© 2007 IBM Corporation
Next Generation Grid: Emerging technologies poised to revolutionize financial markets
Kevin PleiterDirector – Global Financial Services Sector
© 2007 IBM Corporation2 FMA | October, 2007
Companies need to grow
aggressivelyby leveraging
differentiating technology
Financial Market Analytics
© 2007 IBM Corporation3 FMA | October, 2007
Financial Services Sector Trends, 2007-2015
Informed clients are increasing demands
Global integration is reshaping industries
Regulatory burdens are growing
Innovation is becoming imperative
FSS Industry Trends Challenges
Data Explosion
Managing Complexity
Managing Business Integrity
Collaboration & Partnering
Speed & Transparency
© 2007 IBM Corporation4 FMA | October, 2007
0
50
100
150
200
250
300
350
1997 1998 1999 2000 2001 2002 2003 2004
Agency Asset-Backed Corporate Derivatives
Mortgage-Backed Money Market Municipal Repo
Treasury/Sovereign Other
Electronification will target all asset classes…1
0
500
1,000
1,500
2,000
2,500
3,000
3,500
1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004
0
500
1,000
1,500
2,000
2,500
3,000
3,500
4,000
4,500
Average Daily Trade Volume (millions)(10 year CAGR 19%)
Security brokers and services personnel ('000)(10 year CAGR 3%)
Average Daily Trade Volumes vs. Headcount, 1994-2004 (Millions of Shares; Number of Employees (‘000), Volume per Employee)
Number of US and EuropeanElectronic Trading Systems1, 1997-2004
(Number of Electronic Trading Systems by Type of Market Supported)
Note: 1Totals may exceed the actual number of systems listed since some systems cover multiple product linesSource: The Bond Market Association; Reuters; SIA; IBM Institute for Business Value analysis
Volume of daily shares traded per employee (10 year CAGR 15%)
© 2007 IBM Corporation5 FMA | October, 2007
…causing agency margins to die and value to shift toward riskier activities
Note: 1Agency is risk free intermediation of a trade and principal is risk assuming intermediation of a trade Source: IBV profit map model (see slide 18 in this study)
1
Agency vs. Principal1, 2004 and 2015(US $ Billions, Margin Percentage)
Firms unanimously agreed that proprietary trading will be an increasingly important source of profit over the next ten years; regulators indicated they will not prevent it.
020406080
100120140160180200
Principal trading Agency execution0%
10%
20%
30%
40%
50%
60%
70%
2004
2015
RevenueShare50%
Financing and Structured Products Revenue,2004 and 2015
(US $ Billions)
6
18
1
4
0
5
10
15
20
25
2004 2015
Financing
Structured products
CAGR 11%
RevenueShare50%
Revenue Margins
020406080
100120140160180200
Principal trading Agency execution0%
10%
20%
30%
40%
50%
60%
70%
RevenueShare 30%
RevenueShare70%
RevenueShare
50%
RevenueShare50%
© 2007 IBM Corporation6 FMA | October, 2007
The data center energy, power and cooling challenge is a major hurdle for the ongoing success of Financial Services Grids
Power Use
35
30
25
20
15
10
5
0Chiller/
Cooling Tower
InformationTechnology
LightingSwitch/Gen
UninterruptiblePowerSupply
Power Distribution
Unit
ComputerRoom Air
Conditioner
Humidifier
Cooling systems Electrical and building systems
% o
f to
tal
da
ta c
en
ter
ele
ctr
icit
y u
se
© 2007 IBM Corporation7 FMA | October, 2007
These Business and Technology imperatives result in the following Technology trends in financial markets globally
Aggressive adoption of Model and Algo driven trading +
Demand for real time portfolio and risk management +
Massive build out of computing infrastructure
Next Generation Grid
Powered by Hybrid Computing and Stream computing
enabling the use of Data and Analytics as a Weapon
© 2007 IBM Corporation8 FMA | October, 2007
Specifically Financial firms analytic infrastructures are challenged to support ongoing business growth in five dimensions
Application Management
Workload Management
Data Management
Systems Management
Hardware Management
Inconsistent and unrepeatableInability to quickly deploy new apps
High cost modelLack of scalability and Performance
I/O bottle necks and geographic data dispersion issues not addressed
Data center CrisisCost, utilization, power and space
High cost, manual and reactiveNo automated / integrated feedback
© 2007 IBM Corporation9 FMA | October, 2007
The need for speed
Exploit advanced technology
to accelerate performance
© 2007 IBM Corporation10 FMA | October, 2007
Hybrid computing powered Grids incorporating the Cell Broadband Engine™ (Cell/B.E.)
Ten year collaboration with Sony®, Toshiba®
– Extended through 2012
IBM BladeCenter® QS20 system introducedSeptember 2006, QS21 in 2007 and QS22 2008
– “Just another Blade”
– Operates in standard BladeCenter® chassis
– Infiniband, Gig-E and 10GigE support
– Runs Redhat Linux and a host of other standard industry applications
Cell/B.E. Highlights
– Significant performance advantages over traditional systems
– 64-bit Power™
– Streaming Processors
– Low power consumption
– Real-Time
– Security Enabled
© 2007 IBM Corporation11 FMA | October, 2007
Cell use cases in the Financial Service Sector
Various domains
– Fixed Income and FI Derivatives, Equities and Equity Derivatives, Risk Management, Portfolio Trading, Algorithmic Trading and Analytics
Functional areas
– Pricing of Options, Swaps, and Futures
– Interest Rate Modeling, CDO, MBS
– Path Dependent Calculations
– Risk evaluations – VaR, EPE, PCE, etc
– Stream Processing - market and analytic data
© 2007 IBM Corporation12 FMA | October, 2007
Typical Algorithms being deployed in Cell grids
Algorithms typically used include
– Analytical and approximate (crude) solutions through to Black-Scholes
– PDE solvers (e.g.,Crank-Nicolson)
– Householder transformation, Cholesky decomposition, SVD, PCA
– Monte Carlo Simulations
– Statistical Analysis
Standard Library routines being called
– BLAS, LAPACK, FFT
– Random Number Generation (Sobol, MT, Box-Muller, Moro’s Inversion)
– Math Library’s (R, NAG)
© 2007 IBM Corporation13 FMA | October, 2007
Rapidly growing Application Development, Math library and Grid ecosystem for Financial Services
BLAS – DPFFT – DP
Cell Blade Hardware (QS22)
BLAS – SP
Financial Analytic Libraries and ISV applications NAG, Encirq, …
FFT – SP
FSS Stream Workloads ISVs - Encirq
IBM – SystemS IBM – Websphere Front Office
RHEL 5.1, 5.2 with full Cell support
Application Layer
Base SDKProgramming framework
LANL-centric
Hardware / Firmware
Libraries
SP – Single PrecisionDP – Double Precision
1H 2008 Interoperability/Deployment (Windows + Java)
Grid Schedulers via Partners
SOMA CAB – Analytics workstation
LAPACK – SP/DP
Multi-core Cell SDK (SPE Mgmt) + ALF + DACs + Compiler Roadrunner framework
zCell – Back office offload + new workloads
Platform Computing
Proprietary FSS Analytic TradingAnd Risk Applications
IBM – Dynamic Application Virtualization (DAV)Partners
1Q’09 SLES – Cell extensions
RNG/MKL
Available
© 2007 IBM Corporation14 FMA | October, 2007
x86 Linux master cluster
Cell/B.E. accelerator
cluster
Hybrid Supercomputing Reference
Targeting 1.4 petaflop peak, 1.0 petaflop sustained performance
© 2007 IBM Corporation15 FMA | October, 2007
For more information download “How much is a microsecond worth?” at www.ibm.com/financialmarkets
© 2007 IBM Corporation16 FMA | October, 2007