Presentation at Data-Intensive
Computing, Graphs, and
Combinatorics in
Bio-Informatics, Finance,
Linguistics,
and National Security
Tuesday–Wednesday, July 26–27, 2011
Call Girls Near Delhi Pride Hotel, New Delhi|9873777170
Financial Services Data - Use It or Lose It
1. Financial Services Data - Use It or Lose It!
CUNY High Performance Computing Center (HPCC)
Workshop on Data, Graphs and Combinatorics in Bio-Informatics,
Finance, Linguistics and National Security
July 26-27, 2011
Proprietary and Confidential. Not to be distributed or reproduced without permission www.sungard.com/globalservices
2. Why Am I Here?
Professional Personal
14+ years of
SunGard #2 in 2010
development on Wall
FINTECH 100
Street
Intersection of Global
Services Advanced
Products & Services
Technology &
across the FS
Information
spectrum
Management
Practices
Development +
Exploring impacts of Databases + Data
Big Data approach Warehousing +
for past 2+ years Trading + Risk
Management
Proprietary and Confidential. Not to be distributed or reproduced without permission www.sungard.com/globalservices 1
3. Why Are We All Here?
“Interpol and Deutsche Bank, FBI and Scotland Yard;
Business, Numbers, Money, People;
Crime, Travel, Communication, Entertainment;
Computer World!”
- Kraftwerk 1981
Proprietary and Confidential. Not to be distributed or reproduced without permission www.sungard.com/globalservices 2
4. Financial Services & Typical “Intensive” Problems
Wholesale Retail
Portfolio Construction Demand Forecasting
Price Forecasting
Targeted Marketing
Risk Calculation
Compliance Surveillance
Compliance Surveillance
Batch Processing Batch Processing
Proprietary and Confidential. Not to be distributed or reproduced without permission www.sungard.com/globalservices 3
5. Data & Compute Intensity in FS
What creates this Two flavors of intensity Two flavors of intensity
intensity? overload management
•Inter-machine parallelism &
•Algorithmic Complexity •Unacceptable elapsed run time distribution
•Time Series •Grids
•Calculus •Herculean efforts to pare down •BigData
•Graphs size of input data set
•Sharding
•Memory vs. Disk Arbitrage
•Semantic Inefficiency
•Inefficient physical data •Intra-machine hardware
structures acceleration
•Semantic starvation •GPUs
•FPGAs
•Other (Cray, etc…)
Proprietary and Confidential. Not to be distributed or reproduced without permission www.sungard.com/globalservices 4
6. Implementation Constraints in FS
Legacy Burden RDBMS Centricity
• Architectures, • Jailed semantics,
languages, tools, rigidity and specificity
systems, skills
Data Silos Cost of Change
• You don’t own that • Evolution is the only
data, I do answer
Proprietary and Confidential. Not to be distributed or reproduced without permission www.sungard.com/globalservices 5
7. Are We Close to Breaking Free of these Constraints?
Has a New Era Begun?
• Has financial crisis created a new era of information
transparency that requires a new approach to comply?
• Are the economics of “cloud computing” disruptive
enough to truly commoditize computing power?
• Will relational databases one day be seen as the
“stone age” of information management?
Or Are We Plagued By the Current State?
• The cost of maintaining simple data relationships, with
staunch determinism and specificity
• The effort required to design, store and consume richer
information semantics
• The disconnect between computing power and
information abstraction
Proprietary and Confidential. Not to be distributed or reproduced without permission www.sungard.com/globalservices 6
8. The BigData Thesis in FS
• New technology not burdened by the past, with non-linear economic
Why the allure? advantages at scale
Why the marketing • The markets need another source of wealth creation
hype?
How is innovation • By disrupting the past
marketed?
• Replacing 35+ years of relational R&D
What is hype? • Solution for semantically limited representation of data in current state
• Designed for more rapid scalability with “cloud” economics
What is reality? • Reducing the time it takes to process data to both formulate questions and
produce answers at scale
• Can BigData solve existing challenges in FS that couldn’t be solved before?
Theses • Can BigData create new solutions in FS, unachievable with earlier
incarnations of technolgy?
Proprietary and Confidential. Not to be distributed or reproduced without permission www.sungard.com/globalservices 7
9. Is BigData Enough or do we need Smart Data?
Smart Data (?)
• Better relationship
persistence ?
BigData (-)
• Better relationship utility
& consumption?
• Doesn’t address
semantic capability
• Deeper economic
shortfall
impacts?
• Legacy migration cost
• Non-linear return on
BigData (+) investment?
• Re-training effort
• Cheaper to store • Linear return on
investment
• Cheaper to index
• Cheaper to access in bulk
Proprietary and Confidential. Not to be distributed or reproduced without permission www.sungard.com/globalservices 8
10. The Opportunity in FS - A Second Look
Cloud
Economics
• Are Here
Competition Consumer
• Won’t Wait Data Success
Stories
• Are Noted
Digital Exhaust Social Media
• May Also Have • May Have Some
Some Signal Signal in all that
Noise
Proprietary and Confidential. Not to be distributed or reproduced without permission www.sungard.com/globalservices 9
11. Financial Services Data - Use it or Lose it!
Solving linear problems yields linear returns
• Solving for scale doesn’t automatically solve for semantics
• Moore’s Law millionaires are in their 50s and 60s now…do you
really want to bet on a 30 year-old strategy?
Go beyond the RDBMS, Go beyond BigData
• It’s the richness and practical utility of relationships that create
non-linear economic growth and technology returns
• Think Semantics, Think Relationships, Think Graphs
$$$
• A new era of transparency is creating a competitive new era of
monetized “digital exhaust”
• Farm relationships in the data you own or data you control
• Conquer the legacy constraints
Proprietary and Confidential. Not to be distributed or reproduced without permission www.sungard.com/globalservices 10
12. Maximize Your Return on Relationships!
Image Courtesy of @ValdisKrebs
Proprietary and Confidential. Not to be distributed or reproduced without permission www.sungard.com/globalservices 11
13. In Closing, The Challenge To Us All
Our Finite Collective Effort in Industry
& Academia
• Devote effort to game changing solutions with “non-
linear” returns
Our Solution Focus across Academia
& Financial Services
• Addressing adoption constraints should be just as
important as the implementation itself
Financial Services Data
• Use it or lose it!
Proprietary and Confidential. Not to be distributed or reproduced without permission www.sungard.com/globalservices 12
14. Contact Info
John Avery, Partner, SunGard Global Services
John.Avery@SunGard.com
On Twitter
@john_avery
Proprietary and Confidential. Not to be distributed or reproduced without permission www.sungard.com/globalservices 13