Book Image

Building Low Latency Applications with C++

By : Sourav Ghosh
5 (1)
Book Image

Building Low Latency Applications with C++

5 (1)
By: Sourav Ghosh

Overview of this book

C++ is meticulously designed with efficiency, performance, and flexibility as its core objectives. However, real-time low latency applications demand a distinct set of requirements, particularly in terms of performance latencies. With this book, you’ll gain insights into the performance requirements for low latency applications and the C++ features critical to achieving the required performance latencies. You’ll also solidify your understanding of the C++ principles and techniques as you build a low latency system in C++ from scratch. You’ll understand the similarities between such applications, recognize the impact of performance latencies on business, and grasp the reasons behind the extensive efforts invested in minimizing latencies. Using a step-by-step approach, you’ll embark on a low latency app development journey by building an entire electronic trading system, encompassing a matching engine, market data handlers, order gateways, and trading algorithms, all in C++. Additionally, you’ll get to grips with measuring and optimizing the performance of your trading system. By the end of this book, you’ll have a comprehensive understanding of how to design and build low latency applications in C++ from the ground up, while effectively minimizing performance latencies.
Table of Contents (19 chapters)
1
Part 1:Introducing C++ Concepts and Exploring Important Low-Latency Applications
6
Part 2:Building a Live Trading Exchange in C++
10
Part 3:Building Real-Time C++ Algorithmic Trading Systems
14
Part 4:Analyzing and Improving Performance

Subscribing to market data and decoding the market data protocol

The first component we need to build inside the market participants’ trading system is the market data consumer. This component is responsible for subscribing to the multicast stream of public market data updates published by the trading exchange. It needs to decode the market data stream generated by the exchange from the public MDPMarketUpdate format we discussed earlier. Because of the choice of the Simple Binary Encoding (SBE) protocol, the decoding step is straightforward in our application and does not involve any complicated stream decoding logic. Another important responsibility of this component is detecting packet drops on the incremental market data stream and providing mechanisms to recover and synchronize with the market data stream again. This mechanism is also required for trading systems that subscribe to the market data stream after there is a non-empty order book, i.e. after the trading exchange...