Time Synchronization using ML Techniques in Electronic Trading
In modern computer networks, time synchronization is critical because every aspect of managing, securing, planning, and debugging a network involves determining when events happen. Accurately and consistently synchronized clocks are crucial in Electronic Trading networks where latency profiles are measured in microseconds, if not in nanoseconds, and transactions order is important.
Existing time synchronization protocols vary in precision, accuracy and cost. Synchronization algorithms use several different approaches to maintain accuracy, but in general they involve propagating a signal throughout a datacenter. Individual servers will calculate their perceived time, based on the initial signal, anticipated network latency and jitter that impacted the path taken by the signal to the server. In this talk, we will discuss an algorithm, designed and published by Stanford University, which determines the required time corrections not based on a single signal alone, but expands it to rely on several signals received from a designated set of servers. The algorithm leverages machine learning to determine the current time with nanosecond level accuracy.
By attending this session, you will gain an:
* Overview of Electronic Trading technology and why software, hardware and network optimizations are important in highly distributed and real time Electronic Trading systems
* Understanding of the significance of time synchronization to Electronic Trading and Stanford University work on nanosecond level time synchronization using hardware clocks and machine learning techniques
You may also be interested in
Growing up in Honduras, Bolivia, India, Nepal and Indonesia, Thomas believes innovation decision making is one of the most far-reaching...
As a software development manager or a technical leader you are eager to remove impediments so your team can produce...