Big Data - CTOlabs.com

Report
Take Away Thoughts
On Big Data
Some Context You Can Use
A Summary and Discussion of Big Data Presentations at the USGIF
Technology Days
Bob Gourley
CTOlabs.com
June 2012
How Much Data Is there?
All of this afternoon’s presentations reminded me of a quote
that tried to spell out how big the universe is:
"Space," it says, "is big. Really big. You
just won't believe how vastly, hugely,
mindbogglingly big it is.
From the Hitchhikers Guide
Cross-Over in Hype
•
CTOlabs projected in 2008 that the
hype over Cloud Computing was
headed to overtake the hype over
SOA, and it did, in 2010.
•
The hype over Big Data is showing
signs of a faster growth, but we are
unsure at this point at what point the
hype will overtake that of cloud
computing.
•
Big reason to track this: It
underscores that we need to decide
what we mean by this term and not let
others decide for us.
Big Data
•
This is not business as usual.
•
And this is not just lots of data.
•
It is not just exponential growth of data.
•
It is new ways of making sense over data that require changes to existing
architectures.
•
Big Data, the term, in its current use, implies many other things, like:
•
•
Apache Hadoop Framework
•
Commodity hardware leveraging Moore’s law
•
Infinite scalability
•
No data temples
Caution: Big Data, the term, may soon lose all meaning.
I will define Big Data Now
What Is Big Data?
• Big Data is a term applied to data sets whose size is
beyond the ability of legacy approaches to capture,
manage and process the data within a tolerable elapsed
time. Big data sizes are a constantly moving target.
• For the enterprise CTO, speaking of “big data” implies
the need for a strategy for dealing with sensemaking over
large quantities of data.
More Context Follows
Big Data
•
Think of challenges facing humanity: They all need solutions.
•
But first we have to address a key problem:
Amount of data
to Analyze
Area of Need
(and opportunity)
Ability for humans
to analyze data
•
Old ways of doing IT must change. When you hear the
phrase “Big Data” think of this need for change.
The ability to collect, parse, analyze machine data in real time, whether on
premise or in the cloud, will continue to grow.
Ignite Style Presentations
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
Introductory Keynote – “Big Data: Friend or Foe” by Barry Barlow, Director, OnlineGEOINT Services, National
Geospatial-Intelligence Agency
Management of Multi-Terabyte Video Files in Broadcast Television – Steve Atkinson, Director Federal Sales,
Front Porch Digital Inc.
3,000,000 km2 of Satellite Imagery Every Day: Finding the Relevant Pixels – Luke Barrington, CTO, Tomnod;
and John Lucier, Senior Manager Analysis Offerings, DigitalGlobe
The Value of Stream Computing in Big Data – Gabe Chang, Senior Consulting Client IT Architect, IBM
Big Data: Security and Scale in the Era of On-Demand IT – MG (Ret.) John M. Custer, Director, Federal
Missions and Programs, EMC2
Intelligence Products in the Age of Big Data – George Demmy, CTO, TerraGo
Data Intensity and Datacenter Consolidation via Federated Cloud “Big Data” – Eng Lim Goh Ph.D., CTO and
Senior VP, Silicon Graphics International Corp.
A Big Ocean: Handling High Volume Bathymetry – Michael Henheffer, Senior Software Developer Marine
Division, CARIS
Big Features: A WISE Approach to Scalable Geospatial ISR Fusion – Joshua Lieberman, Senior Manager,
Deloitte FAS LLP
OMAR: Open Source Software for Big Data – Mark R. Lucas, Principal Scientist, RadiantBlue Technologies
Interacting with Big Data using Mobile Devices – Kumar Navular, Director, NextGeneration Products,
DigitalGlobe
Predictive Analytics in the Cloud: The Art of the Possible – by Anthony Quartararo, President and CEO,
Spatial Networks Inc.
Big Data: The Diversity Factor – Dan Quinn, Vice President of Sales and Marketing, Progressive Technology
Federal Systems Inc.
Data Harmonization Through the Use of Complex Event Processors – Dennis Groseclose, President,
TransVoyant
Standardizing Web Interfaces to Distribute Wide Area Motion Imagery – Rahul Thakar, VP of Technology,
PIXIA Corp.
Discussant – Bob Gourley, Chief Technology Officer, Crucial Point LLC
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
Recap
Big Data Friend or Foe: Barry Barlow says there is lots of data. Can't analyze it all with people.
Mentioned MapStory. RecordedFuture. Hadoop.
Big Data in Television: DBX is the current format. TB size files. Lessons regarding metadata
and workflow relevant to our world. Suggested use of “proxy workflows.”
Finding Relevant Pixels: New crowdsourcing methods of review.
Stream and Big Data: Interconnected world with fast moving data, analyze in stream.
Game-Changing Tech for Data Explosion: Value comes from analytics. Storage nothing, ot a
tsunami,it is a rising tide.. Use Hadoop. Put compute into storage.
Intel reporting in the age of Big Data: Data must be operatioinalized
Big Ocean: Bathymetry data smartly handled.
Big Features: Context for collaboration while working big data is crtical.
OMAR- Open Source for Big Data: These guys are working really BIG data, in a way that is
smooth and easy to users.
Big Data and Mobile: Well engineerd solutions can make it incredibly easy on users.
Data Intensity and Data Center Consolidation: Think BIG. Keep data where it is. Have a vision.
Predictive Analytics in the Cloud: Know your challenge. User experience is everything.
Activity-Based Intelligence: Must engineer for ABI. Sometimes bring data together.
Global Standardization of Web Interfaces for WAMI: Web services for MI
Diverse Requirements: Says Big Data Not New. Good ways to present to users.
Some discussion items
• Presenters at times used varying definitions of the term “big
data”
• We can’t command that the term always be used the way the
tech community uses it. But we can push back whenever
someone decides to invent their own definition.
What To Do Now:
• Sign up for the Government Big Data Newsletter
• Send me an e-mail, at bob @ crucialpointllc.com
• Or sign up at CTOvision.com
• Leverage the Apache Hadoop framework. Download open
source distribution of CDH4 from Cloudera
• Share your lessons learned, and seek lessons from others on
your Big Data use.
• Be precise on how you use the term, or expect its meaning to
disappear
Backup Slides
Big Data Use Cases in Government
• Security: rapid real time analysis of all relevant data
• Rapid return of geospatial data
• Location based push of data: ads now, but watch for more
• Real time return of relevant search: Google, Cloudera and
USA.gov
• Real time suggestion of topics: Google, Cloudera and USA.gov
• Bioinformatics: Human Genome, Hadoop
• Bioinformatics: Patient location, treatment, outcomes
Big Data and the Special Case of Cyber
• Cyber security has long generated large
quantities of data.
• Enterprises need access to all the data to
look for evidence of
coordinated/sophisticated adversary
action. Old approaches do not enable
that.
• New CDH4 enabled “Big Data”
approaches enable “Enterprise Security
Intelligence” solutions.
• Includes non-cyber data. Rapid computer
emergency response needs all source
data along with the cyber data.
New approaches to cyber security
analysis, including incident
detection, incident response,
forensics and remediation, require
“Big Data” thinking and designs
built to bring all the data together.
The Intent of the Government Big Data Solutions
Award
 Established to help facilitate exchange of best practices, lessons
learned and creative ideas for solutions to hard data challenges
 Special focus on solutions built around Apache Hadoop framework
 Nominees and award winners to be written up in CTOlabs.com
technology reviews
 Award meant to help generate exchange of lessons learned
We established a team of judges, asked them to consider mission impact as primary
criteria, and solicited award nominations via sites frequented by government IT
professionals and solution providers.
The Government Needs More Agility*
“High tech runs three-times faster than normal businesses. And the
government runs three-times slower than normal businesses.
So we have a nine-times gap”
– Andy Grove
 The government can rapidly benefit from the lessons of high tech by
being a faster follower, especially when it comes to Big Data
constructs
 Thesis: If the Big Data community understands more about federal
missions, challenges and successes, we can improve the speed
and effectiveness of federal solutions.
*Among other needs
Most active fed solution areas:
 Federal integrators: Spending internal research and development
funds to create prototypes and full solutions relevant to fed missions
 DoD and IC agencies: Using Big Data approaches to solve “needle
in the haystack” and “connect the dots” problems
 National Labs: Bioinformatics solutions have been put in place by
federal researchers
 OMB and GSA: Ensuring sharing of lessons and solutions. Key
exemplars around web search methods. Solutions inside
government agencies and on citizen facing properties
Big Data solutions are already making a difference in government service to citizens.
Highlighting some of this virtuous work is a goal of our Government Big Data Solutions
Award.
Top Nominees for 2011
 USA Search: Best in class hosted search services over more than 400 gov sites.
Great use of CDH3.
 GCE Federal: Cloud-based financial management solutions. Apache Hadoop,
Hbase, Lucene for Dept of Labor.
 PNNL Bioinformatics: Leading researcher Dr. Taylor of PNNL is advancing
understanding of health, biology, genetics and computing using Apache
Hadoop/MapReduce/HBase.
 SherpaSurfing: Use of CDH as a cybersecurity solution. Ingest packet capture in
any format, analyze trends, find malware, alert.
 US Department of State: Bureau of Counselor Affairs. Large data with important
applications for citizen service and national security.
Each of these are making a difference for government missions right now.
USA Search
 Program of General Services Administration’s (GSA) Office of Citizen
Services and Information Technologies.
 Hosted search services for USA.gov and over 500 other government
websites.
 Solves big data challenges with open source capabilities.
 CDH3 since fall 2010. HDFS, Hadoop and Hive used in cost effective,
resilient, scalable solution.
 Search Results. Search Suggestions. Trend analysis. Analytic dashboards.
Bottom Line: USA Search brings the best of the open source community to multiple
government missions, including direct citizen support
Some Requests
• Sign up for the Government Big Data Newsletter at:
• http://j.mp/ctonews
• Watch for the 2012 Government Big Data Solutions Award
• Stay in touch!
Thank You!
Please give feedback and find more info at:
CTOvision.com

similar documents