SlideShare a Scribd company logo
Hadoop@Visa  Joe Cunningham  Technology Strategy and Innovation October 2, 2009
Agenda 1 Visa 2 Value-Added Information Products 3 Hadoop@Visa – research results VisaNet – an overview
Keeping the Brand Promise Zero Liability go Security Flexibility
Keeping the Brand Promise 3 Global Acceptance Reliability Zero Liability go Security Flexibility Guaranteed Payment
VisaNet Today  VisaNet is the world’s largest, most advanced  payments network. 28 million acceptance locations 130 million authorizations per day 1,600 endpoints Processes transactions in  1 second 16,300 financial institutions Less than  2 seconds unavailability per year Processes in  175 currencies 1.4 million ATMs fast  secure  flexible  reliable  scalable
1 VisaNet – an overview 2 3 Value-Added Information Products Hadoop@Visa – research results Agenda
Visa Processing Architecture  Security/Access Services Message File Web VisaNet Services Integration Common Data Infrastructure Authorization Clearing & Settlement Information Disputes Risk
Value-Added Information Products Information Services Client  Portfolio Analysis Visa Incentive Network Accountholder Transaction Alerts Account Updater Tailored Rewards Risk Management Services Authentication  Advanced Auth Verified by Visa Address Verification Account Monitoring Fraud Detection Compromised Accounts Consumer Alerts Encryption Dynamic Key Exchange Dynamic Cryptograms
1 VisaNet – an overview 2 Value-Added Information Products 3 Hadoop@Visa – research results Agenda
Research Lab Setup Hadoop Systems VM System  Relational Database Custom Analytic Stacks Encryption Processing Management Stack Hadoop #2 ~300Tb / 28 nodes Hadoop #1 ~40Tb / 42 node
Risk Product Use Case Create critical data model elements, such as keys and transaction statistics, which feed our real-time risk-scoring systems Hadoop Cluster Research Sample (Synthetic Transactions) 500 million distinct accounts 100 million transactions per day 200 bytes per transaction  2 years: ~ 73 billion transactions, ~ 36 TB One Month    13 minutes! Input: Transactions Merchant Category   Country / Zip   Output: Key & Statistics MCCZIP Key Statistics related to account, transaction type, approval, fraud, IP address, tx. counts
Financial Enterprise Fit Some key questions we are researching…  What will the Hadoop Solution Stack(s) look like? File System  Transaction Sample System Relational Back-End Analytics Processing Internal vs. External cloud How do I get data into a cloud in a secure way? How does HSM and security integration work in Hadoop? What are the missing pieces?
Why Hadoop@Visa? For our information products, we see promise in combining Visa’s data analytics capability with the power of Hadoop Analyze volumes of data with response times that are not possible today… …  and apply analytic models to individual client, not just client segment But, need to answer questions relating to stack configuration and support…  …  and, eventually must fit with our financial enterprise paradigm: fast  secure  flexible  reliable  scalable
Keeping the Brand Promise Zero Liability go Security Flexibility
Keeping the Brand Promise 3 Global Acceptance Reliability Zero Liability go Security Flexibility Guaranteed Payment

More Related Content

PDF
MuleSoft Event Driven Architecture (EDA Patterns in MuleSoft) - VirtualMuleys63
PDF
Best Practices for Streaming IoT Data with MQTT and Apache Kafka®
PDF
Kubecon 2023 EU - KServe - The State and Future of Cloud-Native Model Serving
PDF
Kafka 101 and Developer Best Practices
PDF
MIPM PCo to Kafka Faurecia SAP co-innovation at Hannover Messe 2017
PPTX
The Top 5 Apache Kafka Use Cases and Architectures in 2022
PPTX
Asynchronous processing in big system
PDF
The delta architecture
MuleSoft Event Driven Architecture (EDA Patterns in MuleSoft) - VirtualMuleys63
Best Practices for Streaming IoT Data with MQTT and Apache Kafka®
Kubecon 2023 EU - KServe - The State and Future of Cloud-Native Model Serving
Kafka 101 and Developer Best Practices
MIPM PCo to Kafka Faurecia SAP co-innovation at Hannover Messe 2017
The Top 5 Apache Kafka Use Cases and Architectures in 2022
Asynchronous processing in big system
The delta architecture

What's hot (20)

PDF
Event Driven Architecture (EDA) Reference Architecture | Anbu Krishnaswamy
PPTX
Flink vs. Spark
PDF
Benefits of Stream Processing and Apache Kafka Use Cases
PPTX
Data Lakehouse Symposium | Day 4
PDF
Real-Life Use Cases & Architectures for Event Streaming with Apache Kafka
PDF
Delta Lake OSS: Create reliable and performant Data Lake by Quentin Ambard
PPT
Mule with workday connectors
PDF
A Reference Architecture for ETL 2.0
PPTX
MuleSoft Architecture Presentation
PDF
Demystifying Service Mesh
PPTX
OCI Overview
PDF
Data Streaming with Apache Kafka in the Defence and Cybersecurity Industry
PDF
Real-time Adaptation of Financial Market Events with Kafka | Cliff Cheng and ...
PPTX
[DSC Europe 22] Overview of the Databricks Platform - Petar Zecevic
PDF
When NOT to use Apache Kafka?
PDF
Kafka for Real-Time Replication between Edge and Hybrid Cloud
PDF
Event-driven Architecture
PDF
Oracle Cloud Infrastructure – Storage
PDF
MuleSoft Sizing Guidelines - VirtualMuleys
PDF
Lecture 6: Infrastructure & Tooling (Full Stack Deep Learning - Spring 2021)
Event Driven Architecture (EDA) Reference Architecture | Anbu Krishnaswamy
Flink vs. Spark
Benefits of Stream Processing and Apache Kafka Use Cases
Data Lakehouse Symposium | Day 4
Real-Life Use Cases & Architectures for Event Streaming with Apache Kafka
Delta Lake OSS: Create reliable and performant Data Lake by Quentin Ambard
Mule with workday connectors
A Reference Architecture for ETL 2.0
MuleSoft Architecture Presentation
Demystifying Service Mesh
OCI Overview
Data Streaming with Apache Kafka in the Defence and Cybersecurity Industry
Real-time Adaptation of Financial Market Events with Kafka | Cliff Cheng and ...
[DSC Europe 22] Overview of the Databricks Platform - Petar Zecevic
When NOT to use Apache Kafka?
Kafka for Real-Time Replication between Edge and Hybrid Cloud
Event-driven Architecture
Oracle Cloud Infrastructure – Storage
MuleSoft Sizing Guidelines - VirtualMuleys
Lecture 6: Infrastructure & Tooling (Full Stack Deep Learning - Spring 2021)
Ad

Viewers also liked (11)

PPTX
Visa net the power of digital currency
PPT
Hw09 Hadoop Applications At Yahoo!
PDF
Threatmetrix スレットメトリックス・グローバル共有インテリジェンスによる VISA 顧客認証サービス(VCAS)のサポート
PDF
Fraud Detection Using A Database Platform
PPTX
Analysis of-credit-card-fault-detection
PPSX
PDF
Big Data Analytic with Hadoop: Customer Stories
PPT
Seminar Presentation Hadoop
PPTX
Hadoop introduction , Why and What is Hadoop ?
PDF
Hadoop Overview & Architecture
 
PDF
How to Become a Thought Leader in Your Niche
Visa net the power of digital currency
Hw09 Hadoop Applications At Yahoo!
Threatmetrix スレットメトリックス・グローバル共有インテリジェンスによる VISA 顧客認証サービス(VCAS)のサポート
Fraud Detection Using A Database Platform
Analysis of-credit-card-fault-detection
Big Data Analytic with Hadoop: Customer Stories
Seminar Presentation Hadoop
Hadoop introduction , Why and What is Hadoop ?
Hadoop Overview & Architecture
 
How to Become a Thought Leader in Your Niche
Ad

Similar to Hw09 Large Scale Transaction Analysis (20)

PPTX
DockerCon 2017 - General Session Day 2 - Ben Golub
PPTX
FINTECH FINAL.pptx
PDF
Fintech Module 8 - Data, Analytics and Strategy
PDF
Confluent & GSI Webinars series - Session 3
PPTX
Blockchain and the investment industry stack
PDF
MIT15-S08S20_class12.pdf
PDF
DataArt Financial Services and Capital Markets
PPTX
The Big Data Ecosystem for Financial Services
PDF
India: A Fintech Nation Due to Democratized Access
PPTX
Mobile & Blockchain: An Intro to the Decentralised World
PPTX
The Future of Apache Hadoop an Enterprise Architecture View
PPTX
Apache Hadoop Summit 2016: The Future of Apache Hadoop an Enterprise Architec...
PDF
Hypermine | 2019 Business Plan
PPTX
Data Fraud Analytics Industry
PPTX
TMT SA Presentation
PDF
ZIGRAM Introduction March 2023
PDF
Account aggregator Hackathon - Masterclass on data science & AI track
PPTX
Understanding the future of payments industry in India.pptx
PPTX
SECURING_THE_PAYMExxxNT_INDUSTRY[1].pptx
DOCX
The India Fintech Industry is Booming | News & Insights
DockerCon 2017 - General Session Day 2 - Ben Golub
FINTECH FINAL.pptx
Fintech Module 8 - Data, Analytics and Strategy
Confluent & GSI Webinars series - Session 3
Blockchain and the investment industry stack
MIT15-S08S20_class12.pdf
DataArt Financial Services and Capital Markets
The Big Data Ecosystem for Financial Services
India: A Fintech Nation Due to Democratized Access
Mobile & Blockchain: An Intro to the Decentralised World
The Future of Apache Hadoop an Enterprise Architecture View
Apache Hadoop Summit 2016: The Future of Apache Hadoop an Enterprise Architec...
Hypermine | 2019 Business Plan
Data Fraud Analytics Industry
TMT SA Presentation
ZIGRAM Introduction March 2023
Account aggregator Hackathon - Masterclass on data science & AI track
Understanding the future of payments industry in India.pptx
SECURING_THE_PAYMExxxNT_INDUSTRY[1].pptx
The India Fintech Industry is Booming | News & Insights

More from Cloudera, Inc. (20)

PPTX
Partner Briefing_January 25 (FINAL).pptx
PPTX
Cloudera Data Impact Awards 2021 - Finalists
PPTX
2020 Cloudera Data Impact Awards Finalists
PPTX
Edc event vienna presentation 1 oct 2019
PPTX
Machine Learning with Limited Labeled Data 4/3/19
PPTX
Data Driven With the Cloudera Modern Data Warehouse 3.19.19
PPTX
Introducing Cloudera DataFlow (CDF) 2.13.19
PPTX
Introducing Cloudera Data Science Workbench for HDP 2.12.19
PPTX
Shortening the Sales Cycle with a Modern Data Warehouse 1.30.19
PPTX
Leveraging the cloud for analytics and machine learning 1.29.19
PPTX
Modernizing the Legacy Data Warehouse – What, Why, and How 1.23.19
PPTX
Leveraging the Cloud for Big Data Analytics 12.11.18
PPTX
Modern Data Warehouse Fundamentals Part 3
PPTX
Modern Data Warehouse Fundamentals Part 2
PPTX
Modern Data Warehouse Fundamentals Part 1
PPTX
Extending Cloudera SDX beyond the Platform
PPTX
Federated Learning: ML with Privacy on the Edge 11.15.18
PPTX
Analyst Webinar: Doing a 180 on Customer 360
PPTX
Build a modern platform for anti-money laundering 9.19.18
PPTX
Introducing the data science sandbox as a service 8.30.18
Partner Briefing_January 25 (FINAL).pptx
Cloudera Data Impact Awards 2021 - Finalists
2020 Cloudera Data Impact Awards Finalists
Edc event vienna presentation 1 oct 2019
Machine Learning with Limited Labeled Data 4/3/19
Data Driven With the Cloudera Modern Data Warehouse 3.19.19
Introducing Cloudera DataFlow (CDF) 2.13.19
Introducing Cloudera Data Science Workbench for HDP 2.12.19
Shortening the Sales Cycle with a Modern Data Warehouse 1.30.19
Leveraging the cloud for analytics and machine learning 1.29.19
Modernizing the Legacy Data Warehouse – What, Why, and How 1.23.19
Leveraging the Cloud for Big Data Analytics 12.11.18
Modern Data Warehouse Fundamentals Part 3
Modern Data Warehouse Fundamentals Part 2
Modern Data Warehouse Fundamentals Part 1
Extending Cloudera SDX beyond the Platform
Federated Learning: ML with Privacy on the Edge 11.15.18
Analyst Webinar: Doing a 180 on Customer 360
Build a modern platform for anti-money laundering 9.19.18
Introducing the data science sandbox as a service 8.30.18

Recently uploaded (20)

PPTX
social-studies-subject-for-high-school-globalization.pptx
PPT
KPMG FA Benefits Report_FINAL_Jan 27_2010.ppt
PPTX
Session 11-13. Working Capital Management and Cash Budget.pptx
PPTX
The discussion on the Economic in transportation .pptx
PDF
THE EFFECT OF FOREIGN AID ON ECONOMIC GROWTH IN ETHIOPIA
PDF
Spending, Allocation Choices, and Aging THROUGH Retirement. Are all of these ...
PDF
Mathematical Economics 23lec03slides.pdf
PPTX
Unilever_Financial_Analysis_Presentation.pptx
PPTX
OAT_ORI_Fed Independence_August 2025.pptx
PPTX
Introduction to Customs (June 2025) v1.pptx
PDF
Dialnet-DynamicHedgingOfPricesOfNaturalGasInMexico-8788871.pdf
PPTX
Who’s winning the race to be the world’s first trillionaire.pptx
PDF
Bitcoin Layer August 2025: Power Laws of Bitcoin: The Core and Bubbles
PDF
HCWM AND HAI FOR BHCM STUDENTS(1).Pdf and ptts
PDF
Lecture1.pdf buss1040 uses economics introduction
PDF
Understanding University Research Expenditures (1)_compressed.pdf
PDF
Unkipdf.pdf of work in the economy we are
PDF
Dr Tran Quoc Bao the first Vietnamese speaker at GITEX DigiHealth Conference ...
PDF
Corporate Finance Fundamentals - Course Presentation.pdf
PDF
Predicting Customer Bankruptcy Using Machine Learning Algorithm research pape...
social-studies-subject-for-high-school-globalization.pptx
KPMG FA Benefits Report_FINAL_Jan 27_2010.ppt
Session 11-13. Working Capital Management and Cash Budget.pptx
The discussion on the Economic in transportation .pptx
THE EFFECT OF FOREIGN AID ON ECONOMIC GROWTH IN ETHIOPIA
Spending, Allocation Choices, and Aging THROUGH Retirement. Are all of these ...
Mathematical Economics 23lec03slides.pdf
Unilever_Financial_Analysis_Presentation.pptx
OAT_ORI_Fed Independence_August 2025.pptx
Introduction to Customs (June 2025) v1.pptx
Dialnet-DynamicHedgingOfPricesOfNaturalGasInMexico-8788871.pdf
Who’s winning the race to be the world’s first trillionaire.pptx
Bitcoin Layer August 2025: Power Laws of Bitcoin: The Core and Bubbles
HCWM AND HAI FOR BHCM STUDENTS(1).Pdf and ptts
Lecture1.pdf buss1040 uses economics introduction
Understanding University Research Expenditures (1)_compressed.pdf
Unkipdf.pdf of work in the economy we are
Dr Tran Quoc Bao the first Vietnamese speaker at GITEX DigiHealth Conference ...
Corporate Finance Fundamentals - Course Presentation.pdf
Predicting Customer Bankruptcy Using Machine Learning Algorithm research pape...

Hw09 Large Scale Transaction Analysis

  • 1. Hadoop@Visa Joe Cunningham Technology Strategy and Innovation October 2, 2009
  • 2. Agenda 1 Visa 2 Value-Added Information Products 3 Hadoop@Visa – research results VisaNet – an overview
  • 3. Keeping the Brand Promise Zero Liability go Security Flexibility
  • 4. Keeping the Brand Promise 3 Global Acceptance Reliability Zero Liability go Security Flexibility Guaranteed Payment
  • 5. VisaNet Today VisaNet is the world’s largest, most advanced payments network. 28 million acceptance locations 130 million authorizations per day 1,600 endpoints Processes transactions in 1 second 16,300 financial institutions Less than 2 seconds unavailability per year Processes in 175 currencies 1.4 million ATMs fast secure flexible reliable scalable
  • 6. 1 VisaNet – an overview 2 3 Value-Added Information Products Hadoop@Visa – research results Agenda
  • 7. Visa Processing Architecture Security/Access Services Message File Web VisaNet Services Integration Common Data Infrastructure Authorization Clearing & Settlement Information Disputes Risk
  • 8. Value-Added Information Products Information Services Client Portfolio Analysis Visa Incentive Network Accountholder Transaction Alerts Account Updater Tailored Rewards Risk Management Services Authentication Advanced Auth Verified by Visa Address Verification Account Monitoring Fraud Detection Compromised Accounts Consumer Alerts Encryption Dynamic Key Exchange Dynamic Cryptograms
  • 9. 1 VisaNet – an overview 2 Value-Added Information Products 3 Hadoop@Visa – research results Agenda
  • 10. Research Lab Setup Hadoop Systems VM System Relational Database Custom Analytic Stacks Encryption Processing Management Stack Hadoop #2 ~300Tb / 28 nodes Hadoop #1 ~40Tb / 42 node
  • 11. Risk Product Use Case Create critical data model elements, such as keys and transaction statistics, which feed our real-time risk-scoring systems Hadoop Cluster Research Sample (Synthetic Transactions) 500 million distinct accounts 100 million transactions per day 200 bytes per transaction 2 years: ~ 73 billion transactions, ~ 36 TB One Month  13 minutes! Input: Transactions Merchant Category Country / Zip Output: Key & Statistics MCCZIP Key Statistics related to account, transaction type, approval, fraud, IP address, tx. counts
  • 12. Financial Enterprise Fit Some key questions we are researching… What will the Hadoop Solution Stack(s) look like? File System Transaction Sample System Relational Back-End Analytics Processing Internal vs. External cloud How do I get data into a cloud in a secure way? How does HSM and security integration work in Hadoop? What are the missing pieces?
  • 13. Why Hadoop@Visa? For our information products, we see promise in combining Visa’s data analytics capability with the power of Hadoop Analyze volumes of data with response times that are not possible today… … and apply analytic models to individual client, not just client segment But, need to answer questions relating to stack configuration and support… … and, eventually must fit with our financial enterprise paradigm: fast secure flexible reliable scalable
  • 14. Keeping the Brand Promise Zero Liability go Security Flexibility
  • 15. Keeping the Brand Promise 3 Global Acceptance Reliability Zero Liability go Security Flexibility Guaranteed Payment

Editor's Notes

  • #6: Fast – On average, VisaNet fully processes transactions in approximately 1 second. Secure – VisaNet supports multiple defense layers to prevent network intrusions, protect against account fraud, and to make stolen card data unusable for criminalsincluding data encryption, sophisticated intrusion detection systems and neural network technology capable of analyzing each transaction in real-time. Over the last two years, Visa has spent more than $100 million on data security enhancements. Reliable – VisaNet runs multiple, parallel processing systems in highly redundant data centers. This ensures near 100 percent availability. In the past 15 years, VisaNet experienced less than two seconds average unavailability per year. State-of-the-art monitoring tools are in place to detect network availability and client-specific processing anomalies, such as high rates of decline. Scalable – VisaNet authorizes an average of 130 million transactions a day, totaling more than 92 billion authorization. On Dec. 21, 2008, Visa’s busiest day last year, VisaNet authorized more than 200 million transactions. Based on recent stress tests, VisaNet is capable of authorizing over a half billion transactions per day.