Business, Finance and Information Marketplaces

Companies are now using powerful analytics to predict customers’ and competitors' customers’ behavior in order to better tailor their products, product pricing and services.  However, understanding the customer means relying on data from multiple data sources, hence a problem of data variability. Geo-spatial data, social media, voice, video are just some of the examples of unstructured data that contribute to a company’s knowing the customer today and predicting their future behaviors.

Companies are also using "information marketplace" platforms that correlate data streams from services of other work packages and derive insights from these correlations through domain specific applications. This marketplace provides the infrastructure for advertising, provisioning and billing raw data, services and access to applications.

Moving one layer of abstraction from the companies to the financial markets, High-frequency trading (HFT) and algorithmic trading, the practice of using automated financial tools to rapidly trade large volumes of securities, has given rise to a rapidly escalating speed war between companies seeking any marginal advantage over its competitor.

One hardware direction that we will consider with respect to Big Data is the growing use of FPGAs which we already see coming online with respect to the previously mentioned HFT speed war.  The potential advantage of FPGAs is their ability to combine high bandwidth I/O with computational horsepower together with real-time response times, and European company, Maxeler has already achieved some success in this area.