Global Trading Platform Vamsi Chemitiganti Chief Architect Financial Services Derrick Kittler Middleware SA Manager, Lead 11 June 2013 -
2 What's this talk about? What it is... and... What it is not...
3 Agenda Introductions Securities Industry background & trends Terminology Requirements Architectural Components Representative Architecture The road ahead Questions
4 Securities Industry The securities industry is one of the most important industries in the world, contributing 10% to the US GDP alone* By itself, the industry would be the 10th largest country in the world in terms of GDP Provides vital functions in terms of huge numbers of employed workers, capital access to govt & entrepreneurs Repository of household savings and assets; high equity ownership among US households Post financial crisis, huge emphasis on oversight and supervision
5 Trends in the Securities Industry Increasing Volumes globally more issuers and more trading Changing marketplaces and trading avenues Shrinking Margins leading to Cost & Efficiency pressures Complex business processes Increasing Risks Regulatory Pressures Cost pressures
6 Overview of financial markets
7 Industry Terminology... Equities, Fixed Income and Options OMS (Order Mgmt System) FIX MDD (Market Data Distribution) Broker Front, Back and Mid Office Buy Side and Sell Side Marketplace
8 Trade & Transaction Lifecycle key workflows
9 Lifecycle of a Trade Ref. DTCC
10 Business Requirements Architecture must offer support front,mid & back office trading capabilities Display trade and blotter UIs for trade management. Support trading as a service Support a wide variety of financial products and formats. The first version will focus on equities Provide support for order capture, trading, crossing Provide the ability to cross buy and sell side market orders when both side orders are detected in the system) Auto route and execute orders based on accounts, quantity and real time market data Other order routing requirements..too many to list here
11 System Requirements Ability to support open messaging, complex events, workflow, business rules and integration capabilities GTP should provide realtime, low latency messaging capabilities Ability to add support for FIX,FpML,Market Data, JMS adapters etc. Support multiple versions of FIX and ability to route orders based on data within FIX Ability to run in both embedded as well as standalone mode with built in support for HA and Performance Other order routing requirements..too many to list here Need to provide rich Integration into Existing Enterprise Management Systems and Processes
12 Global Trading Platform Functional Architecture
13 Design Tenets SOA based approach is key all core business functions modeled as SOA services ESB to interconnect all market participants Open messaging standard - AMQP chosen as the transport protocol of choice for performance and industry reasons FIX (Financial Information Exchange) run over AMQP JBOSS Fuse ESB chosen as the service bus JBOSS BRMS provides Rules,CEP and BPM under a single umbrella All data management services can be modeled as data services in a future version
14 Key components JBoss Fuse - mediation, routing and transformation Red Hat Messaging AMQP, flow control, guaranteed delivery Camel connectors QuickFIX, Bindy, etc... from("quickfix:examples/inprocess.cfg?sessionid=fix.4.2:market TRADER&exchangePattern=InOut").filter(header(QuickfixjEndpoint.ME SSAGE_TYPE_KEY).isEqualTo(MsgType.ORDER_STATUS_REQUEST)).bean(new MarketOrderStatusService()); JBoss BRMS Rules for order mgmt, routing, crossing,matching BPM for error handling escalation and auditing
15 Deployment Arch deepdive Started with embedded model. BPM and Rules embedded in existing Tomcat, IBM and BEA application servers BPM as a library than as a server Need to provide standalone BPMS and EDM Need to build out a shared infrastructure to host different process applications Centralize business process knowledge per group in a set of servers Applications access servers via webservice calls. Detailed performance & failover testing Decision server interface
16 Global Trading Platform Deployment Architecture
17 Global Trading Platform Data Flow
18 The road ahead Further model all core business functions as SOA services Introduce CEP semantics as trading volumes increase and as more clients and instruments are brought on Create tiers of AMQP brokers by consumer type and service offered BPM and EDM as Service Implement in memory caching to improve lookup speed as well as to introduce in memory processing of data Looking at AMQP 1.0 and at the Proton project as well Beginning to examine modeling data elements as data services