17 June 2019© MARKLOGIC CORPORATION

Global Markets Beyond Legislations

Edwin Raubenheimer, IT Solutions Engineer

Building ABN AMRO’s Agile Data Foundation

ABN AMROGlobal Markets Beyond LegislationsBuilding a data driven and agile foundation

Edwin RaubenheimerSolution Engineer – Global MarketsMay 15th 2019

EDWIN RAUBENHEIMER

ABN AMRO Global Markets

Solution Engineer

Data and E-Cluster

9 Years - Banking – Data, Finance, Risk and Regulations19 Years - Software Engineering

Experience

17/2019

OUR JOURNEYBuilding a Data Driven and Agile Foundation

GLOBAL MARKETSNot about shouting

WHAT DO WE DO IN GLOBAL MARKETS?CUSTOMER FACILITATION

Hedge our daily position against the exchange or other parties to mitigate risk from marketfluctuations

• FX (Spot, Forward)• Fixed Income (Bonds, CP etc)• Equities (Shares of stock)• Commodities (Oil, etc)• Interest Rate Derivatives• Loan Deposits (Money Market)

Equity Capital Market

Issuing/Listing Shares(Primary Market)

Research

Generating Capital for companies via Equities (Shares)

Selling Existing Shares(Secondary Market)

Debt Capital Market

Creating, structuring and execution of Debt

Instruments

Advisory Services

Advising, generating or structuring capital-raising solutions via Debt

Instruments (Bonds) for corporate and financial institutions

Primary Market Secondary Market ABN AMRO

Financial InstitutionsCorporates

Public Institutions

Private and Retail Clients

Investment and Hedging Customer Risk

Options Swaps

Product Offerings

Hedging Bank Risk

GLOBAL MARKETS DATA TYPES

TRADE& Post Trade

Market Static

Instrument Static

Issuers and Guarantors

Request for Quote

or Sale

Trading Venues

Orders

Over the Counter

Market Data Front Office Data Back Office Data Functions Data

Etc..

Dealer Network/platform

or Voice

Exchange/ Trading facilities

Pricing Models

Common Market Pricing

Data and Curves

Streaming Pricing Data

Market Snapshots

per venue / instrument

Time and Sales per venue /

instrument

Confirmation and

Settlementsinstructions

Risk

Market and Credit

ExposurePositions

and Valuations

Regulatory

Payments

Finance and Risk

Product Control

Models

THERE IS A NEED FOR SPEEDTHE CHANGING WORLD IN WHICH WE OPERATE – NEW WAY OF THINKING

New/Changing technology

KEY CRITERIA FOR WHERE WE WANT TO MOVE

No big design upfront Must enable a gradual

implementation of data management in GM, starting from simple store-and-forward, growing towards a sophisticated data management platform.

Query off-loading The operational systems

provided a golden copythat enables off-loading of all requests for GM data, reducing load on expensive operational systems.

Easy upload / raw data Data ingestion must be based

on raw data (preferably in XML or JSON), without modification (no filtering / transformation), and using any technology (e.g. messages, files, ESB, APIs).

Central data modelling The ODS must be optimised

for achieving full control on the complex governance of how data is modelled and published.

Multiple delivery models Data delivery must support

multiple models (i.e. rawbut also modelled data) depending on knowledge of the consumer.

Lineage in published data Data (i.e. documents)

published by the ODS always include the data lineage back to the source.

Security/Rules based Regulatory compliance (e.g.

GDPR) is ensured by applying rules to queries (e.g. retention, access) that are stored as documents as well.

No direct interfaces Data providers deliver

once to the ODS, and not directly to consumers. The ODS controls and manages the distribution of data to all data consumers.

GLOBAL MARKETSBecoming more efficient…

TECHNICAL DECOUPLE

FUNCTIONAL DECOUPLE

14

DATA MOVEMENT PATTERNS USED IN GLOBAL MARKETS

PROVIDING ENTERPRISE

Streaming Pipeline

ESB / API Gateway

GLOBAL MARKETS CONSUMING GLOBAL MARKETS

Raw Data Store

PROVIDING GLOBAL MARKETS

CONSUMING

Analytics & operational use cases

MI/BI Reporting

Enterprise use cases (e.g. reporting)

IntegratedData Store

Application

Application

Application

Other Bounded Context(s), e.g. Payments, Credits

Application

Application

ESB/

API

Gat

eway

Application

Streaming Pipeline Application

ApplicationApplication

Legend: Raw data Harmonised data Raw dataevent stream

Harmonised dataevent stream

Global Markets

HarmonisedData

Central Metadata

Raw DataStore

Global Markets

HarmonisedData(IDS)

People Systems Processes Regulations

Catalogue

Order and Trading Systems

Financial Systems(Positions)

Pricing Engines

Market Sources

Reference SourcesCollectionsDocuments

Attributes / Tags

EnvelopeMeta data

Triples

Domain ModelEntities and Relationships

Attributes

DatesCalculationsCanonical

XMLJsonCSV

A

B

C

Modelled Data

Triples

Requests

Ingest

Raw Documents Harmonised Data

Information

Delivery

Hindsight View

Data Consumers

Reporting

Finance andRisk

Regulatory

Product Control

Business

Operations

DescriptiveAnalytics(What?)

Diagnostic Analytics(Why?)

OLAP

Data Analytics

Business Intelligence

Quantitative Models

MachineLearning

AI/CIDeep

Learning/Mining

Self ServingExplanatory

ModelsPredictiveAnalytics Making it happen

Data Orchestration

Monitoring

04 LESSONS LEARNED SUMMARYOur experience and challenges

DO’S AND DONT’S

EXPERIMENT EARLY, GET PATTERNS THAT WORK FOR YOU. ML IS TOO POWERFUL

NEW PARADIGM, COMMUNICATE OFTEN, WHY? WHAT?

TRAINING… TRAINING… BOTH TECHNOLOGY AND MIND SET. INVEST…

DATA IS CHIEF, UNDERSTAND YOUR DATA

USE TOOLS TO ACCELERATE DEVELOPMENT AND MANAGEMENT

DO’S

DO’S AND DONT’S

DON’T CREATE A BLACK BOXCONNECTION TO COMMERCIAL

VISUALISATION TOOLS ARE LIMITED

DON'T EXPECT ML TO INCREASE THE SPEED OF DELIVERY IN THE SHORT TERM

DONT’SDON’T USE EXCEL FOR SOURCE TO

TARGET

DON'T BUILD SOLUTIONS ON EXPERIMENTS, START CLEAN

DON’T DUMP RAW DATA INTO ML AND EXPECT MAGIC

Thank You