Benchmarking GPU-Accelerated Databases Against Traditional Databases; A Sociotechnical Analysis of High Frequency Trading

Databases systems play a critical role in the ability to organize and analyze data. In an increasingly data-inundated world, databases must continue evolving to provide sufficient compute power to analyze large datasets. GPU-accelerated databases are a particular type that leverage the highly-parallel nature graphical process units to perform rapid fast queries. In this paper, the performance of a GPU-accelerated database, Kinetica, is benchmarked against PostgreSQL, a popular open- sourced database, for various queries using a standardized database benchmarking framework: TCP-DS. GPU-accelerated databases show promise in database performance for complex, OLAP query workloads.

This paper analyzes the relationship between high frequency traders (HFT), exchanges, regulators, and market participants. Specifically, the paper analyzes the causes behind the mainstream distrust towards HFT by examining the ambiguous role of HFT as beneficial liquidity providers and destructive price manipulators. Through policy analysis, the paper explores both proposed and current laws proposed for enforcing market-enhancing activities among HFT while deterring manipulative practices. Through network analysis, the unforeseen consequences of certain regulations and how exchanges work with HFT to change the way investors and market participants approach the market are examined. The Social Construction of Technology (SCOT) framework is used in this paper to understand the relationship between HFT, regulators, and exchanges and how each entity is capable of far-reaching economic impacts on society, specifically by influencing the capabilities of modern financial markets. The intent of this paper is to provide a framework to understanding the fundamental role of HFT in society and to suggest a methodical approach in designing proper regulation for HFT such that their activity can enhance financial markets. The research in this paper is important for STS and engineering, as it illustrates how computational technology is leveraged to fulfill economic needs while introducing unique social challenges.