After reading the bestselling book of Yuval Noah Harari (Homo Deus: A Brief History of Tomorrow) I was glad to conclude that building systems which are artificially conscious, is still a long way to go. However, the upcoming years artificial intelligence (AI) will impact many different types of work, especially where intelligence can be automated.
A lot of systems and software are available on the market to create AI-applications, but most of them expect the data on which AI can be executed ready to be processed. Unfortunately this is often not the case. AI-applications (as any application which create insight based on data) are the end of a chain starting with data created by systems and/or devices.
Between these two opposites a number of process steps and capabilities are required to produce reliable and valid insights. Besides this, the same set of data might be used differently, for instance to create a report as well as for simulation purposes. Implementing AI should therefore be an integrated part of an organizations’ “data-centric” or “insights-driven” strategy.
Within The Open Platform 3.0 Big Data Project, of which I’m a member, we recognized that a reference architecture describing standards which help organizations setting up this “insights-driven” strategy, was needed. We all agreed that a standard should be based on a Business Data Lake as a provisioning environment for data to create insights.
This standard, called ‘Open Business Data Lake Conceptual Framework' (O-BDL) has been published in March 2017 and is a first step towards a Reference Architecture for an Open Business Data Lake. An O-BDL is presented a platform providing enterprise capabilities to:
• Consolidate and preserve enterprise data sets along the data lifecycle
• Assemble preserved enterprise data sets by applying suitable structural definitions and transformations
• Consume these assembled data to explore data and/or to create (AI)-insights
As part of the O-BDL a set of concepts (data, ingestion, processing and data management related) are described, including a number of business scenarios which can be enabled by O-BDL implementations, such as the off-load of Enterprise Data Warehouse (EDW), the creation of Big Data Apps, the Data-Driven Enterprise and an ecosystem of data driven enterprises.
Next Wednesday July 19th I’ll present the O-BDL at the Open Group Conference in Ottawa. In my next blog I’ll describe the characteristics of an O-BDL.