project Tools Used Data Loading Real-time via Flume Conf scripts Batch via Sqoop cmd-line exec Discovery phase Exploitation phase Data prep via R scripts, Python scripts etc Data analysis via R scripts, Python scripts, Pig, Spark etc Sharing output via Hive tables, Impala tables, HDFS files etc Data Export Batch via Sqoop cmd-line exec a.k.a. data munging a.k.a. the magic

56 Data Loading into Hadoop Default load type is real-time, streaming loads Batch / bulk loads only typically used to seed system Variety of sources including web log activity, event streams Target is typically HDFS (Hive) or HBase Data typically lands in raw state Lots of files and events, need to be filtered/aggregated Typically semi-structured (JSON, logs etc) High volume, high velocity - Which is why we use Hadoop rather than RBDMS (speed vs. ACID trade-off) Economics of Hadoop means its often possible to archive all incoming data at detail level Real-Time Logs / Events File / Unstructured Imports Loading Stage

57 Apache Flume : Distributed Transport for Log Activity Apache Flume is the standard way to transport log files from source through to target Initial use-case was webserver log files, but can transport any file from A>B Does not do data transformation, but can send to multiple targets / target types Mechanisms and checks to ensure successful transport of entries Has a concept of agents, sinks and channels Agents collect and forward log data Sinks store it in final destination Channels store log data en-route Simple configuration through INI files Handled outside of ODI12c

58 Apache Kafka : Reliable, Message-Based Developed by LinkedIn, designed to address Flume issues around reliability, throughput (though many of those issues have been addressed since) Designed for persistent messages as the common use case Website messages, events etc vs. log file entries Consumer (pull) rather than Producer (push) model Supports multiple consumers per message queue More complex to set up than Flume, and can use Flume as a consumer of messages But gaining popularity, especially alongside Spark Streaming

59 GoldenGate for Continuous Streaming to Hadoop Oracle GoldenGate is also an option, for streaming RDBMS transactions to Hadoop Leverages GoldenGate & HDFS / Hive Java APIs Sample Implementations on MOS Doc.ID (HDFS) and (Hive) Likely to be formal part of GoldenGate in future release - but usable now Can also integrate with Flume for delivery to HDFS - see MOS Doc.ID

60 NoSQL Databases Family of database types that reject tabular storage, SQL access and ACID compliance Useful as a way of landing data quickly + supporting random cell-level access by ETL process Focus is on scalability, speed and schema-on-read Oracle NoSQL Database - speed and scalability Apache HBase - speed, scalability and Hadoop MongoDB - native storage of JSON documents May or may not run on Hadoop, but associated with it Great choice for high-velocity data capture CRUD approach vs write-once/read many in HDFS

61 ODI on Hadoop - Big Data Projects Discover ETL Tools ODI provides an excellent framework for running Hadoop ETL jobs ELT approach pushes transformations down to Hadoop - leveraging power of cluster Hive, HBase, Sqoop and OLH/ODCH KMs provide native Hadoop loading / transformation Whilst still preserving RDBMS push-down Extensible to cover Pig, Spark etc Process orchestration Data quality / error handling Metadata and model-driven

62 ODI on Hadoop - How Well Does It Work? Very good for set-based processing of Hadoop data (HiveQL) Can run python, R etc scripts as procedures Brings metadata and team-based ETL development to Hadoop Process orchestration, error-handling etc Rapid innovation from the ODI Product Dev team - Spark KMs etc coming soon But requires Hadoop devs to learn ODI, or add ODI developer to the project

63 Options for Sharing Hadoop Output with Wider Audience During the discovery phase of a Hadoop project, audience are likely technical Most comfortable with data analyst tools, command-line, low-level access to the data During the exploitation phase, audience will be less technical Emphasis on graphical tools, and integration with wider reporting toolset + metadata Three main options for visualising and sharing Hadoop data 1.Coming Soon - Oracle Big Data Discovery (Endeca on Hadoop) 2.OBIEE reporting against Hadoop direct using Hive/Impala, or Oracle Big Data SQL 3.OBIEE reporting against an export of the Hadoop data, on Exalytics / RDBMS

64 Oracle Business Analytics and Big Data Sources OBIEE 11g can also make use of big data sources OBIEE supports Hive/Hadoop as a data source Oracle R Enterprise can expose R models through DB functions, columns Oracle Exalytics has InfiniBand connectivity to Oracle BDA Endeca Information Discovery can analyze unstructured and semi-structured sources Increasingly tighter-integration between OBIEE and Endeca

65 New in OBIEE : Hadoop Connectivity through Hive MapReduce jobs are typically written in Java, but Hive can make this simpler Hive is a query environment over Hadoop/MapReduce to support SQL-like queries Hive server accepts HiveQL queries via HiveODBC or HiveJDBC, automatically creates MapReduce jobs against data previously loaded into the Hive HDFS tables Approach used by ODI and OBIEE to gain access to Hadoop data Allows Hadoop data to be accessed just like any other data source

66 Importing Hadoop/Hive Metadata into RPD HiveODBC driver has to be installed into Windows environment, so that BI Administration tool can connect to Hive and return table metadata Import as ODBC datasource, change physical DB type to Apache Hadoop afterwards Note that OBIEE queries cannot span >1 Hive schema (no table prefixes) 2 1 3

67 OBIEE / HiveServer2 ODBC Driver Issue Most customers using BDAs are using CDH4 or CDH5 - which uses HiveServer2 OBIEE only ships/supports HiveServer1 ODBC drivers But OBIEE on Windows can use the Cloudera HiveServer2 ODBC drivers which isn t supported by Oracle but works!

68 Dealing with Hadoop / Hive Latency Option 1 : Impala Hadoop access through Hive can be slow - due to inherent latency in Hive Hive queries use MapReduce in the background to query Hadoop Spins-up Java VM on each query Generates MapReduce job Runs and collates the answer Great for large, distributed queries but not so good for speed-of-thought dashboards

69 Dealing with Hadoop / Hive Latency Option 1 : Use Impala Hive is slow - because it s meant to be used for batch-mode queries Many companies / projects are trying to improve Hive - one of which is Cloudera Cloudera Impala is an open-source but commercially-sponsored in-memory MPP platform Replaces Hive and MapReduce in the Hadoop stack Can we use this, instead of Hive, to access Hadoop? It will need to work with OBIEE Warning - it won t be a supported data source (yet )

70 How Impala Works A replacement for Hive, but uses Hive concepts and data dictionary (metastore) MPP (Massively Parallel Processing) query engine that runs within Hadoop Uses same file formats, security, resource management as Hadoop Impala Processes queries in-memory Hadoop Accesses standard HDFS file data HDFS etc Option to use Apache AVRO, RCFile, LZO or Parquet (column-store) Designed for interactive, real-time SQL-like access to Hadoop Impala Hadoop HDFS etc BI Server Presentation Svr Impala Hadoop HDFS etc Cloudera Impala ODBC Driver Impala Hadoop HDFS etc Impala Hadoop HDFS etc Multi-Node Hadoop Cluster

71 Connecting OBIEE to Cloudera Impala Warning - unsupported source - limited testing and no support from MOS Requires Cloudera Impala ODBC drivers - Windows or Linux (RHEL etc/sles) - 32/64 bit ODBC Driver / DSN connection steps similar to Hive

72 So Does Impala Work, as a Hive Substitute? With ORDER BY disabled in DB features, it appears to But not extensively tested by me, or Oracle But it s certainly interesting Reduces 30s, 180s queries down to 1s, 10s etc Impala, or one of the competitor projects (Drill, Dremel etc) assumed to be the real-time query replacement for Hive, in time Oracle announced planned support for Impala at OOW watch this space

73 Dealing with Hadoop / Hive Latency Option 2 : Export to Data Mart In most cases, for general reporting access, exporting into RDBMS makes sense Export Hive data from Hadoop into Oracle Data Mart or Data Warehouse Use Oracle RDBMS for high-value data analysis, full access to RBDMS optimisations Potentially use Exalytics for in-memory RBDMS access RDBMS Imports Real-Time Logs / Events Loading Stage Processing Stage Store / Export Stage File Exports RDBMS Exports File / Unstructured Imports

74 Dealing with Hadoop / Hive Latency Option 3 : Big Data SQL Preferred solution for customers with Oracle Big Data Appliance is Big Data SQL Oracle SQL Access to both relational, and Hive/NoSQL data sources Exadata-type SmartScan against Hadoop datasets Response-time equivalent to Impala or Hive on Tez No issues around HiveQL limitations Insulates end-users around differences between Oracle and Hive datasets

75 Oracle Big Data SQL Part of Oracle Big Data 4.0 (BDA-only) Also requires Oracle Database 12c, Oracle Exadata Database Machine Extends Oracle Data Dictionary to cover Hive Extends Oracle SQL and SmartScan to Hadoop Extends Oracle Security Model over Hadoop Fine-grained access control Data redaction, data masking Uses fast c-based readers where possible (vs. Hive MapReduce generation) Map Hadoop parallelism to Oracle PQ Big Data SQL engine works on top of YARN Like Spark, Tez, MR2 SmartScan Exadata Storage Servers SQL Queries Exadata Database Server Hadoop Cluster SmartScan Oracle Big Data SQL

76 View Hive Table Metadata in the Oracle Data Dictionary Oracle Database 12c with Big Data SQL option can view Hive table metadata Linked by Exadata configuration steps to one or more BDA clusters DBA_HIVE_TABLES and USER_HIVE_TABLES exposes Hive metadata Oracle SQL*Developer 4.0.3, with Cloudera Hive drivers, can connect to Hive metastore SQL> col database_name for a30 SQL> col table_name for a30 SQL> select database_name, table_name 2 from dba_hive_tables; DATABASE_NAME TABLE_NAME default access_per_post default access_per_post_categories default access_per_post_full default apachelog default categories default countries default cust default hive_raw_apache_access_log

77 Big Data SQL Server Dataflow Read data from HDFS Data Node Direct-path reads C-based readers when possible Use native Hadoop classes otherwise Translate bytes to Oracle Apply SmartScan to Oracle bytes Apply filters Project columns Parse JSON/XML Score models

78 Hive Access through Oracle External Tables + Hive Driver Big Data SQL accesses Hive tables through external table mechanism ORACLE_HIVE external table type imports Hive metastore metadata ORACLE_HDFS requires metadata to be specified Access parameters cluster and tablename specify Hive table source and BDA cluster CREATE TABLE access_per_post_categories( hostname varchar2(100), request_date varchar2(100), post_id varchar2(10), title varchar2(200), author varchar2(100), category varchar2(100), ip_integer number) organization external (type oracle_hive default directory default_dir access parameters(com.oracle.bigdata.tablename=default.access_per_post_categories));

79 Use Rich Oracle SQL Dialect over Hadoop (Hive) Data Ranking Functions rank, dense_rank, cume_dist, percent_rank, ntile Window Aggregate Functions Avg, sum, min, max, count, variance, first_value, last_value LAG/LEAD Functions Reporting Aggregate Functions Sum, Avg, ratio_to_report Statistical Aggregates Correlation, linear regression family, covariance Linear Regression Fitting of ordinary-least-squares regression line to set of number pairs Descriptive Statistics Correlations Pearson s correlation coefficients Crosstabs Chi squared, phi coefficinet Hypothesis Testing Student t-test, Bionomal test Distribution Anderson-Darling test - etc.

80 Leverages Hive Metastore for Hadoop Java Access Classes As with other next-gen SQL access layers, uses common Hive metastore table metadata Provides route to underlying Hadoop data for Oracle Big Data SQL c-based SmartScan

81 Extending SmartScan, and Oracle SQL, Across All Data Brings query-offloading features of Exadata to Oracle Big Data Appliance Query across both Oracle and Hadoop sources Intelligent query optimisation applies SmartScan close to ALL data Use same SQL dialect across both sources Apply same security rules, policies, user access rights across both sources

82 Example : Using Big Data SQL to Add Dimensions to Hive Data We want to add country and post details to a Hive table containing page accesses Post and Country details are stored in Oracle RBDMS reference tables Hive Weblog Activity table Oracle Dimension lookup tables Combined output in report form

83 Create ORACLE_HIVE External Table over Hive Table Use the ORACLE_HIVE access driver type to create Oracle external table over Hive table ACCESS_PER_POST_EXTTAB now appears in Oracle data dictionary

84 Import Oracle Tables, Create RPD joining Tables Together No need to use Hive ODBC drivers - Oracle OCI connection instead No issue around HiveServer1 vs HiveServer2; also Big Data SQL handles authentication with Hadoop cluster in background, Kerberos etc Transparent to OBIEE - all appear as Oracle tables Join across schemas if required

85 Create Physical Data Model from Imported Table Metadata Join ORACLE_HIVE external table containing log data, to reference tables from Oracle DB

86 Create Business Model and Presentation Layers Map incoming physical tables into a star schema Add aggregation method for fact measures Add logical keys for logical dimension tables Remove columns from fact table that aren t measures

87 Create Initial Analyses Against Combined Dataset Create analyses using full SQL features Access to Oracle RDBMS Advanced Analytics functions through EVALUATE, EVALUATE_AGGR etc Big Data SQL SmartScan feature provides fast, ad-hoc access to Hive data, avoiding MapReduce

88 Oracle / Hive Query Federation at the RDBMS Level Oracle Big Data SQL feature (not BI Server) takes care of query federation SQL required for fact table (web log activity) access sent to Big Data SQL agent on BDA Only columns (projection) and rows (filtering) required to answer query sent back to Exadata Storage Indexes used on both Exadata Storage Servers and BDA nodes to skip block reads for irrelevant data HDFS caching used to speed-up access to commonly-used HDFS data

89 Access to Full Set of Oracle Join Types No longer restricted to HiveQL equi-joins - Big Data SQL supports all Oracle join operators Use to join Hive data (using View over external table) to a IP range country lookup table using BETWEEN join operator

90 Add In Time Dimension Table Enables time-series reporting; pre-req for forecasting (linear regression-type queries) Map to Date field in view over ORACLE_HIVE table Convert incoming Hive STRING field to Oracle DATE for better time-series manipulation

91 Now Enables Time-Series Reporting Incl. Country Lookups

92 What About Oracle Big Data SQL and ODI12c? Hive, and MapReduce, are well suited to batch-type ETL jobs, but Not all join types are available in Hive - joins must be equality joins Any data from external Oracle RDBMS sources has to be staged in Hadoop before joining Limited set of HiveQL functions vs. Oracle SQL Oracle-based mappings have to import Hive data into DB before accessing it

93 Combining Oracle and Hadoop (Hive) Data in Mappings Example scenario : log data in Hadoop needs to be enriched with customer data in Oracle Hadoop (Hive) contains log activity and customer etc IDs Reference / customer data held in Oracle RBDMS How do we create a mapping that joins both datasets? movieapp_log_odistage.custid = CUSTOMER.CUSTID

94 Options for Importing Oracle / RDBMS Data into Hadoop Could export RBDMS data to file, and load using IKM File to Hive Oracle Big Data Connectors only export to Oracle, not import to Hadoop One option is to use Apache Sqoop, and new IKM SQL to Hive-HBase-File knowledge module Hadoop-native, automatically runs in parallel Uses native JDBC drivers, or OraOop (for example) Bi-directional in-and-out of Hadoop to RDBMS Join performed in Hive, using HiveQL With HiveQL limitations (only equi-joins) movieapp_log_odistage.custid = customer.custid Sqoop extract

95 New Option - Using Oracle Big Data SQL Oracle Big Data SQL provides ability for Exadata to reference Hive tables Use feature to create join in Oracle, bringing across Hive data through ORACLE_HIVE table

96 Oracle Big Data SQL and Data Integration Gives us the ability to easily bring in Hadoop (Hive) data into Oracle-based mappings Allows us to create Hive-based mappings that use Oracle SQL for transforms, joins Faster access to Hive data for real-time ETL scenarios Through Hive, bring NoSQL and semi-structured data access to Oracle ETL projects For our scenario - join weblog + customer data in Oracle RDBMS, no need to stage in Hive

97 Using Big Data SQL in an ODI12c Mapping By default, Hive table has to be exposed as an ORACLE_HIVE external table in Oracle first Then register that Oracle external table in ODI repository + model 3 1 External table creation in Oracle 2 Register in ODI Model Logical Mapping using just Oracle tables

98 Custom KM : LKM Hive to Oracle (Big Data SQL) ODI12c Big Data SQL example on BigDataLite VM uses a custom KM for Big Data SQL LKM Hive to Oracle (Big Data SQL) - KM code downloadable from java.net Allows Hive+Oracle joins by auto-creating ORACLE_HIVE extttab definition to enable Big Data SQL Hive table access

99 ODI12c Mapping Creates Temp Exttab, Joins to Oracle Big Data SQL Hive External Table created as temp object 1 Register in ODI Model 3 Main integration SQL routines uses regular Oracle SQL join 4 Hive table AP uses LKM Hive to Oracle (Big Data SQL) 2 IKM Oracle Insert

100 Finally What Keeps the CIO Awake at Night Security and Privacy Regulations Are we analysing and sharing data in compliance with privacy regulations? - And if we are - would customers think our use of it is ethical? Do I know if the data in my Hadoop cluster is *really* secure?

101 Hadoop Security By Default Connections between Hadoop services, and by users to services, aren t authenticated Security is fragmented : HDFS, Hive, OS user accounts, Hue, CM all separate models No single place to define security policies, groups, access rights No single tool to audit access and permissions By default, everything is open and trusted - reflects roots in academia, R&D, marketing depts

102 Secured Hadoop : Kerberos, Sentry, Data Encryption etc Available for most Hadoop distributions, part of core Hadoop Kerberos Authentication - enables service-to-service, and client-to-service authentication using MIT Kerberos or MS AD Kerberos Apache Sentry - Role-based Access Control for Hive, Impala and HDFS (CDH5.3+) Transparent at-rest HDFS encryption (CDH5.3+) Closes security loopholes, goes some way to Oracle-type data security

103 Oracle Big Data SQL : Single RBDMS/Hadoop Security Model Potential to extend Oracle security model over Hadoop (Hive) data Masking / Redaction VPD FGAC

104 Summary Hadoop and Oracle Big Data Appliance are increasingly appearing in BI+DW Projects Gives DW projects the ability to store more data, cheaper and more flexibly than before Enables non-relational (SQL) query tools and analysis techniques (R, Spark etc) Extends BI s capability to report and analyze across wider data sources Maturity varies widely in terms of tool maturity, and Oracle integration with Hadoop Trend is for Oracle to productize big data, creating tools + products around Oracle BDA We are probably at early stages - but very interesting times to be an Oracle BI+DW dev!

105 Thank You for Attending! Thank you for attending this presentation, and more information can be found at Contact us at or Look out for our book, Oracle Business Intelligence Developers Guide out now! Follow-us on Twitter or Facebook (facebook.com/rittmanmead)