IBM doesn’t release a whole lot of news detailing the ins and outs of its Db2 database division. Indeed, database vendors including Oracle, SAP, Microsoft and open source champions such as Apache Cassandra, MongoDB and Redis tend to take up more column spaces in the tech-business press. Each vendor is constantly trying to outstrip the other in terms of its database’s supposed ability to scale upwards in size, handle deeper analytics complexity and work increasingly effectively (marcoms people like to say ‘seamlessly’) on different types of data pipeline jobs, all in different cloud environments.
What brought IBM forward to start talking about Db2 then? As is the case with the lion’s share of technology augmentations currently playing out at the end of this decade -- it’s all about adding Artificial Intelligence (AI) via the Machine Learning (ML) methods that fuel it.
How smart is a database?
Databases are already quite clever i.e. they can store data in different shapes, formats, vectors, directions and degrees of structured, semi-structured or unstructured rawness. The trouble is, we all depend on them a lot (either directly, or at the backend of the apps we use everyday) and so they now need to get a whole lot smarter. Because of this need for extra smartness, IBM is adding AI into Db2 through data science methods that will inject Machine Learning into it.
“Today, expectations for the database are higher,” said Prasun Mahapatra, senior database administrator at Micro Focus, and a Db2 user. “Databases must be smarter. Building out Db2 as the new AI database will enable users to optimize everything from the usage of data structures, memory and disk, to the most complex queries.”
What IBM has actually done is to provide built-in support for ‘data science’ development in the Db2 version 11.5 release. Data science is sometimes thought of as the three-way intersection point between mathematics, technology hacking & programming, and business strategy. This video explains data science in relatively simple terms -- essentially it is a way of getting computers to deduce and work out more from the data crunching tasks we give them.
How much smarter is an AI/ML database?
So if AI + ML is helping computing systems to get increasingly smarter, how does it get into a database? IBM is building that channel by adding new ‘drivers’ (software connections that manage interfaces) for multiple open source programming languages and frameworks that will allow developers to analyze and build Machine Learning models into software applications that make use of Db2. In theory, at least, such apps should require less management, be more resilient to outages and help to improve user productivity.
Also new with Db2 11.5 is the Augmented Data Explorer, a new natural language querying feature that is designed to give developers a traditional search engine-like experience. Users can pose questions to Db2 and receive results in data visualizations and summaries written in natural language, for easy understanding. The tool itself is containerized for ease of deployment and management and features dynamic visualization, which can help speed the exploration of datasets when building applications.
“In addition to these advances, Db2 users and developers will be able to take advantage of the new capabilities in IBM’s Data Virtualization technology – already available with IBM Cloud Private for Data. Data Virtualization is designed to enable users to easily search across diverse data sources. With it, developers and data engineers can focus on development, minimizing time spent on Extract, Transform and Load (ETL) processes that are associated with moving data. Db2 now also includes blockchain support that helps application developers pull data directly from a blockchain and combine that data with other data sources for analytics or dashboards,” noted IBM, in a press statement
With the addition of Db2 11.5, IBM has streamlined the portfolio to three editions that share the same common code base: Db2 Community, Db2 Standard, and Db2 Advanced. IBM Db2 is a no charge (and perpetually free) download for trials and developers and is intended for a single application developer to use to design, build, test and prototype applications for deployment on Db2 client or server platforms.
AI everywhere
Vice president for IBM data and AI Daniel Hernandez has said the changes IBM has made here are ‘reflective’ of IBM’s clients’ growth patterns and the acceleration of AI adoption we all expect. It’s true; barely a day (perhaps an hour) goes past without the technology industry now telling us that it has engineered more AI into every niche of its coalface.
The database used to be just that i.e. a base of data, a traditional system of record and a data storage repository.
That time has now passed, we expect intelligence every level of the modern technology stack from front-end user interfaces with natural language and speech recognition, throughout our selection of applications and now all the way back to through into our data management layer.
