EnterpriseDB (EDB), the leading Postgres data and AI company, is announcing significant enhancements to EDB Postgres AI (EDB PG AI), the platform designed to help enterprises deploy and scale AI securely and compliantly across their Postgres environments. Now, featuring a new low-/no-code AI pipeline creation experience and hybrid management, EDB aims to meet customers where they are as the complexities of the AI era continue to grow.
As one of the most popular database systems in the world, EDB is uniquely positioned to supply AI with what it needs most—data, according to Jozef de Vries, chief product engineering officer at EDB. With this strategic advantage, EDB expands the power of AI to the Postgres ecosystem, “enabling customers to stay in those environments [and] not complicate their data estate,” said de Vries.
In alignment with this goal, EDB is introducing a new low-/no-code experience for developers and business users to intuitively build intelligent applications. This new capability expands EDB into the application services space, now allowing “our users to directly start building AI applications within our platform system itself…connected directly into the Postgres environment that they've been working with,” de Vries explained.
Through a simple point-and-click interface and a low-code software development kit (SDK), users can rapidly create applications in as little as days, according to EDB. Setting up AI pipelines—capable of syncing embeddings with source data for a continuously fresh AI knowledgebase—takes just five lines of code, eliminating costly infrastructure management.
Further, with the assistance of NVIDIA accelerated computing, NVIDIA NIM microservices, and NVIDIA NeMo Retriever, EDB PG AI integrates with NVIDIA’s extraction, embedding, and re-ranking microservices as well as the latter company’s AI model extension capabilities.
With these NVIDIA integrations and frameworks, “The goal here is to extend that notion of data sovereignty [and] data privacy to even the LLMs [large language models] that our platform and the end users’ AI applications are interacting with,” noted de Vries. “For our customers who don't want to send their data out to Open AI LLMs, for example, they can invoke our model serving capabilities, pull down these NVIDIA NIM LLM packages and run those LLMs locally, thus keeping their data within their sovereign or secure environment.”
EDB PG AI also paves the way for greater observability and management with support for a new market paradigm: intentional hybridity. As the race to the cloud slows—due to costs, data access and ownership challenges, data security concerns, etc.—many organizations are embracing a purposefully hybrid environment to mitigate common cloud complexities. Yet, even as the push for cloud decelerates, the AI surge ever-increases.
"It's really this interesting point in time in the market that we are finding ourselves, where we have this progression…[in] taking the foot off the gas pedal towards racing to the cloud but also putting the foot more down on the gas pedal and racing towards AI,” pointed out de Vries. “In letting all of that control and all that data go to cloud providers, they're in a situation where the more they want to embrace AI technologies, they have to figure out what access, what control they're able to relinquish.”
As a means to address this widening priority gap, EDB PG AI is introducing hybrid management, offering comprehensive visibility across the entire Postgres estate, including real-time insights across hundreds of databases. Without needing any DBA expertise, users can rapidly identify and resolve issues up to 5x faster, boosting application performance by up to 8x, according to EDB. EDB PG AI’s hybrid management allows teams to intuitively optimize infrastructure, increasing productivity by up to 30% while reducing TCO by up to 6x compared to legacy systems.
“We're providing this hybrid platform solution so customers have an answer to the strategies that they see going forward that don't entail 100% of focus on the cloud, while also including capabilities into the platform at the data management layer, at the LLM layer, at the application layer, so they can still embrace these new, emerging AI technologies, without having to compromise their data security or their data sovereignty,” de Vries explained.
Other new enhancements for EDB PG AI include:
- Data security at all layers through Transparent Data Encryption (TDE), supply chain security, and hardened container images from Iron Ban, featuring additional security features such as role-based access control with fine-grained permissions down to row level, robust audit logging for real-time threat detection, and data redaction to limit sensitive information exposure
- Purpose-built PG AI Analytics Engine increasing query performance across Postgres by independently scaling from storage and being optimized for columnar formats, including Iceberg and Delta tables
- A universal, secure data store that supports all types of data models—SQL, vector, JSON, time-series, key-value, and more—for building applications with structured, semi-structured, and unstructured data types
According to William McKnight, president, McKnight Consulting Group, “McKnight Consulting Group compared the integrated EDB Postgres AI platform against a usual DIY approach using AWS. Across all the eight critical components of an enterprise AI factory, EDB Postgres AI reduced overall complexity by 67%, delivered a 3x faster design-to-delivery capability (from seven months to nine weeks), and a 38% reduction in maintenance complexity and costs.”
To learn more about EDB’s latest updates, please visit https://www.enterprisedb.com/.