How has the Generative AI (GenAI) boom impacted database technologies since November 2022?
Since 2016-17, there has been a new category of databases called vector databases, which basically enables similarity search. But they struggled to find funding. They had niche use cases for similarity match until the GenAI revolution, when suddenly people discovered a good use case for them would be to augment interactions with the Large Language Models (LLMs) using your private content. So, if you are an enterprise and you have a vector database and you have corresponding documents or images that you want element to process, you can use the vector database to augment the LLM. This created a massive sudden immediate need for vector database technology.
How did this impact Oracle Database?
What we found in Oracle Database was the big weakness that these single purpose vector databases have is that’s all they can do. They do not really have the business data. Most business applications need business data as well as vector data to run real world queries. Hence, we decided we need to add vector technology into Oracle Database that already has the business data. So, we had a very good starting point and we know how to build indexes and large data representations. It was a very quick pivot.
What traction has Oracle Database seen in the past year or more in India?
India is a massive hypergrowth economy. I moved to the United States 30 years ago. But I am from Calcutta. Every time I come here; I am impressed by the new use cases that people have found for Oracle Database. And Oracle has a very big footprint in India. It is one of our highest growth markets. Banking in India is in a different league because of UPI and the volume of payments. So, Oracle is very successful in India, we partner with some of the largest banks, telcos and government agencies in India. It is still very early to look at adoption of vector technology because it is going to be available when 23C is fully available. We have a limited availability programme right now. Worldwide, we have gotten a lot of interest from customers. But we expect India to be a place where this is a technology which will be very widely adopted.
India must be difficult given the sheer scale required…
It is about the scale and volume. Further, the regulatory landscape in India is stringent for valid reasons, primarily to combat fraud and ensure system stability and uptime for banks. Unlike more established Western economies, India's fast-growing economy necessitates constantly evolving regulations. This dynamic environment poses some challenges, especially for the financial services sector. However, amidst these challenges, every company I talk to is actively addressing regulatory compliance by establishing specialised task forces or centres of expertise. It is worth noting that India faces immediate and pressing issues, but there's a concerted effort from businesses to adapt to this.
How does Oracle address the issue of ‘scale’ in India?
When it comes to scaling, we have a diverse range of technologies at our disposal, perhaps more varied than any other database. We can scale out the database. First of all, we can run on the world's largest computers. I distinctly recall running a workload on a system with 3000 processors and Oracle performed flawlessly up to 3000 cores. Additionally, we can scale out across multiple computers using a technology known as Real Application Clusters, which seamlessly distributes the workload across machines. Additionally, we can scale out across multiple computers using a technology known as Real Application Clusters, which seamlessly distributes the workload across machines.
Further, for even greater scalability, we offer the option to partition the Oracle Database into shards, each functioning as a sub-database while providing a unified global perspective of a distributed database. This versatility in scaling options is crucial, especially in regions like India, where there's a growing demand for scalability and continuous availability. Businesses require 24/7 access to data and cannot afford transaction downtimes due to regulatory constraints. These demands are only expected to intensify over time, particularly with the ongoing transition of cash transactions to digital platforms like ATMs and mobile payments, which will further increase the influx of data into our systems.
Also Read: Oracle Cloud Infra Sees 50% YoY Growth In India During Q2FY24