Distributed Statistics Generator System

The purpose of the project is to load a number of channels with various stock ticker data statistics that can be consumed selectively by multiple trading engines. These trading engines will then update the stocks in their portfolios as necessary and print out statistics as they are updated.

The system is largely structured into three layers: a data producer, a statistics engine (consumer/producer), and a set of training engines (consumers).

The source of ticker data for this system is a set of csv files representing bid prices, bid quantities, ask prices, and ask quantities for three companies: Microsoft, Oracle, and IBM. These csv files are read into system via the DataProducer project (from the data inbox) on a one-second delay.

Usage Notes:

• First, ensure that Apache ActiveMQ is running tcp://localhost:61616
• Begin by executing the producer class DataProducer.java in the DataProducer project. This will load csv ticker data in to the respective ticker data message queues
• Next, run the StatisticsSubscriber.java classes in each of the TradingEngine projects. These subscribers will subscribe to the company stats topics they are interested in to then update their statistics
• Finally, run the StatisticsStream.java class in the StatisticsEngine project. This system will consume from the ticker data message queues, run the statistics, and then push the statistics onto the relevant topics to be consumed by the trading engines
• Note that each of the trading engines will print out a statistics report on its portfolio(s) when it receives new information. These reports are printed to standard output. It may be necessary to stop the execution in order to inspect the output.

Enterprise Integration Patterns Used:

1. Message/Message Broker/Message Endpoint/Message Channel: Implemented via Camel and used to form message queues/topics for the ticker data and statistics

2. Message Router: Used to route ticker messages to their appropriate company queues

3. Publish-Subscribe Channel: Utilized for trading engines to subscribe to statistics topics for the companies in which they have stocks

4. Point-To-Point Channel: Utilized for ticker data consumer to push messages onto a queue to be directly consumed by the statistics engine

5. Message Translator: Utilized for translating stock ticker csv data into java hash map objects

6. Invalid Message Channel: Channel which is fed invalid ticker data (values must be numeric and positive)

7. Delay EIP: Utilized in liu of polling consumer. Delays action on message received based on given time.

Design Patterns Used:

1. Singleton: Utilized to force each trading engine to only have a single instance of "Reporting Engine". Also implemented for "Statistics Engine"

2. Strategy: Used in Statistics Engine to streamline implementation of various statistic-calculating algorithms

3. Template Method: Used to provide “skeleton” of steps for implementing statistic-calculating algorithms

4. Composite: Structure used to represent a trading engine's portfolio(s) and stocks

5. Iterator: Used to iterate through the portfolio composite structure

System Diagram: