@joe_Caserta
Big Data: Setting up a Big Data Lake
Joe Caserta
President
Caserta Concepts
September 17, 2015 - New York City
NEW YORK

@joe_Caserta

@joe_Caserta
Launched Big Data practice
Co-author, with Ralph Kimball, The
Data Warehouse ETL Toolkit
Data Analysis, Data Warehousing and
Business Intelligence since 1996
Began consulting database programing
and data modeling 25+ years hands-on experience
building database solutions
Founded
Caserta Concepts
Web log analytics solution published in
Intelligent Enterprise
Launched Data Science, Data
Interaction and Cloud practices
Laser focus on extending Data
Analytics with Big Data solutions
1986
2004
1996
2009
2001
2013
2012
2014
Dedicated to Data Governance
Techniques on Big Data (Innovation)
Top 20 Big Data
Consulting - CIO Review
Top 20 Most Powerful
Big Data consulting firms
Launched Big Data Warehousing
(BDW) Meetup NYC: 3,000+ Members
2015 Awarded for getting data out
of SAP for data analytics
Established best practices for big data
ecosystem implementations
Caserta Timeline
Awarded Top
Healthcare Analytics
Solution Provider

@joe_Caserta
About Caserta Concepts
• Consulting firm with focused expertise on Data Innovation, using Modern Data
Engineering approaches to solve highly complex business data challenges
• Award-winning company
• Internationally recognized work force
• Mentoring, Training, Knowledge Transfer
• Strategy, Architecture, Implementation
• An Innovation Partner
• Transformative Data Strategies
• Modern Data Engineering
• Advanced Architecture
• Leaders in architecting and implementing enterprise data solutions
• Data Warehousing
• Business Intelligence
• Big Data Analytics
• Data Science
• Data on the Cloud
• Data Interaction & Visualization
• Strategic Consulting
• Technical Design
• Build & Deploy Solutions

@joe_Caserta
Awards and Recognitions

@joe_Caserta
Client Portfolio
Retail/eCommerce
& Manufacturing
Digital Media/AdTech
Education & Services
Finance. Healthcare
& Insurance

@joe_Caserta
Partners

@joe_Caserta
Caserta Innovation Lab (CIL)
• Internal laboratory established to test & develop solution concepts and
ideas
• Used to accelerate client projects
• Examples:
• Search (SOLR) based BI
• Big Data Governance Toolkit
• Text Analytics on Social Network Data
• Continuous Integration / End-to-end streaming (Spark)
• Recommendation Engine Optimization
• Relationship Intelligence (Graph DB/Search)
• Others (confidential)
• CIL is hosted on

@joe_Caserta
Community
New York City
3,000+ members
Free Knowledge Sharing

@joe_Caserta
As a Mindful Cyborg, Chris utilizes up to 700 sensors, devices, applications, and
services to track, analyze, and optimize as many areas of his existence.
This quantification enables him to see the connections of otherwise invisible data,
resulting in dramatic upgrades to his health, productivity, and quality of life.
The Future is Today

@joe_Caserta
The Progression of Data Analytics
Descriptive
Analytics
Diagnostic
Analytics
Predictive
Analytics
Prescriptive
Analytics
What
happened?
Why did it
happen?
What will
happen?
How can we make
It happen?
Data Analytics Sophistication
BusinessValue
Source: Gartner
Reports  Correlations  Predictions  Recommendations
Cognitive Computing / Cognitive Data Analytics


@joe_Caserta
Innovation is the only sustainable competitive advantage a company can have
Innovations may fail, but companies that don’t innovate will fail

@joe_Caserta
What’s New in Modern Data Engineering?

@joe_Caserta
What you need to know (according to Joe)
Hadoop Distribution: Apache, Cloudera, Hortonworks, MapR, IBM
 Tools:
 Hive: Map data to structures and use SQL-like queries
 Pig: Data transformation language for big data
 Sqoop: Extracts external sources and loads Hadoop
 Storm: Real-time ETL
 Spark: General-purpose cluster computing framework
 NoSQL:
 Document: MongoDB, CouchDB
 Graph: Neo4j, Titan
 Key Value: Riak, Redis
 Columnar: Cassandra, Hbase
 Search: Lucene, Solr, ElasticSearch
 Languages: Python, Java, R, Scala

@joe_Caserta
Enrollments
Claims
Finance
ETL
Ad-Hoc Query
Horizontally Scalable Environment - Optimized for Analytics
Big Data Lake
Canned Reporting
Big Data Analytics
NoSQL
Databases
ETL
Ad-Hoc/Canned
Reporting
Traditional BI
Spark MapReduce Pig/Hive
N1 N2 N4N3 N5
Hadoop Distributed File System (HDFS)
Traditional
EDW
Others…
The Evolution of Modern Data Engineering
Data Science

@joe_Caserta
How We’ve Built Data Warehouses
•Design – Top Down / Bottom Up
• Customer Interviews and requirements gathering
• Data Profiling
•Extract Transform Load data from source to data
warehouse
•Create Facts and Dimensions
•Put a BI tool on top
•Develop reports
•Data Governance

@joe_Caserta
The Traditional Conversation
• Kimball Vs. Inmon
• Dimensional vs. 3rd Normal Form
• What hardware do we need (that will be ready in 6 months)
• Oracle vs SQL Server, Postgres or MySQL if we were brave
(and cheap)
• Which ETL tool should we BUY  Informatica, Datastage?
• Which BI tool should we sit on top  Business Objects,
Cognos?

@joe_Caserta
The New Conversation
• Do we need a Data Warehouse at all?
• If we do, does it need to be relational?
• Should we leverage Hadoop or NoSQL?
• Which platform and language are we going to code in?
• Which bleeding edge Apache Project should we put in
production!

@joe_Caserta
Why Change?
New technologies are great and all.. But what drives our
adoption of new technologies and techniques?
• Data has changed – Semistructured, Unstructured, Sparse
and evolving schema
• Volumes have changed  GB to TB to PB workloads
• Cracks in the Armor of Traditional Data Warehousing
approach!
AND MOST IMPORTANTLY:
Companies that innovate to leverage their data win!

@joe_Caserta
Cracks in the Data Warehouse Armor
• Onboarding new data is difficult!
• Data structures are rigid!
• Data Governance is slow!
• Disconnected from business needs:
“Hey – I need to munge some new data to see if it has value”
Wait! We have to….
Profile, analyze and conform the data
Change data models and load it into dimensional models
Build a semantic layer – that nobody is going to use
Create a dashboard we hope someone will notice
..and then you can have at it 3-6 months later to see if it has value!

@joe_Caserta
Is Anyone Surprised?
DWs have 70% FAILURE RATE
• Semi-scientific analysis has proven the majority of data analytic
projects fail..
• And of those that don’t fail, only a fraction are deemed a
“success”, others just finish!
• Data is just REALLY hard, especially without the right strategy
What do we think the Data Governance failure rate is?

@joe_Caserta
Is Traditional Warehousing All Wrong?
NO!
The concept of a Data Warehouse is sound:
•Consolidating data from disparate source systems
•Clean and conformed reference data
•Clean and integrated business facts
•Data governance (a more pragmatic version)
We can be more successful by acknowledging the
EDW can’t solve all problems.

@joe_Caserta
So what’s missing?
The Data Lake
A storage and processing layer for all data
• Store anything: source data, semi-structured,
unstructured, structured
• Keep it as long as needed
• Support a number of processing workloads
• Scale-out
..and here is where Hadoop
can help us!

@joe_Caserta
Hadoop (Typically) Powers the Data Lake
Hadoop Provides us:
• Distributed storage  HDFS
• Resource Management  YARN
• Many workloads, not just Map Reduce

@joe_Caserta
Governing Big Data
 Before Data Governance
 Users trying to produce reports from raw source data
 No Data Conformance
 No Master Data Management
 No Data Quality processes
 No Trust: Two analysts were almost guaranteed to come up
with two different sets of numbers!
 Before Big Data Governance
 We can put “anything” in Hadoop
 We can analyze anything
 We’re scientists, we don’t need IT, we make the rules
 Rule #1: Dumping data into Hadoop with no repeatable process, procedure, or
governance will create a mess
 Rule #2: Information harvested from an ungoverned systems will take us back to
the old days: No Trust = Not Actionable

@joe_Caserta
•This is the ‘people’ part. Establishing Enterprise Data Council,
Data Stewards, etc.Organization
•Definitions, lineage (where does this data come from),
business definitions, technical metadataMetadata
•Identify and control sensitive data, regulatory compliancePrivacy/Security
•Data must be complete and correct. Measure, improve,
certify
Data Quality and
Monitoring
•Policies around data frequency, source availability, etc.Business Process Integration
•Ensure consistent business critical data i.e. Members,
Providers, Agents, etc.Master Data Management
•Data retention, purge schedule, storage/archiving
Information Lifecycle
Management (ILM)
Data Governance
• Add Big Data to overall framework and assign responsibility
• Add data scientists to the Stewardship program
• Assign stewards to new data sets (twitter, call center logs, etc.)
• Graph databases are more flexible than relational
• Lower latency service required
• Distributed data quality and matching algorithms
• Data Quality and Monitoring (probably home grown, drools?)
• Quality checks not only SQL: machine learning, Pig and Map Reduce
• Acting on large dataset quality checks may require distribution
• Larger scale
• New datatypes
• Integrate with Hive Metastore, HCatalog, home grown tables
• Secure and mask multiple data types (not just tabular)
• Deletes are more uncommon (unless there is regulatory requirement)
• Take advantage of compression and archiving (like AWS Glacier)
• Data detection and masking on unstructured data upon ingest
• Near-zero latency, DevOps, Core component of business operations
for Big Data

@joe_Caserta
Making it Right
 The promise is an “agile” data culture where communities of users are encouraged
to explore new datasets in new ways
 New tools
 External data
 Data blending
 Decentralization
 With all the V’s, data scientists, new tools, new data we must rely LESS on HUMANS
 We need more systemic administration
 We need systems, tools to help with big data governance
 This space is EXTREMELY immature!
 Steps towards Data Governance for the Data Lake
1. Establish difference between traditional data and big data governance
2. Establish basic rules for where new data governance can be applied
3. Establish processes for graduating the products of data science to
governance
4. Establish a set of tools to make governing Big Data feasible

@joe_Caserta
Process Architecture
Communication
Organization
IFP
Governance
Administration
Compliance
Reporting
Standards
Value Proposition
Risk/Reward
Information
Accountabilities
Stewardship
Architecture
Enterprise Data
Council
Data Integrity
Metrics
Control Mechanisms
Principles and
Standards
Information Usability
Communication
BDG provides vision, oversight and accountability for leveraging
corporate information assets to create competitive advantage,
and accelerate the vision of integrated delivery.
Value Creation
• Acts on Requirements
Build Capabilities
• Does the Work
• Responsible for adherence
Governance
Committees
Data Stewards
Project Teams
Enterprise
Data Council
• Executive Oversight
• Prioritizes work
Drives change
Accountable for results
Definitions
Data Governance for the Data Lake

@joe_Caserta
Data Lake Governance Realities
 Full data governance can only be applied to “Structured” data
 The data must have a known and well documented schema
 This can include materialized endpoints such as files or tables OR
projections such as a Hive table
 Governed structured data must have:
 A known schema with Metadata
 A known and certified lineage
 A monitored, quality test, managed process for ingestion and
transformation
 A governed usage  Data isn’t just for enterprise BI tools anymore
 We talk about unstructured data in Hadoop but more-so it’s semi-
structured/structured with a definable schema.
 Even in the case of unstructured data, structure must be
extracted/applied in just about every case imaginable before analysis
can be performed.

@joe_Caserta
Modern Data Quality Priorities
Be
Corrective
Be Fast
Be
Transparent
Be Thorough

@joe_Caserta
Data Quality Priorities
Data Quality
SpeedtoValue
Fast
Slow
Raw Refined

@joe_Caserta
The Data Scientists Can Help!
 Data Science to Big Data Warehouse mapping
 Full Data Governance Requirements
 Provide full process lineage
 Data certification process by data stewards and business owners
 Ongoing Data Quality monitoring that includes Quality Checks
 Provide requirements for Data Lake
 Proper metadata established:
 Catalog
 Data Definitions
 Lineage
 Quality monitoring
 Know and validate data
completeness

@joe_Caserta
Big
Data
Warehouse
Data Science Workspace
Data Lake – Integrated Sandbox
Landing Area – Source Data in “Full Fidelity”
The Big Data Pyramid
Metadata  Catalog
ILM  who has access,
how long do we
“manage it”
Raw machine data
collection, collect
everything
Data is ready to be turned into
information: organized, well
defined, complete.
Agile business insight through data-
munging, machine learning, blending
with external data, development of
to-be BDW facts
Metadata  Catalog
ILM  who has access, how long do we
“manage it”
Data Quality and Monitoring 
Monitoring of completeness of data
Metadata  Catalog
ILM  who has access, how long do we “manage it”
Data Quality and Monitoring  Monitoring of
completeness of data
 Data has different governance demands at each tier
 Only top tier of the pyramid is fully governed
 We refer to this as the Trusted tier of the Big Data Warehouse.
Fully Data Governed ( trusted)
User community arbitrary queries and
reporting
Usage Pattern Data Governance

@joe_Caserta
Peeling back the layer… The Landing Area
•Source data in it’s full fidelity
•Programmatically Loaded
•Partitioned for data processing
•No governance other than catalog and ILM (Security
and Retention)
•Consumers: Data Scientists, ETL Processes,
Applications

@joe_Caserta
Data Lake
•Enriched, lightly integrated
•Data has been is accessible in the Hive Metastore
• Either processed into tabular relations
• Or via Hive Serdes directly upon Raw Data
•Partitioned for data access
•Governance additionally includes a guarantee of
completeness
•Consumers: Data Scientists, ETL Processes,
Applications, Data Analysts

@joe_Caserta
A Note On Unstructured Data
• A Structure must be extracted/applied in just about every
case imaginable before analysis can be performed.
• Full data governance can only be applied to “Structured”
data
• This can include materialized endpoints such as files or
tables OR projections such as a Hive table
• Governed structured data must have:
• A known schema with Metadata
• A known and certified lineage
• A monitored, quality test, managed process for ingestion
and transformation

@joe_Caserta
Data Science Workspace
•No barrier for onboarding and analysis of new data
•Blending of new data with entire Data Lake,
including the Big Data Warehouse
•Data Scientists enrich data with insight
•Consumers: Data Scientists (cool cats) only!

@joe_Caserta
Big Data Warehouse
•Data is Fully Governed
•Data is Structured
•Partitioned/tuned for data access
•Governance includes a guarantee of completeness
and accuracy
•Consumers: Data Scientists, ETL Processes,
Applications, Data Analysts, and Business Users
(the masses)
Big
Data
Warehouse

@joe_Caserta
The Refinery
BDW
Data Science
Workspace
Data Lake
Landing Area
Cool
new
data
New
Insights
•The feedback loop between Data Science and Data
Warehouse is critical
•Successful work products of science must Graduate
into the appropriate layers of the Data Lake

@joe_Caserta
Big Data Warehouse Technology?
“Polyglot Persistence - where any decent sized
enterprise will have a variety of different data
storage technologies for different kinds of data.
There will still be large amounts of it managed in
relational stores, but increasingly we'll be first asking
how we want to manipulate the data and only then
figuring out what technology is the best bet for it…”
- Martin Fowler (http://martinfowler.com)
Abridged Version: Use the right tool for the job!

@joe_Caserta
Polyglot Warehouse
We promote the concept that the Big Data
Warehouse may live in one or more platforms
•Full Hadoop Solutions
•Hadoop plus MPP or Relational
Supplemental technologies:
•NoSQL: Columnar, Key value, Timeseries, Graph
•Search Technologies

@joe_Caserta
Hadoop is the Data Warehouse?
•Hadoop can be the entire data pyramid platform for
including landing, data lake and the Big Data
Warehouse
•Especially serves as the Data Lake and “Refinery”
•Query engines such as Hive, and Impala provide SQL
support

@joe_Caserta
More Typical: Hadoop + Relational
•Hadoop is the platform for the Data Lake and
Refinery
•The Active Set is federated out into MPP or
Relational Platforms  Presentation Layer
•Serves as a good model when there is existing MPP
or Relational Data Warehouse in place

@joe_Caserta
On the Cloud
AWS and other cloud providers present a very
powerful design pattern:
•S3 serves as the storage layer for the Data Lake
•EMR (Elastic Hadoop) provides the Refinery, most
clusters can be ephemeral
•The Active Set is stored into Redshift MPP or
Relational Platforms
Eliminate massive on premise appliance footprint

@joe_Caserta
Data Warehousing is not Dead!
• The principles of Data Warehousing still makes sense
• Recognize gaps in feature/functionality of the Relational
Database, and traditional Data Warehousing
• Believe in the Data Lake and accept Tunable Governance
• Think Polyglot Warehouse and use the right tool for the job

@joe_Caserta
What skills are needed?
Modern Data
Engineering/Data
Preparation
Domain
Knowledge/Business
Expertise
Advanced
Mathematics/
Statistics

@joe_Caserta
What about the tools I have?
People, Processes and Business commitment is still critical!
Caution: Some Assembly Required
The V’s require robust tooling:
Some of the most hopeful tools are
brand new or in incubation!
Enterprise big data implementations
typically combine products with some
custom built components

@joe_Caserta
Use Cases
• Real-Time Trade Data Analytics
• Comply with Dodd-Frank
• Electronic Medical Record Analytics
• Save lives?

@joe_Caserta
High Volume Trade Data Project
• The equity trading arm of a large US bank needed to scale its
infrastructure to enable the ability to process/parse trade data real-time
and calculate aggregations/statistics
~ 1.4Million/second ~12 Billion messages/day ~240 Billon/month
• The solution needed to map the raw data to a data model in memory or
low latency (for real-time), while persisting mapped data to disk (for end
of day reporting).
• The proposed solution also needed to handle
ad-hoc data requests for data analytics.

@joe_Caserta
The Data
• Primarily FIX messages: Financial Information Exchange
• Established in early 90's as a standard for trade data communication
widely used throughout the industry
• Basically a delimited file of variable attribute-value pairs
• Looks something like this:
8=FIX.4.2 | 9=178 | 35=8 | 49=PHLX | 56=PERS | 52=20071123-05:30:00.000 |
11=ATOMNOCCC9990900 | 20=3 | 150=E | 39=E | 55=MSFT | 167=CS | 54=1 | 38=15 | 40=2 |
44=15 | 58=PHLX EQUITY TESTING | 59=0 | 47=C | 32=0 | 31=0 | 151=15 | 14=0 | 6=0 |
10=128 |
• A single trade can be comprised of 100's of such messages, although
typical trades have about a dozen

@joe_Caserta
Data Quality
Rules Engine
Storm Cluster
Trade
Data
d3.js Real-time
Analytics
Hadoop Cluster
Low Latency
Analytics
Atomic data
Aggregates
Event Monitors
• The Kafka messaging system is used for ingestion
• Storm is used for real-time ETL and outputs atomic data and
derived data needed for analytics
• Redis is used as a reference data lookup cache
• Real time analytics are produced from the aggregated data.
• Higher latency ad-hoc analytics are done in Hadoop using Pig
and Hive
Kafka
High Volume Real-time Analytics
Solution Architecture

@joe_Caserta
Electronic Medical Records (EMR) Analytics
Hadoop Data LakeEdge Node
`
100k
files
variant 1..n
…
variant 1..n
HDFS
Put
Netezza
DW
Sqoop
Pig EMR
Processor
UDF
Library
Provider
table
(parquet)
Member table
(parquet)
Python Wrapper
Provider
table
Member
table
Forqlift
Sequenc
e Files
…
variant 1..n
Sequenc
e Files
…
15 More
Entities
(parquet)
More
Dimensions
And
Facts
• Receive Electronic Medial Records from various providers in various formats
• Address Hadoop ‘small file’ problem
• No barrier for onboarding and analysis of new data
• Blend new data with Data Lake and Big Data Warehouse
• Machine Learning
• Text Analytics
• Natural Language Processing
• Reporting
• Ad-hoc queries
• File ingestion
• Information Lifecycle Mgmt

@joe_Caserta
Some Thoughts – Enable the Future
 Big Data requires the convergence of
data governance, advanced data
engineering, data science and business
smarts
 Make sure your data can be trusted
and people can be held accountable for
impact caused by low data quality.
It takes a village to achieve all the tasks
required for effective big data strategy
& execution
 Get experts that have done it before!
Achieve the impossible…..
… everything is impossible until someone does it!

@joe_Caserta
Workshops: www.casertaconcepts.com/training
Sept 21-22 (2 days), Agile Data Warehousing
taught by Lawrence Corr
Sept 23-24 (2 days), ETL Architecture and Design
taught by Joe Caserta
(Big Data module added)
SAVE $300 by using discount code: DAMANYC
Agile DW & ETL Training in NYC, 2015
New York Executive Conference Center
1601 Broadway @48th St.
New York, NY 10019

@joe_Caserta
Recommended Reading

@joe_Caserta
Thank You / Q&A
Joe Caserta
President, Caserta Concepts
joe@casertaconcepts.com
(914) 261-3648
@joe_Caserta