Generative artificial intelligence may be the hottest topic in enterprise information technology circles these days, but it was only part of the story at this past week’s massive Amazon Web Services re:Invent conference in Las Vegas, both within Amazon.com Inc.’s AWS and among exhibitors.

AWS certainly touted its gen AI offerings. In particular, it is taking a “bring-your-own-model” approach to gen AI, hoping customers will leverage the Amazon Bedrock service with their foundation model of choice.

The company hopes customers will choose one of the Amazon Titan family of models. However, unlike competitors hawking a single model family, AWS welcomes several popular foundation models.

Perhaps the most exciting gen AI news from AWS is the Amazon Q gen AI-powered assistant. Q is surprisingly diverse and powerful, providing assistance to many people in different roles in their organizations, from human resources to finance to IT. Developers can even use Q to generate SQL or build data pipelines with nothing but natural language prompts.

Where did Q get its name? The AWS party line is that they named it after James Bond’s Q: inventive but exasperated. But the buzz at the show was that the name came from Star Trek’s Q: omnipotent but juvenile. You be the judge.

While gen AI certainly stole the spotlight, my vote for the most intriguing news from AWS was its progress extending serverless functionality to various data services, including the Aurora relational database service, the Redshift cloud data warehouse, and the ElastiCache caching service.

Unlike AWS Lambda serverless compute, which leverages the Firecracker microVM technology under the covers, architecting serverless database technologies requires rethinking sharding. Sharding means splitting up large data stores across instances, a decidedly nonserverless approach. To address this problem, AWS invented Caspian, AWS’ secret sauce for extending its popular serverless architecture to its data infrastructure services.

”Caspian is a combination of innovations that span a hypervisor, a heat management planning [memory allocation] system, and a few changes to database engines themselves,” Peter DeSantis (pictured), senior vice president of AWS utility computing products, explained in his keynote. “Together these innovations enable Aurora Serverless databases to resize in milliseconds in response to changing load.”

To leverage Caspian to solve the problem of sharding, AWS rolled out a routing and orchestration layer, or ROL, limitless database, which allows the serverless infrastructure to handle sharding while providing a single endpoint to consumers of the service.

Delivering on the promise of serverless at the scale AWS customers require involves many moving parts, from Caspian to the ROL limitless database to new clock technology, to keep all the pieces in sync. Overcoming this complexity is why serverless data services get my vote for the most important news from AWS at re:Invent.

Highlights on the exhibit floor

AWS was by no means the only story in town. More than 400 exhibitors joined AWS at re:Invent. Here are my standouts:

Giant Swarm GmbH offers a management and operating platform for hybrid and multicloud deployments that extends to fleets of Kubernetes clusters.

This platform provides comprehensive management across on-premises, edge and cloud-based Kubernetes environments and includes observability, scalability and network capabilities. Giant Swarm’s technology is entirely open source, supporting the company’s subscription-based business model.

Why Giant Swarm stands out: Platform engineering teams leverage Giant Swarm to support developers, giving them the operational and management capabilities necessary to implement microservices in hybrid environments quickly and safely.

ChaosSearch Inc. turns Amazon S3 storage service into a real-time analytics database that supports streaming data at scale.

The company places its entire data lake into a single S3 bucket, leveraging S3’s massive scalability and low cost. These economics enable ChaosSearch to maintain all data as hot, meaning the data are ready for real-time queries on a continual basis. The ChaosSearch data lake also supports popular APIs for ease of use and leverages S3 for caching, providing performance at low cost.

Why ChaosSearch stands out: By leveraging S3, ChaosSearch can offer real-time analytics functionality at a small fraction of the cost of competitors. Such economics are especially important with massive data sets, for example, for AI training.

Stonebranch Inc. offers a universal integration platform that enables automations across legacy and modern applications.

Originally a mainframe job scheduler with agents for AS/400 and Windows applications as well as z/OS apps on the mainframe, Stonebranch now offers agentless automation via APIs for cloud-based applications. Stonebranch integrates with modern data infrastructure technologies including DataBricks and Snowflake and features Open Telemetry-compatible observability capabilities for integration with Dynatrace, DataDog and similar platforms.

Why Stonebranch stands out: Other automation technologies focus either on legacy environments or modern, cloud-based applications. Stonebranch bridges both worlds, enabling its customers to create complex automations that span hybrid environments.

Massdriver Inc. has joined the crowded platform engineering marketplace with an offering that enables platform engineering teams to build their own internal development platforms, or IDPs.

Unlike competing platforms, however, Massdriver provides a visual canvas for assembling prebuilt bundles of DevOps tools and custom modules into workflows, simplifying the creation and use of the IDPs.

Why Massdriver stands out: By combining platform engineering, cloud operations and DevOps into a single tool via a visual canvas, Massdriver enables teams to streamline cloud-native application development across the software lifecycle.

ThatDot Inc. has leveraged DARPA research to implement Quine, a streaming graph database technology with previously unachievable scale. Quine also supports continuous query monitoring so that any change to data in the graph will kick off all relevant queries in real-time.

Complementing Quine is thatDot’s novelty detection capability. While typical machine learning-based anomaly detection works with numerical data, thatDot Novelty Detector works with categorical data – that is, any non-numerical information. As a result, thatDot can detect anomalies in a wide range of behaviors.

Why thatDot stands out: Quine and Novelty Detector are disruptive technologies in their own right, but when customers use them together, the combination is especially powerful. ThatDot is set to disrupt the entire AIOps market with this technology.

DayZero Software dba Superblocks offers a low-code platform for developers to build internal tools for employees at their organizations. Superblocks is especially well-suited for mission-critical apps at scale, for example, customer support applications.

Unlike other low-code platforms that downplay the need for coding, Superblocks is explicitly extensible. It combines developer-focused visual low-code functionality with support for coding in multiple popular languages.

Why Superblocks stands out: Developers select programming constructs from pull-down menus, giving developers a quick and intuitive way to build business logic in a visual environment, while also enabling them to write code when necessary.

Don’t mistake the sizzle for the steak

Gen AI may have accounted for much of the sizzle at this year’s re:Invent, but as the examples above illustrate, there is plenty of innovation across the cloud landscape that isn’t part of the AI hype.

For its part, AWS has both ends covered, positioning itself as a gen AI tooling and platform leader while also expanding its innovation in the serverless segment.

Even so, the cloud leader had no monopoly on innovation, as the six re:Invent exhibitors above illustrate. For anyone wary of the excessive hype around gen AI, this diversity of innovation should be an encouraging sign as we enter 2024.

Jason Bloomberg is managing partner at Intellyx, which advises business leaders and technology vendors on their digital transformation strategies. He wrote this article for SiliconANGLE. Disclosure: Dynatrace is a former Intellyx customer. None of the other organizations mentioned in this article is an Intellyx customer. AWS covered the author’s expenses at re:Invent, a standard industry practice.

Photo: AWS