An Algorithmic Trading Management System Based On Parallel Computing

Algorithmic trading has dominated trading technologies in financial markets. It not only makes the trading volume and trading channels increase exponentially at a dramatic rate but also avoid the operation errors by manual trading. In order to do arbitrage, algorithmic trading needs to calculate the correlations among multi markets. And in order to hidden the trading purpose, a big order needs to split into many small orders and trades via several channels. So algorithmic trading uses many trading strategies to do arbitrage and control the position. Therefore the development of algorithmic trading has a high cost on the management of a number of different quote and trading interfaces and trading strategies, which are non-core modules.The algorithmic trading management system is a platform to integrate quote and trading interfaces, sub accounts management system, risk warming system, trading strategies management system and pre-computing database. It has several functions as follows:1. It makes the development of algorithmic trading focus on the algorithm rather than other non-core modules, which in turn reduces the developing cost.2. It can hidden the trading purpose and prevent the counterparties to discover the trading trend.3. It provides the basis support for the arbitrage in multi markets.4. It makes one trading algorithm can quickly spread in several markets.5. It employs distributed build thus the trading algorithm can increase their parallel running speed with the increase of the number of computers.The demand analysis illustrates the functionality demand and non-functionality demand. The program design is according to the demand details, and one demand has several solutions and the optimal solution is determined by the user or automatically chosen by the computer. The automated test instruction explains the necessity and provides the evaluations for performance, reliability and functionality. The automated deployment instruction provides thermal expansions and automatically dynamic deploy solutions.