enterprise information management and business...
TRANSCRIPT
Enterprise Information Management and Business Intelligence Initiatives at the Federal Reserve
XXXIV Meeting on Central Bank Systematization
7-9 September 2011
Kenneth Buckley
Associate Director
Division of Reserve Bank Operations and Payment Systems
Agenda
• Background and challenges
• Enterprise Information Management (EIM)
responses and strategies
• EIM at the Federal Reserve Bank of New
York (FRB NY)
• Business Intelligence initiatives
• General outlook and next steps
2
Background and challenges - The Federal Reserve System is comprised of 12 geographically dispersed banks
3
Regionally Located Business Lines:
•Research Departments at all Banks and the Board
•Markets Group at FRB New York
Centrally operated -
national mainframe, server,
network data centers
Regional IT Operations:
•Application
development
•Local server operations
•Local IS operationsTo varying degrees, these separate organizations all
had largely separate data efforts and supporting
infrastructure in place or under consideration
• The Federal Reserve collects and uses a wide variety of information from internal and external
sources to support its central bank mission to:
evaluate the condition of financial institutions;
forecast economic conditions
execute monetary policy; and
operate domestic payment and US Treasury services.
• Data supports monitoring of market risks
banking failures
business cycles & market conditions
liquidity and credit constraints
• In the mid-1990s, significant changes occurred in financial markets
products and institutions became more complex
markets consolidated and cross-border economies became interdependent
repeal of the Glass-Steagall Act removed separation between investment and depository banks
• 2008 Financial crisis
broad investor distrust of virtually all forms of private credit - especially complex credit products
largely unforecasted by standard market indicators
increased risk from decline in asset prices and sharp and sudden intensification and rapid global
spread
international pressure on central banks to act quickly
Background and challenges - Structural changes in market and banking environments
4
Background and challenges - Gaps in data and knowledge management tools
• Central bank response
provision of short-term liquidity to banks and other financial institutions
currency swap agreements with other central banks to assist in providing liquidity in their jurisdictions
• Operational challenges resulting from the financial crisis
expanding the balance sheet to account for new asset classes from new lending facilities
understanding complex financial instruments
managing the level of lending demand and complex collateral instruments
need for enhanced risk analytics
diverse information sources for asset valuation and risk management
business processes requiring external partners
• Problems with data management in the Federal Reserve became more visible:
data sources and repositories were disparately organized and managed
Individuals could create, acquire, and store data without much standardization
ownership blurred and silos emerged leading to constraints on sharing and controlling data
data not managed consistently and overall quality suffered or was hard to determine – often no “single
point of truth”
inadequate and inconsistent enterprise guidelines, policies, governance, and controls
Many efforts local in scope and focus, resulting in duplication of data, limited availability, and lack of
clear responsibility /ownership for data integrity, consistency, quality, availability
lots of data available, but that did not commensurately improve intelligence or insight
5
Data
Background and challenges – inability to meet the service demands from knowledge workers
Information KnowledgeActions & Outcomes
Integrate Access Report Analyze Understand Decide Act Produce Measure
Learn
Ability to Create Business Intelligence
Ability to leverage Information Assets
6
EIM responses and strategies
1. Create and evolve Federal Reserve
Enterprise Information Management
(FedEIM) framework
• Comprehensive data management strategy
comprised of integrated data management
disciplines
• Led by centralized Enterprise Architecture
function
• Primary objectives:
1. Improve operational efficiency, promote
transparency, and (most importantly)
help enable better business insight and
decision making
2. Be responsive to changing regulatory
priorities, business conditions, and
industry directions
• Need to structure, describe, access, and
govern data as assets across functional
areas, regardless of organizational or
technical boundaries
• Establish a business view of the technology
framework
7
EIM responses and strategies
2. Proceed with data governance and organizational alignment efforts
• Evolve organizational thinking and approach – effective data management is an
organizational discipline, and not just a technical issue
• Establishing roles, responsibilities, and organizational structures that can adapt to changing
conditions
3. Rationalize and coordinate data procurement and tools acquisition
• Processes and organization to support:
efficient / shared acquisition of data
data management tools
• Ways to better identify and manage user demands
8
EIM responses and strategies
• General strategy is to rely on and foster a combined regional and local approach
evolve over time to more appropriate mix
realize benefits and economies of national approach without losing value of allowing local
flexibility and innovation
• Numerous regional efforts proceeding across Federal Reserve System today under the
guidance of national FedEIM framework
• Business functions at FRB New York and Board of Governors are two of largest
“consumers” of data today, and both have taken significant steps to advance data
management capabilities and coordination
• FRB NY’s efforts currently focused on
establishing a formal Data Management Office (DMO);
enhancing and institutionalizing standard processes and related governance;
improving flexibility and responsiveness to changing business needs;
participating (with rest of FRS) in efforts to rationalize tools and data sources
9
EIM at the FRB NY - Strategy
• Situational Analysis
assessed the state of data assets and corporate data management in Markets function to
determine future structures needed to support strategic data management needs
concluded that a dedicated organization with a strong governance structure was needed to
execute the future state plan and support ongoing operations
• Established the Data Management Office (DMO) as the “organizational glue” bonding
business, operations, and technology
drive the „Data Management strategy‟ (long-term while supporting day-to-day deliverables)
establish data governance; identify data ownership; engage data stewardship
establish “authoritative sources” of analytics, reference and market data
define policies to govern data acquisition, storage and access
• Key tactical DMO objectives
10
improve collaboration between Market
Operations, Bank Supervision, and
Economic Research functions
rationalize relationships with market data
vendors and service providers
collaborate with Markets Technology to
implement the necessary applications
provide ongoing support for data
acquisition, data storage and data delivery
systems
EIM at the FRB NY - DMO Functional Organization Structure
11
• The DMO is a BUSINESS function liaising between “the consumers” and the “suppliers”
• The DMO is made up of Market Data subject matter experts, Data Management Business Analysts, Project Managers and QA specialists dedicated to the delivery of data management solutions
Consumers
Data Management Office (DMO)
Data Providers
Services
DMO• Market Data Specialists
• Data Management BAs
• Project Managers
• QA Specialists
• Market Data Vendor
Admin Specialists
(contracts)
RM’s
Domestic Markets
Treasury Markets
Domestic MM & Reserve
Disc Window & Collateral
Market Analysis
Foreign Exchange
Facilities Market Data Terminal Providers
Internal sourcing
Market Data Feed Providers
Markets Business Technology
Automation Group Technology
Automation Group QA
Market Data Service Providers
Business driven
Technology enabled
Operations supported
• Acquisition of
“golden” copy
data from
external sources
• Acquisition of
“golden copy”
data from
INTERNAL
operations
• Control and
governance over
purchases
• Require “Data
Review” for all
new systems to
ensure proper
use of shared
repositories
• Extract and
transform data
from internal and
external sources
that fit business
needs
• Understanding
of system data
flows
• Perform quality
checks,
exception
handling and
data remediation
• Identify existing
“authoritative
sources” of data
• Define future
data repositories
• Define the
strategic
repository
infrastructure
• Determine data
ownership -
stewardship
• Implement
proactive
processes to
maintain quality,
timeliness and
completeness
data (Data
Quality)
• Provide
Centralized
“Customer
Service” to
consuming
systems
• Define Data
Standards (Data
Models, Data
Dictionary,
Metadata, etc.)
• Define the
methods for
quick and
organized data
access (i.e.:
Portal; Excel; BI
tools; SAS;
Access)
• Partner with
consuming
applications to
ensure data is
“fit for purpose”
• Ensure ease of
access to data
(provide end-to-
end solutions)
Data Creation
& AcquisitionProcessing
Data
“Factories”Maintenance
Data
DistributionConsumption
Governance
Data Management is a multi-faceted function composed of a series of “component disciplines” that track to the Data Management Lifecycle
EIM at the FRB NY - DMO Scope
12
EIM at the FRB NY– Data/Tools Acquisition
Rationalizing and coordinating data procurement and tools acquisition
• Established a cross-bank committee with representatives from Markets, Supervision,
Research, and others to monitor and track all market data purchases
• Developed and now maintain a complete inventory of all purchased data into an
industry standard market data procurement tool (FITS)
• Worked with market data vendors to strategically identify and acquire data,
collaborating across functions and with other Reserve Banks and the Board of
Governors
• Participating in enterprise-wide efforts to pilot and standardize data management and
business intelligence tools (e.g., Informatica and FITS)
13
Business intelligence initiatives
• As a result of its experience during the financial crisis, the Federal Reserve is committed to
improving the responsiveness and flexibility of its BI tools for analysis and decision making.
• Experience with BI tools manipulating large data sets has yielded the following lessons learned:
traditional software products do not maximize the performance of the hardware
traditional methods for statistical and computational analysis of large datasets can be expensive and
challenging
transferring large amounts of data from one location to another can be slow
issues related to managing large data are analytical and not transactional (parallel vs. serial)
large data issues are a hardware/software problem
• Numerous BI initiatives underway today are evaluating practical applications of several different
tool sets, including,
GreenPlum DW appliance for the RADAR platform (provides large DB consumer loan analytical capabilities
using massively parallel processing technology)
TIBCO Spotfire, intended as a general purpose and provides powerful self service tools for data access and
manipulation, reporting and analysis, and collaboration
14
BI initiatives - RADAR
RADAR (Risk Assessment, Data Analysis, & Research)
• RADAR is a data warehouse (as well as separate securities evaluation service) used by Bank
Supervision as well as Research functions
broad array of US consumer credit data - credit cards, auto loans, student loans, mortgages
currently over 9 TB and 350 million records
• Generate and view data by various parameters, maps, charts Geography, various demographics, time periods, or loan status.
Options for customizing the views of data, including the ability to develop maps or charts
• Developed by staff in Philadelphia and Kansas City Reserve Banks and provides both preset and
ad-hoc query capabilities
• Platform is high performance clustered Linux on Apache/Tomcat web servers, coupled with
Greenplum database and 15 commercial/open source software analysis tools (SAS, Matlab,
C/C++, etc.), as well as various in-house developed query and reporting tools
• While the data warehouse is mainly used for bank surveillance purposes, it has also proven useful
in the Fed's community development initiatives
15
BI initiatives - RADAR
16
Mortgage
Delinquencies
by County: 90+
Past Due,
Foreclosure, All
50 States,
201009
BI initiatives – National Cash Data Warehouse
• Provides ability to aggregate and view depository institution and Reserve Bank
currency activity
• The NCDW, along with the associated data analysis and reporting tool, provides
important functionality to the Cash business to:
monitor and enforce certain business rules,
support billing for customer activity that exceeds rule-based parameters, and
support analyses that serve as a loss control function for the business.
• Migrating to a Microsoft SQL 2008 platform with Business Objects reporting tools
17
General outlook and next steps
• EIM and BI will continue to be high priority objectives in our IT and business line
strategies. These objectives will influence other strategic initiatives such as
collaboration, mobility, and analytic tools for knowledge workers.
• Tactical next steps include:
building on successes and applying lessons learned, continuing to support locally driven DM
efforts, and evolving the enterprise wide programs that fit needs and realities
further improve analytic capabilities
creating standard directory-based interfaces for analysts using tools like Spotfire that
can free them from concern with physical DB and technologies used
more complex and larger volumes of data
further defining policies and standards for data governance, and supporting their broad
implementation and acceptance as a fundamental (and typical) aspect of how we conduct
our central bank activities
rationalizing toolsets and technologies and, by applying standard reference architectures,
transition over time to platforms that optimize performance and reduce operating costs
current initial efforts underway include moving to consolidated operating environments
for Informatica
establishing enterprise wide standard taxonomies or metadata that cross current
organizational and/or business area boundaries
improving specialized skill sets associated with data roles and BI technology that can reduce
time to market for delivering solutions
19