tibco streambase: real-time analytics: the last mile in big data

28
1 Enabling Real Time Analytics in the Capital Markets Richard Tibbetts CTO, TIBCO StreamBase Real Time Analytics: The Last Mile in Big Data

Upload: tibco-spotfire

Post on 29-Nov-2014

1.442 views

Category:

Business


0 download

DESCRIPTION

TIBCO StreamBase: Real-Time Analytics: The Last Mile in Big Data Richard Tibbetts, CTO, TIBCO StreamBase ANALYTICS AND VISUALIZATION FOR THE FINANCIAL ENTERPRISE CONFERENCE June 25, 2013 The Langham Hotel Boston, MA

TRANSCRIPT

Page 1: TIBCO StreamBase: Real-Time Analytics: The Last Mile in Big Data

1

Enabling Real Time Analytics in the Capital Markets

Richard Tibbetts

CTO, TIBCO StreamBase

Real Time Analytics: The Last Mile in

Big Data

Page 2: TIBCO StreamBase: Real-Time Analytics: The Last Mile in Big Data

2

Real Time Visibility: The Last Mile in Big Data

The Last Mile:

Delivery to

usage site

• Expensive

• Purpose Built

• Non-Optional

Page 3: TIBCO StreamBase: Real-Time Analytics: The Last Mile in Big Data

3

Big Data: Everyone Talks about Volume

In the past 10 years, the amount of

data to be processed has exploded.

Mobile devices, sensors, market

data, video, audio, and social media

are among the sources of this

explosion of data.

Page 4: TIBCO StreamBase: Real-Time Analytics: The Last Mile in Big Data

4

There can be an overwhelming volume

Page 5: TIBCO StreamBase: Real-Time Analytics: The Last Mile in Big Data

5

Most solutions focus on accumulating and analyzing

Page 6: TIBCO StreamBase: Real-Time Analytics: The Last Mile in Big Data

6

Especially analyzing things which already happened

Page 7: TIBCO StreamBase: Real-Time Analytics: The Last Mile in Big Data

7

But how will you access the data when you need it?

Page 8: TIBCO StreamBase: Real-Time Analytics: The Last Mile in Big Data

8

Closing the loop – creating value with big data

Accumulate – Capture and store the data

• Use the right technology to capture data as it happens (e.g. Flume)

• Load it into appropriate price/performance storage (e.g. HDFS)

Analyze – Form and test hypotheses

• Discover mathematical relations and patterns (e.g. R)

• Crunch enormous volumes of data to validate and tune (e.g. Hadoop)

Access – Use the knowledge at the point of decision

• Get the data to the people who need it, when they need it

• Present it in the way that they can consume it

Automate – Optimize even small decisions with data

• Control systems with a closed feedback loop

• Monitor and focus on the exceptions, automate the common actions

Page 9: TIBCO StreamBase: Real-Time Analytics: The Last Mile in Big Data

9

Real Time Analytics

StreamBase Event Processing Platform

Page 10: TIBCO StreamBase: Real-Time Analytics: The Last Mile in Big Data

10

StreamBase EventFlow - Integration, Analytics, Reaction

Visual programming integrates components developed by team members with different skills

• Developers, Quants, Data Scientists, Traders, Business Analysts

Connectivity

Data Infrastructure

Quantitative Analytics

Audit Logging

Business Strategy

Off the Shelf Components

Page 11: TIBCO StreamBase: Real-Time Analytics: The Last Mile in Big Data

11

Applications Integration

Normalization Aggregation Correlation

Rules Alerts

Automation

Adapters and

Handlers

Adapters and

Handlers

StreamBase Server(s)

StreamBase Studio for Developing EventFlow Applications

Data Management

Persistence Stores

Logs

Market Data

Trade Data

Static Data

Systems Data

Performance Benchmarks

Automation

Desktop

Alerts

StreamBase Platform for Complex Event Processing

Inputs Outputs

Page 12: TIBCO StreamBase: Real-Time Analytics: The Last Mile in Big Data

12

StreamBase Studio: Eclipse Visual Development

Visual operators

Algorithm library

Schemas, threading, etc.

Event Application Code

Page 13: TIBCO StreamBase: Real-Time Analytics: The Last Mile in Big Data

13

StreamBase Studio: Agile Development, Mixed Teams

Feed Simulation

Refactoring

Seamless Java & visual EventFlow Debugger

Visual Diff/Merge

Unit Testing

Java Operators & Function Wizards

Page 14: TIBCO StreamBase: Real-Time Analytics: The Last Mile in Big Data

14

API’s

C++, Java, .NET, Python

Messaging

TIBCO RV & EMS, JMS, MQ Series, RMDS, Solace, Tervela, Wombat, 29West

Additional General Purpose Adapters

FIX, STAMP, JDBC, POP3, IRC, HTTP, SMTP, IM, XML, RSS, Twitter

StreamBase Connectivity and Visualization Options

Market Data/Venue Samples

Visualization

Eclipse RCP

Adobe Flex

Java Swing

.NET

MS Excel

SL RTView

Databases

TR Velocity

Page 15: TIBCO StreamBase: Real-Time Analytics: The Last Mile in Big Data

15

Kosmos: StreamBase for High Frequency Trading

Who is Kosmos?

• One of Brazil’s leading hedge funds which employs both quantitative and qualitative investment strategies

StreamBase for HFT

• Connectivity to wide selection of equities and FX data sources

• Building next generation trading system

• Equities, FX and derivatives

• Selected StreamBase over other CEP providers

Why StreamBase?

• Rapid time-to-market

• Flexibility

• Developer empowerment

Page 16: TIBCO StreamBase: Real-Time Analytics: The Last Mile in Big Data

16

SUNGARD: Pre-Trade Risk Solution

Business Driver:

• Fast Response to Regulatory Change

Project:

• Enhancement of Assent Liquidity Solutions and Brass Order Management System

Asset classes:

• Equities

• Derivatives

• Fixed income

Why StreamBase?

• Maturity of StreamBase’s platform and success in the CEP market

• Graphical event-flow language and visual environment that enable developers and business users to work more efficiently

• Superior performance to reduce latency issues

“CEP... is very good for rapid development of applications which are deployed in high-volume, low-latency type environments.

The products offered by Assent are constantly being challenged with lots of market data, orders and execution updates. We want to use StreamBase to augment and enhance some of the products that we have.”

Chris Lees, VP Liquidity Services, SunGard

Page 17: TIBCO StreamBase: Real-Time Analytics: The Last Mile in Big Data

17

BNY ConvergEx: StreamBase for Monitoring

BNY ConvergEx: The Convergence of Brokerage and Technology

• Top 10 NYSE liquidity provider

• Vortex, an electronic ATS / dark pool

• A leader in algorithmic trading and OMS technology

• Deployed StreamBase for monitoring of U.S. trading operations

BNY ConvergEx Deploys StreamBase for U.S. Trading Operations

• Intelligent monitoring of trading operations, including its dark pool Vortex

• Real-time market interaction monitoring

• Real-time credit risk correlation

• System health monitoring

• Real-time analysis of market feeds for latency and response time

“The implementation of StreamBase allows us to monitor our risk… and assess the health of our entire U.S. trading operations in real-time.” Joseph Weisbord, Managing Director, BNY ConvergEx Group

Page 18: TIBCO StreamBase: Real-Time Analytics: The Last Mile in Big Data

18

CityIndex: StreamBase for Options Pricing and Monitoring

Who is CityIndex?

• Leading global provider of retail trading services, including CFDs, margined FX and spread betting

StreamBase for Pricing and Monitoring

• Real-time access to prices

• Rapid application development and deployment

Why StreamBase?

• Rapid time-to-market

• High performance

• Ease-of-use

Watch what Toby Young has to say about StreamBase

(Requires 2D barcode scanner or click to view)

Page 19: TIBCO StreamBase: Real-Time Analytics: The Last Mile in Big Data

19

Internet

Hadoop

MR

MR

MR

MR

MR

MR

MR

MR

MR

MR

MR

MR

Map Reduce

StreamBase CEP

Monitoring

Automation

Filtering

Alerting

Real-Time Applications

Data Warehouse

Hadoop Flume MySQL

HDFS

Analytic

Analytic

Analytic

Analytic

Operational Monitoring

LiveView

Historical Analytics

Analysts

Operational Staff

Fraud Analysts

Fraud detection

Fraud, Security

Management Control

Site Metrics

Bu

sin

ess I

nte

llig

ence

Real Time Big Data: Virtual Goods Phishing and Fraud

Page 20: TIBCO StreamBase: Real-Time Analytics: The Last Mile in Big Data

20

Operational Visibility for Real Time Analytics

StreamBase LiveView

Page 21: TIBCO StreamBase: Real-Time Analytics: The Last Mile in Big Data

21

Example End User Experience

Real-Time Graphs

Active Tables

Ad-hoc query

Notifications

End-user configurability

Page 22: TIBCO StreamBase: Real-Time Analytics: The Last Mile in Big Data

22

StreamBase LV Use Case - Real-Time Risk Management

Real-time exposure by asset class

Real-time transaction detail

Real-time exposure by desk

Page 23: TIBCO StreamBase: Real-Time Analytics: The Last Mile in Big Data

23

EventFlow and R: Pre-Deal Real Time VAR Checks

Normalize

Integrate with R Simulations

Loop each trade through all scenarios

Calculate per-trade, per-scenario Risk

Calculate Risk: 95%,99%,X%

Accept/Reject proposed trade, minimize risk to

counterparty

Aggregate

Configure parameters for back testing

Connect

Page 24: TIBCO StreamBase: Real-Time Analytics: The Last Mile in Big Data

24

StreamBase LiveView Use Case - Real-Time TCA

Real-Time TCA Alerts

Live Orders, TCA, FIX

Which orders are slipping?

Page 25: TIBCO StreamBase: Real-Time Analytics: The Last Mile in Big Data

25

LiveView Server – Data Source and Query Clients

Message Bus

Market Data

News

Orders &

Executions

Data

LiveView Server

Snapshot Results

Continuous Query Processor

Query

Continuous Alerting

Active Tables Active Tables

Continuous Updates

Sources Clients

Maintain Data Image

Respond to Continuous Queries

Materialize analysis

Pre-aggregate data

Manage alerts

Page 26: TIBCO StreamBase: Real-Time Analytics: The Last Mile in Big Data

26

Clients: Desktop, Web, Excel, Mobile, Custom .NET, etc

Page 27: TIBCO StreamBase: Real-Time Analytics: The Last Mile in Big Data

27

GAIN Capital: FX Pricing, Risk, Surveillance and Compliance

GAIN Capital

• Improved risk management, surveillance and compliance system

• Rapid applications design, development and deployment

• Faster execution

GAIN’s developers

• Intuitive development environment

• Library of application examples and help files

• Responsive support team

GAIN’s Forex.com clients

• Tighter spreads for customers through improved pricing and market making

StreamBase/GAIN Institutional Clients

• Rapid FX applications design, development and deployment

• Allow developers to focus on strategies and not the plumbing

• High frequency trading connectivity and order execution to all FX liquidity providers including tier 1 and tier 2 banks

“StreamBase Event Flow programming language is 50% faster than traditional programming techniques.” Andrew Haines, CIO, GAIN Capital

Watch what Andrew Haines has to say about StreamBase on TABB Forum

(Requires 2D barcode scanner or click)

Page 28: TIBCO StreamBase: Real-Time Analytics: The Last Mile in Big Data

28

Summary and Next Steps

Once you have accumulated, analyzed, and accessed big data you will have insight

Operationalizing insight on real time data is Event Processing

StreamBase’s Event Processing Platform is best of breed

• Natural and powerful EventFlow authoring environment

• Extensible integrations with analytics

• High performance deployment server

• Rich portfolio of connectivity with focus on capital markets

• LiveView for operational monitoring and alerting

Download and try out http://streambase.tibco.com

Questions?