iet nw region - payment hub design
DESCRIPTION
A presentation to the UK IET North West Region on design issues for Payments Hubs.TRANSCRIPT
Payment Hub Design(…or how IT helps solve the credit crunch)
Gary Farrow
IET / BCS 7th Sept 2010
2
Agenda
• Payment Hub Overview
• Capability Model
• Payments Hub Spectrum
• Design Issues– Service Granularity
– Technology Selection
– Canonical Data Zone
• Traversing the Spectrum
• Liquidity Problem
• Review
3
Agenda
• Payment Hub Overview
• Capability Model
• Payments Hub Spectrum
• Design Issues– Service Granularity
– Technology Selection
– Canonical Data Zone
• Traversing the Spectrum
• Liquidity Problem
• Review
4
Payment Hub Architecture
Core Banking
Ledgers
Legacy
Core Banking
Ledgers
Finacle
Payments Hub
Gateways
Payments
Payments Payments
Schemes/3rd Parties
Channels
Payments
BACS FPS CCCCLLink/
ATM/
VISA
Business
Partners
Characteristics
• Integration Services– Routing
– Transformation
– High throughput
• Payments Business Services– Liquidity monitoring
– Liquidity information provision
– Scheme cap monitoring
– Point of control of other services
• Payments Process Execution– Payments process orchestration
– State management
– View of payments state
Decouples ledgers from mechanics of payments processing
5
Payments Hub Justification
• Reduces integration complexity
– SxL problem reduced to S+L
– Complexity reduction since not all schemes connect to all ledger
– Complexity increase due to financial controls
• Supports ledger migration roadmap
• Supports business strategy of acquisition
• Architectural simplicity
– Single payments process
– Use of a single ‘canonical’ data form (ISO20022)
– Master data management issue reduced (CIF, ISCD)
Schemes
Accounting Ledgers / Product Systems
Payments Hub
BusinessLine
Treasury
Mortgage
NewPlatform
Legacy
Inter GL
AcquiredBank
Total Number of Ledgers
6
7
8
9
2008 2009 2010 2011 2012
Cards
Case for a Payments Hub is compelling
6
Agenda
• Payment Hub Overview
• Capability Model
• Payments Hub Spectrum
• Design Issues– Service Granularity
– Technology Selection
– Canonical Data Zone
• Traversing the Spectrum
• Liquidity Problem
• Review
7
Enrichment Services
Destination Determination
Router
Liquidity Monitor
Flow Control State Machine
Fraud Service(Façade)
Transformation
AML Service(Façade)
Payments Capability Model
Excess Management
Funds ControlPayment Message
Origination
Account Validations
Account Posting
Settlement Control
Accounts
Mandate Management
Intelligent Routing
RepairServices
Almanac
DiaryServices
Capabilities are apportioned but how ?
8
Agenda
• Payment Hub Overview
• Capability Model
• Payments Hub Spectrum
• Design Issues– Service Granularity
– Technology Selection
– Canonical Data Zone
• Traversing the Spectrum
• Liquidity Problem
• Review
9
Payments Hub Spectrum
•Routing•Transformation
•Payments Almanac•Diary management•Enrichment
PureMiddleware
Payments ‘Engine’
Increasing Functional Richness
•Record Validation•File validation•Payments repository•Bank reconciliation data
•Intelligent scheme routing
•Mandate management•Payment origination•Customer reconciliation data
•Payments state management•Coordination of ‘value-add’ services
•Fraud•AML•Auto-Repair
•End-end process control•Store & Forward
Position on Spectrum
10
Agenda
• Payment Hub Overview
• Capability Model
• Payments Hub Spectrum
• Design Issues– Service Granularity
– Technology Selection
– Canonical Data Zone
• Traversing the Spectrum
• Liquidity Problem
• Review
11
Service Granularity
• Service granularity refers to the nature of the service interactions between Hub and Ledgers
• Scenario shows a simple payment pattern
• Granularity affected by:– Capability of Hub
– Nature of product systems
• Modern banking packages have rich capability– Coarse grained services
• Legacy systems provide disaggregated services– Fine grained services
Coarse Grained
Fine Grained
12
Technology Selection
Middleware Framework Engine
Pure middleware product supporting messaging/ transformationStateless
Pre-build sub-flows, Payments Repository, Scheme transformations, Stateless/fulGrey box component
Pre-built scheme levelprocessingBlack box component,Stateless/ful
Low technology costRelative low cost implementation
Standard based, Modular
Low implementation cost if‘out of the box’
Building enhanced functions is costly
Customisation can still be extensive
Customisation cycle slowBlack box componentHigh product cost
IBM MQ / Message Broker, Oracle (BEA),
IBM Enterprise Payments Platform, Clear2Pay, Dovetail
Fundtech
Payments Hub Spectrum
13
Canonical Data Zone
• Canonical Data is a standardised representation of the key data– ISO2002 credit / debit transfers
– Base24 real-time payments
• Canonical Data Zone is the architectural layers that make use of such data– Objective: Reuse of common
processing steps
• Hub should process payments using canonical data
• Ideal world target system should also process payments in the same canonical format
The Canonical Zone
SchemeSpecific Format
SystemSpecific Format
Canonical data standard is best defined as an extended version of a recognised standard
14
Canonical Data Design Issues
• Legacy integration is a neat example– Three necessary transformations
– BACS-Canonical-Legacy
• Package integration– Core banking system vendors offer
scheme modules that already use scheme specific format
– Transform BACS-Canonical-BACS (-Internal)
• Package design principle– Use ‘out of the box’
– Scheme specific interfaces
– Driver against Hub
• Hub design principle– Integration ‘heavy lifting’
– Minimise transformations
– Ideal Canonical Data is an enriched standard form
– Driver against package principle
Universal Scenario• Merger / acquisition• Migration• Different systems for
different product types
BACSSTD18
LegacyFormat
BACSSTD18
The Canonical Zone
15
Architectural Tension
Package Vendor
Client
System Integrator
Low costSpeed to marketFuture proofing
Module SalesCustomisation developmentCountry Specific Modules
Package principleArchitectural Elegance / flexibilityDeliverabilityMinimise risk
16
Agenda
• Payment Hub Overview
• Capability Model
• Payments Hub Spectrum
• Design Issues– Service Granularity
– Technology Selection
– Canonical Data Zone
• Traversing the Spectrum
• Liquidity Problem
• Review
17
Moving through the Spectrum
•Solution Governance•Policy compliance•Paradigm alignment
Pure Middleware Payments ‘engine’
•Risk & Compliance
•Payments Services Directive
•Product (re-) Selection•Supplier re-selection
Position on Spectrum
Implications of moving through the spectrum are significant
18
Agenda
• Payment Hub Overview
• Capability Model
• Payments Hub Spectrum
• Design Issues– Service Granularity
– Technology Selection
– Canonical Data Zone
• Traversing the Spectrum
• Liquidity Problem
• Review
19
Liquidity Monitoring
BACS
-40
-20
0
20
40
60
CCCCL
-150
-100
-50
0
50
ATM/LINK/VISA
-100
0
100
200
FPS
0
50
100
150
Net Liquidity Position
-100
0
100
200
Liquidity varies– Intra-day
– Inter-day
– Monthly cycles due to corporate bureaux services
Settlement risk
– Distorted due to Agency Banks
– E.g .Northern Rock
Treasury
– Monitors Cash / Liquidity position periodically
– Plans for Cash Management based on known liquidity position
– Optimises financial investment and borrowing
£(Millions)
Liquidity monitoring and information services are required for pro-active management of liquidity
20
Consequences of Poor Liquidity Management
One lump or two?
Payments Schemes
– Deferred Net Settlement
– Underwritten by the Bank of England
– Require daily settlement payments via Real Time Gross Settlement
Once settlement figure are known scheme participating Banks have ~20 minutes to make a CHAPS payment
Missing a settlement payment is not desirable
– Repetition will result in scheme expulsion
– CxO will be ‘invited for tea’ at the Bank of England
Payment Hub is the architectural component to provide liquidity monitoring services
21
Agenda
• Payment Hub Overview
• Capability Model
• Payments Hub Spectrum
• Design Issues– Service Granularity
– Technology Selection
– Canonical Data Zone
• Traversing the Spectrum
• Liquidity Problem
• Review
22
Review
• Payment Hub Overview Understand what a Payments Hub is Benefits it can provide
• Capability Model• Introduced the Payments Hub Spectrum• Design Issues in placement on the Spectrum
– Service Granularity– Technology Selection– Canonical Data Zone
• Issues in traversing the Spectrum
• Liquidity Problem– How a Hub can monitor liquidity
• Solved the Credit Crunch? Not quite….. But by designing Payments Hubs for our major banks I hope IT make a small but important
contribution
23
Thank You
Merci
Grazie
Gracias
Obrigado
Danke
Japanese
English
French
Russian
GermanItalian
Spanish
Brazilian Portuguese
Arabic
Traditional Chinese
Simplified Chinese
Thai