Blogs

How enterprises benefit from a database powered by and built for AI

Offering Management, IBM Cloud and Cognitive Software, IBM

More is expected of databases today than ever before. They must simultaneously operate with extreme efficiency while opening the door to innovative opportunities like artificial intelligence (AI). For this reason, businesses should consider a database that is both powered by, and built for, AI. It should be powered by AI to help increase optimization and usability. And it should be built for AI to help developers and data scientists leverage the latest technology to drive deep insights.

The premier example of such a database is Db2 11.5, which IBM announced as available today. Many companies currently rely on Db2 for supporting their mission-critical applications due to its high reliability, resiliency and security. With this new version, IBM aims to transform Db2 into the AI database for IT organizations and developers by adding machine learning (ML) and additional capabilities in updates to come.

Powered by AI

A database powered by AI enables businesses to focus better on value-added tasks. Instead of manually optimizing the database for improved cost savings, ML can help continually boost efficiency in a more automated way. AI also helps users arrive at insights more easily. In particular, you can more readily use systems that employ natural language in their search functionality or reports. As a result, users can put more time into acting on insight, rather than dealing with the underlying database.Db2 11.5 is the database powered by, and built for, AI

Recognizing this, Db2 11.5 will feature a tech preview of a machine learning-based optimizer. ML Optimizer will supplement the existing cost-based optimizer to further enhance and improve performance. As a tech preview, it will initially focus on improving cardinality estimation. Another tech preview of a new, no-cost tool – Augmented Data Explorer (ADE) – will also be available for Db2 11.5. ADE uses natural language querying, faceted search and natural language generation capabilities to make it easy for data scientists, data engineers and business analysts to easily explore the data stored in Db2. Try out the ADE beta today.

Built for AI

It is not enough for a database to be optimized using AI. Developers need to easily leverage AI in their applications as well. Support for popular languages and frameworks allows them to save time that might have otherwise been wasted figuring out something new or proprietary. Data federation capabilities open new sources of data for AI models. And, access to compatible code patterns for AI application development means users can jump into their projects immediately and return a finished product more quickly.

Db2 11.5 offers all of these features. It has native language support for Python, Ruby, Go, Java, PHP, Node.js and Sequelize, and it supports popular frameworks such as Visual Studio Code and Jupyter notebooks. Db2 11.5 will also feature blockchain federation support to help application developers pull data directly from blockchain sources and combine that with other data sources for analytics or dashboards. In addition, code patterns are now available to accelerate AI application development.

Simple to upgrade

To get the most out of AI database capabilities, organizations must choose a database that fits the size and scope of their projects. This includes the ability to easily grow with the company as well. In so doing, they help ensure that they will have the storage and computing power needed to fully leverage AI models.

Db2 recently consolidated their editions to three to make it easy for customers to adopt and upgrade their deployments. Users can start out with the community edition at no cost, which has the full capabilities of Db2—including high availability, disaster recovery and pureScale. It is available as a native install, docker install or managed service in the cloud. Then if users want to increase their memory, cores, or database size they can move up to the standard or advanced editions. 

Additional improvements

As they add new AI capabilities, businesses cannot ignore the need for more traditional database features like resiliency and performance. Newer databases have improved these beyond the already impressive functionality in previous versions.

For example, Db2 11.5 has made enhancements to pureScale, its advanced clustering capability for better resiliency, including:

  •       Support for currently-committed semantics across members
  •       Improved cluster-wide free space management
  •       Better performance of load/drop with range partitioned tables
  •       Default enablement of online index creation and extent claim
  •       Security feature enhancements

IBM has made many performance and usability improvements to Db2, including:

  •       Compression enhancements
  •       Built-in spatial support
  •       ETL optimizations
  •       Additional SIMD exploitation
  •       Warehouse query optimization
  •       LOB support to the columnar engine, commonly referred to as the BLU engine
  •       Support for 4K sector size storage, benefitting modern storage devices
  •       Support for external tables, making it easy to load or archive data from Db2 

As businesses get better at managing and leveraging a wider variety of data, they should look for ways to adopt and exploit AI with their database. To do that, they’ll need one that uses AI to drive efficiency, supports the use of AI applications, can be upgraded easily and has a history of performance and resiliency.

Db2 11.5, the AI database has been purposefully built to address all of these needs and will continue to make updates to its AI capabilities throughout the year. With this new vision, Db2, the core component of the IBM Hybrid Data Management Platform will also enable customers accelerate their AI application development, while automating some of their data management.

Explore everything Db2 11.5 has to offer.