Key Takeaways from Our Webinar with Dataversity
Every January, a lot of folks in this industry do their best to guess which way the data management wind will blow this year. Donna Burbank, data architect and Managing Director at Global Data Strategy, and Lalit Ahuja, CTO of GridGain discussed this on a recent webinar. Donna also presented some of the findings from a recent Dataversity survey on Data Management Trends in 2025.
Lalit started things off with a walkthrough of the complex process involved in meeting the real-time data needs of modern enterprise applications, especially for AI or advanced data analytics. There’s an entire computation sequence that often happens – streaming data transformation, feature extraction, vector embedding, and applying machine learning models for inference.
He expressed a strong opinion that simplifying architecture is essential. There’s no reason why that entire in-memory compute grid can’t be combined with the low latency cache, the data hub, AI data store, and durable long-term storage to all seamlessly merge into a single technology solution. It would vastly simplify architecture and reduce decision application latency.
Key Takeaways from Lalit Ahuja’s presentation:
- Data-driven enterprises need real-time processing.
- Combining streaming and transactional data with historical context and complex computations can minimize latency.
- Unified real-time data platforms deliver business outcomes while simplifying the enterprise data ecosystem.
Donna continued the conversation by saying that, with all the discussion about AI in every information outlet, it’s becoming difficult to separate hype from reality. The good news is that Dataversity conducted a 46-page Trends in Data Management survey with hundreds of global respondents, allowing us to look at solid data on this subject.
One interesting finding she presented was that “saving cost and increasing efficiency” was the number two business reason for implementing data management, after “reporting and analytic insights” in the top spot. Many times, organizations think of the data management and analytics teams as cost sinks, but an excellent reason to have those teams is to save the organization money.
Graph showing responses to the question: What are your main business goals and drivers for implementing Data Management in your organization?
It was interesting to see some of the things the survey respondents said about the relationship between AI and data management efforts. A lot of times, the focus on LLMs, AI, and machine learning in the media overshadows the fact that none of these new cutting-edge technologies work unless they are fed the right data. Data management is “critical to ensure the reliability, transparency, and ethical use of data in AI systems.”
Legacy technology, such as mainframe applications, became a hot topic for discussion when the Dataversity survey showed 30% of respondents still had mainframes as a major part of their data management and analytics landscape. Lalit, an ex-financial services data manager, pointed out that the risk of switching to a new technology is often high in an industry that is very risk averse. Every data manager at a bank tends to kick the can down the road, delaying any action around that for yet another year. “Not on my watch.” And mainframe applications just keep chugging along, working without issue, often without anyone left at the company who knows how they work.
Some of the most interesting points made during the webinar were during the Q&A. A lively discussion started when Lalit mentioned that he thought relational databases would, one day soon, become legacy technology. When Donna asked why he thought that, he pointed out that a lot of data is unstructured nowadays and doesn’t necessarily fit in the old boxes like third normal form (3NF). Relational databases have their limitations. “People are expecting more flexibility in how they access data, how they manage data, how they work with it,” Lalit added.
Donna countered that folks will always need to be able to find the single source of truth for some core parts of the business. “Where’s my customer’s address?” Organizations need that. Her belief is that over time, databases will evolve, rather than fade away.
Lalit agreed that for some uses, relational databases will continue to be needed, but as more and more use cases demand real-time and data that isn’t traditionally structured, they will fade in importance. After all, mainframes are still chugging along, doing a lot of work in organizations, too.
Overall, it was interesting to both see the survey results signaling where our industry is headed, and hear two great industry minds with different perspectives discuss the state of where we are and where they each believe we’ll go from here.
Check out the on-demand Trends in Data Architecture 2025 webinar for yourself.