Oracle CloudWorld ran this year from the 9^th through the 12^th of September at the Venetian convention center in Las Vegas Nevada, USA. It was jam packed with exciting product and cloud service announcements, keynote addresses, customer panels, and demos. Vendor trade shows can be overwhelming to some or complete snooze fests to others. Neither was the case at Oracle CloudWorld 2024. In fact, the show had a quite perceptible and persistent electric buzz going on. The air was palpable with attendee excitement.

After keynotes and breakout sessions, attendees could be heard discussing among themselves how they can take advantage of these new Oracle innovations. That electric atmosphere kept amping up after the technical sessions that provided useful pragmatic implementation processes, deep-dives, and demos.

What was causing all this excitement? It began on Monday, 9 September 2024, with Oracle’s press release of their strategic partnership with Amazon Web Services (AWS). This announcement has massive implications that will be discussed shortly. On top of the AWS partnership was Oracle’s very impressive blowout Q1 FY2025 financial results. Thoses two anouncements alone would have been enough to create huge buzz. But those announcements were simply the start of a noteworthy week. 

To get a feel for the “why” behind the buzz requires a step back to look at the overall strategies Larry Ellison, CTO Oracle laid out at his Oracle CloudWorld 2024 keynote. They are:

1. Multicloud
2. GenAI Everywhere
3. Security

Multicloud

Oracle’s partnership with AWS – Oracle Database@AWS – is the latest proof point of Oracle’s ironclad commitment to Multicloud. AWS is the third major public cloud provider – after Azure and Google Cloud – to partner with Oracle Cloud Infrastructure to deliver Oracle’s highly advanced database services running on Exadata. Services include Exadata Database Service and workload-optimized Autonomous Database – Autonomous Transactional Database, Autonomous Data Warehouse, and Autonomous JSON Database. 

Oracle essentially puts a deployment of Oracle Cloud Infrastructure (OCI) inside their partner’s cloud. This colocation provides the same unparalleled performance, capabilities, and security as in the Oracle Cloud. Oracle worked with AWS, Microsoft Azure, and Google Cloud (GCP) to deliver seamless integration with each cloud’s user interfaces and user authentication making those Oracle Database services appear and provision as native to each specific public cloud. Oracle then makes those services available through the partner’s marketplace at the same price and performance as in the Oracle Cloud.   

Data can move freely without egress fees between the Oracle databases in Oracle and AWS, or Azure, or GCP. This Oracle multicloud strategy simplifies Enterprises’ migration to any of the major public cloud service providers by enabling them to use the cloud and cloud services of their choice with the most advanced databases on which they run their business. Multicloud eliminates the tough choices Enterprises have historically had to make.

Further proof of Oracle’s multicloud commitment is their announcement of new Google Cloud and Azure regions where Oracle’s Database services are now available with more regions planned in the near future.

This is a very big deal because it finally breaches public cloud walled gardens. Organizations with on-premises data centers moved away from walled gardens decades ago. And although most IT organizations have a primary public cloud service provider, they really don’t want to be locked in. They want multicloud options if the walls – such as egress fees – can be torn down. Oracle is the first to do so.

GenAI Everywhere

AI and especially GenAI has been the buzz de jour since OpenAI came out with the first public GenAI LLM. The blush has somewhat come off the rose at this point. Every IT organization is determining how to take advantage of it; however, there is friction. Friction from data gravity – where data originally lands is where it tends to stay. That means the data the LLM needs access to for training or RAG too frequently has to be moved or migrated from where it currently exists to where the LLM can access it. There’s additional friction from a lack of integration with mission-critical or business-critical applications. That doesn’t take into consideration the friction from the cost of the GPU hardware needed to train the models.

It is these frictions that Oracle addressed in a big way at Oracle CloudWorld 2024. Starting from the top of the stack down through the infrastructure. Here are the proof points.

Mission-Critical/Business-Critical Applications GenAI LLM Integration

Oracle Fusion Applications

Oracle announced 50+ new AI agents across Oracle Fusion applications including ERP, HCM, SCM, and CX. These AI agents are powered by the latest GenAI evolution and are specifically architected to quickly jumpstart productivity. They can effectively take over the frequent, repetitive tasks that human beings do not do well with. This enables users to focus their valuable time on more strategic tasks and initiatives. 

These new AI agents enable end-to-end business process automation, provide actionable insights, provide content, and make recommendations.  

Oracle Fusion Application Cloud

• Oracle Cloud ERP – Integrated AI allows several financial processes such as management reporting and cash forecasting to be optimized while increasing sustainable analytics performance and accuracy. A powerful aspect of Oracle Cloud ERP is the predictive cash forecasting. It uses AI models that accelerate insights while creating prescriptive/constant cash forecasting based on the user’s periodic timeframes of choice. Its narrative reporting includes financial performance, variance explanations, and impacting trends means management reports and insights occur much faster accelerating smarter decisions.

Then there’s the incredibly powerful Oracle Fusion Cloud Sustainability automated transaction records. This allows users to leverage the integrated AI for classification rules, sustainability, and metadata attributes that automatically create activity records while adding transactions to a sustainability ledger.

• Oracle Cloud HCM – Integrated AI enables a more complete organization skill and talent catalog, accelerates insights into those skills and talents, then aligns them more effectively with business priorities. The AI powered data enrichment combines the data in the organization’s catalog with 3^rd party sources.
• Oracle Cloud SCM – Improves manufacturing and operational performance via AI powered shift reporting plus the smart ops workbench. These identify the issues impacting production with real-time insights. The integrated GenAI in the Oracle Order Management then assists in the writing of order acknowledgement emails and order change history notes.
• Oracle Cloud CX – Integrated GenAI increases capitalization of sales opportunities, makes upsell and cross-sell conversion more common. It enriches personalization by assisting email authoring, activity summaries, executive summaries for quotes and proposals, and while also answering contract related questions. The Oracle CX Unity AI integration has the ability to provide next best action recommendations based on signal detection according to role title, and aggregated topic engagement.
• Oracle Fusion Data Intelligence – Is getting a GenAI powered development assistant. It’s architected to assist developers by streamlining configurations. It provides a guided process to speed up third-party data source additions.

Mission-Critical/Business-Critical Database GenAI LLM Integration

Oracle Database 23ai

Oracle Database 23ai, comes with JSON Relational Duality Views, Vector Embedings, Vector Indexing, AI Vector Search, and APEX. All of these capabilities greatly simplify GenAI development. What’s new from Oracle CloudWorld 2024 is generative development for enterprise or GenDev. 

What’s exciting about GenDev is it’s a smart AI-centric application development infrastructure. This infrastructure is unlike any other in that it tightly integrates with all of those marvelous Oracle Database 23ai capabilities.

Oracle Autonomous Database

Oracle Autonomous Databases provides additional GenAI capability, performance, and functionality on top of Oracle Database 23ai. Capabilities such as “Select AI.”  Select AI integrates with a number of LLMs, and allows users to interact with their databases using natural language. It then converts natural language to highly efficient SQL queries, simplifying and automating the rapid use of GenAI LLMs.

By learning effective SQL query patterns from curated training data, LLMs can produce more efficient queries that run faster. As part of Oracle Autonomous Database, Select AI inherits all security and authentication features of the database. And because it’s integrated with Oracle SQL it’s available through any SQL IDE, SQL Developer Web, Oracle Application Express (APEX), and Oracle Machine Learning Notebooks.

Select AI has been generally available in the Oracle Autonomous Databases since Oracle CloudWorld 2023. What’s new from Oracle CloudWorld 2024 is:

• Select AI RAG (Retrieval Augmented Generation) – automates use of AI Vector Search to augment Generative AI with up-to-date corporate, often private, data to yield much more accurate and contextually relevant responses
• Graph-based RAG
• AI data pipelines: prepare, load, and transform data using natural language, as well as create AI pipelines to generate  text and image vector embeddings.
• Data Load that automatically detects Personally Identifiable Information (PII) data and can also enrich data using AI expressions.
• Table Assistant: easily edits data and columns using natural language. 
• Synthetic data generation. 
• Run models directly in the Autonomous Database – Import ONNX transformer models to enable in-database vector embedding and leveraging ready-made and pretrained GenAI models.
• Run models directly in the Autonomous Database powered by GPUs – AI data operation performance accelerated by fully-managed access to NVIDIA GPUs. Initially, customers can take advantage of Oracle Machine Learning Notebooks that use GPU-enabled Python packages for resource-intensive workloads. It’s much simpler, faster, and integrated.

Autonomous Database is now integrated with 35 different LLMs across 7 providers, including but not limited to OCI Generative AI, OpenAI, Azure OpenAI, Cohere, GCP Gemini, Claude, Hugging Face. Customers choose the LLM that works best for them. The Oracle partnerships with AWS, Azure, and Google, has made the Autonomous Database in its full glory available in their clouds as well at the same price as on Oracle Cloud.

Oracle HeatWave GenAI

Ever since Oracle HeatWave was introduced in 2020 it has been a huge success in the market. The newest in-database capability is HeatWave GenAI, now available on AWS in addition to OCI, is even more impressive than previous capabilities: HeatWave MySQL, HeatWave AutoML, and HeatWave Lakehouse.

HeatWave GenAI comes loaded with in-database LLMs, automated in-database vector stores, scale-out vector processing, and a built-in HeatWave Chatbot. It also comes with AutoPilot. AutoPilot delivers machine learning (ML) powered automation that  improves performance and price-performance while increasing the productivity of both database administrators and developers.

Oracle announced at Oracle CloudWorld:

• Always free tier on OCI.
• GenAI, Lakehouse, Javascript, Autopilot indexing, Private link, MySQL HA, availability on AWS.
  • It’s been available on all OCI regions.
  • And Azure regions with high speed low latency links to OCI regions.
• GenAI multi-lingual support, LLM batch processing, automatic vector update.
• AutoML storage of large ML models, data drift, semi-supervised log anomaly detection. 
• Lakehouse writing results to object store, automatic change propagation.
• MySQL support for Javascript, Autopilot indexing, Hypergraph optimizer, bulk ingest, Ops Insight.

Mission-Critical/Business-Critical Application Development

Oracle  AI RAG Agent

The AI RAG agent is an integral part of the OCI GenAI Agents service. First announced as a beta in January, it became GA at Oracle CloudWorld 2024. The purpose of the AI RAG agent is to help enterprises quickly and confidently build conversational interfaces to their enterprise data.

Oracle Code Assist – Beta

Oracle Code Assist is a GenAI-powered programming assistant. Just released to beta trials. It’s currently optimized for Java and should enhance application development on OCI. 

OCI Kubernetes Engine AI Enhancements

New capabilities on the OCI Kubernetes are designed to simplify AI workloads and cloud-native applications on Kubernetes at scale.

OCI GenAI Infrastructure

Oracle is now taking customer orders on massive GenAI LLM infrastructures with 131,072 NVIDIA Blackwell GPUs in its OCI Supercluster—announced as the world’s first Zettascale computing cluster. OCI Superclusters are designed for LLM training – including the largest LLMs in the world with massive data sets. Configured with either NVIDIA B200 or GB200 NVL72 liquid-cooled bare-metal instances, the newest OCI Superclusters can use NVLink and NVLink Switch allowing up to 72 Blackwell GPUs to communicate with each other at an aggregate bandwidth of 129.6 TBps in a single NVLink domain. Blackwell GPUs are forecasted to be available in the first half of 2025. Pricing has not been released yet.

Security

The rise of privacy laws, regulations, data breaches, and ransomware has made security top-of-mind for every Chief Security Officer (CSO). Oracle has always been known for its security chops and bonifides in its databases, applications, and OCI. Oracle’s very first customer more than 4 decades ago was the CIA and they still use Oracle. Cyber security is an absolute must for all data environments.

The latest cloud security innovation from Oracle is “Zero Trust Packet Routing” or ZPR – pronounced zipper. ZPR is based on the open standard that Applied Invention and Oracle designed in 2023. ZPR tackles data security in a novel way It controls network traffic based on the resources and data services accessed, while empowering OCI customers to simply describe, using natural language, what can talk to what using security attributes and policies. For example, ZPR can ensure OCI customer databases are only ever reachable by API endpoints irrespective of IP setting changes, TCP ports, moves to a public subnet, etc. 

What makes ZPR so powerful is that it separates the network security from the underlying infrastructure. Its policies are independent of the network configuration. Those policies can be written before, after, or during network configuration. Security constraints survive any network architectural changes since they are evaluated independent of network configurations. ZPR protects OCI customers from the most common security compromises, network misconfigurations.

Conclusion

Oracle CloudWorld 2024 was a big hit for everyone that attended. The atmosphere was electric. Attendees were pumped by Oracle’s commitment to Multicloud, GenAI Everywhere, and Security. These themes were not just visions or vaporware. They were backed up with real products, services, demos, and most importantly, production customers. 

Oracle CloudWorld 2024, by any measure was also highly consequential for the marketplace. Oracle clearly demonstrated its engineering and development engine is firing on all cylinders delivering world-class innovations in multiple essential fields.