Joshua Resnick - Fotolia
IBM to deliver refurbished Db2 for the AI and cloud era
IBM has a tuned-up version of Db2 planned, featuring a handful of AI and machine learning capabilities to make it easier for users to send and manage Db2 data across clouds.
IBM is readying a revitalized version of Db2 with new AI-based functions and greater transparency across clouds from multiple vendors, enabling the venerable database to better participate in modern environments.
The upcoming Db2 11.5.5 has been reworked to take advantage of IBM's Cloud Pak for Data offering and its extended services, as well as the extended services of IBM's partner network. Among those extended services from IBM and its partners is an array of open source applications and tools.
By working more seamlessly with the clouds from AWS and Microsoft, IT teams can not only avoid vendor lock-in but also more efficiently put together a hybrid cloud strategy, according to IBM documents shared with TechTarget.
"What [IBM] wants to do is make it easier for Db2 users to move and manage data more easily between on-premise and multiple clouds from IBM and AWS and Azure," said one source familiar with the company's plans. "The new management console they have [with version 11.5.5] should save IT people time, and the new management console gives them a global view of all those environments."
One of the goals of the new version is to give Db2 users a reason to stay with the product in the face of competitive offerings from AWS and Microsoft, according to an IT analyst. Both have woven their database offerings into key AI and cloud technologies.
"This [new Db2] offering is on the right track, but I'm not sure it will encourage users to leave whatever database they already have," one longtime IT consultant said. "The issue with databases is once you've invested in one, you have a lot of custom code built in and it is hard to move to another platform."
IBM is using AI and machine learning to speed queries through the new ML Optimizer, currently in preview. The ML Optimizer uses machine learning and neural networks to learn from past experiences, which helps improve decisions using feedback from query executions and statistics.
The offering provides users with cardinality estimation, an essential capability for successful query optimization. The technology assists database administrators in creating join conditions and expressions, a new machine learning-based workload management feature that provides the ability to create predictive models for query cost, CPU and RAM utilization, and machine learning infrastructure, enabling model training.
IBM is expected to make Db2 11.5.5 public in early to mid-December, according to sources familiar the product.
IBM DB2 containers in OpenShift
Responding to corporate users' growing focus on containers, IBM has incorporated the Db2 C2C (Click2Containerize) tool, which permits users to analyze an existing, non-containerized Db2 database to determine if it is suitable for containerization with OpenShift. After users have gone through the prescribed process of registering an existing database, they can then generate the shell script that provides the framework for containerizing the database. Once this is complete, users still need administrator access to an OpenShift cluster to execute the generated script.
"IBM purposely did not want to use the word migration, with all the negative connotations of that word," said one source. "But essentially, this technology allows users to move existing offerings over, making them hybrid cloud services by just clicking. This avoids some of the manual work associated with migrations. They will position this as an accelerated path over to a hybrid cloud."
Db2 version 11.5.5 appears to complement the technology offered by Instana, a startup IBM acquired last week. Instana's software automatically monitors and collects information about application components running on both public and private clouds, as well as mobile devices and on-premises platforms, including IBM mainframes.
"I would describe [Instana] as a fortuitous acquisition, given how its capabilities fit with Db2 [11.5.5]," said another IT analyst. "It is built using a machine learning model, so it can detect a problem and autocorrect it before everything collapses. They have a very impressive young team there."
Catering to the development community, the new version contains support for REST APIs for both static and dynamic SQL; the ability to run Gremlin queries on Db2 Graph so developers can uncover insights hidden in large amounts of data; and the ability to create machine-learning pipelines as Python user-defined functions, thereby allowing developers to deploy training pipelines directly on the server. The updated version also has new drivers for Node.js and Go.
Bundled with version 11.5.5 is an upgraded Db2 Data Management Console that allows the monitoring of hundreds of databases across Db2 from a single screen. Admins can examine a range of database objects including tables, schemas and stored procedures. They can also write and run SQL in an integrated SQL editor, according to the IBM documents.
The console will come free with all versions of Db2 and Cloud Pak for Data.
IBM officials did not respond to a request for comment.
