The JSE, founded in 1887, is the largest stock exchange in Africa. It ranks as one of the 20 largest stock exchanges in the world by market capitalization. The JSE has operated as a marketplace for the trading of financial products for over 120 years. During this time, the JSE has evolved from a traditional floor-based outcry market to a modern securities exchange providing electronic trading, clearing, and settlement in underlying and derivative instruments for a variety of asset classes, and has extensive surveillance and regulation capabilities. The JSE also operates the South African government bond market on behalf of the National Treasury and is a major provider of financial information.
A Real-Time, Single Source of Truth
The JSE maintains over 180 disparate data sources across its data landscape. Additionally, over 120 different applications need to operate in perfect harmony to ensure that there is accurate settlement. Many are small in data volumes but high in complexity, so data integration is a top priority. The data had to be sourced and integrated from a variety of heterogeneous data systems, including relational databases, flat files, and APIs, to serve the needs of the various functions. Using traditional batch-oriented extract, transform, and load (ETL) processes was cumbersome and inefficient. Additionally, the different versions of data entities across different systems created confusion and skepticism around data authenticity. In this high-speed trading environment, it is imperative to remain 100% accurate for 100% of the time, so the JSE looked beyond the traditional ETL and batch processing solutions for data integration.
To consolidate its intricate data landscape and aggregate data in real time, the JSE implemented the Denodo Platform, a data integration and data management solution built on data virtualization, to build a logical data layer. All of the relevant information sourced from these systems is compiled into base views on the Denodo Platform. The JSE has built more than 1,700 base views on top of these data sources, which are processed by transformation rules to produce derived and interface views. The data integration layer processes about 2 billion rows a month at the JSE.