As its annual re:Invent conference draws to a close, Amazon Web Services Inc. left no doubt in anyone’s mind that it’s investing deeply in the technologies, tools and applications that it will need to stay the clear industry leader in cloud computing.

If you attended this year’s re:Invent in Las Vegas, your brain is probably on the verge of exploding. AWS jam-packed the keynotes and other sessions with dozens of announcements of new and enhanced cloud services that deepen its differentiation with respect to Microsoft Azure, Google Cloud Platform and other public cloud providers.

As detailed in my day one and day two blogs, AWS’ announcements ranged across its entire solution portfolio, addressing established and emerging enterprise cloud computing requirements in such areas as:

• boosting cloud storage efficiencies,
• simplifying cloud data movement,
• enforcing enterprise controls on cloud data workloads,
• accelerating the pipeline for high-performance cloud apps,
• speeding development of apps for cloud-to-edge deployment,
• enhancing the manageability of apps in serverless, hybrid and edge clouds,
• driving innovative AI into every cloud application,
• building and optimizing diverse data workloads in the cloud, and
• managing rich cloud-native applications across complex deployments.

And that was just scratching the surface. In those prior writeups, I was primarily discussing the announcements that AWS Chief Executive Andy Jassy discussed in his keynote and that were the subject of formal press releases. In fact, there were other announcements, and none of a trivial nature, that were published in blogs written by AWS’ deep bench of technical experts. In addition to those, on day three, some announcements were presented by Amazon Chief Technology Officer Werner Vogels in his morning keynote. Many of these had a developer focus, including:

• Launching a fully managed open-source streaming service: AWS announced the public preview of Amazon Managed Streaming for Kafka. This is a fully managed service that runs applications on Apache Kafka in the AWS Cloud. The user needs no infrastructure management expertise in Apache Kafka, which is the most widely adopted open-source streaming platform. AWS operates highly available and secure Apache Kafka clusters. The service makes it easy for users to migrate existing Kafka applications to the AWS cloud without code changes. It supports Apache Kafka version 1.1.1 and offers Amazon EC2 M5 instances as Apache Kafka brokers.
• Providing new open-source tools for serverless application developers: The company introduced AWS Toolkits for PyCharm (available now), IntelliJ (in preview) and Visual Studio Code (in preview). Distributed under open-source Apache License v2, these new open-source toolkits enable development of serverless applications. They support full creation, step-through debugging and deployment of serverless functions in the integrated development and language of the user’s choice. Developers can invoke Lambda functions locally or remotely, test serverless code locally in a Lambda-like execution environment and deploy applications to the AWS region of their choice. They can also use and customize sample function payloads from different event sources, including Amazon Simple Storage Service, Amazon API Gateway and Amazon Simple Notification Service.
• Publishing an API and runtimes for sharing, discovering and deploying libraries and serverless lambda functions: AWS announced general availability of Lambda Layers and the Lambda Runtime AP. Designed to simplify serverless application development, the features enable developers to use their favorite languages when authoring Lambda functions. Lambda Layers enables centralized management of code and data shared across multiple Lambda functions. Lambda Runtime API enables developers to use any programming language for developing lambda function. AWS is making C++, Rust and Ruby open-source runtimes available now and is working with the following partners to provide open source runtimes Erlang and Elixir (Alert Logic), Cobol (Blu Age), Node.js (NodeSource N|Solid) and PHP (Stackery). The runtimes and layers are available in all regions where Lambda is available.
• Enabling application load balancing on serverless functions: The cloud provider announced a limited preview of AWS Lambda serverless functions as targets for application load balancers. It requires configuration of application load balancers as HTTP/S front-ends for requests coming from web browsers and clients. Information technology administrators can apply this feature by associating Lambda functions with application load balancers on the Amazon EC2 or AWS Lambda management console.
• Support for more complex distributed cloud programming patterns: AWS announced support for a fully managed workflow with eight new AWS service integrations for AWS Step Functions (generally available); accelerated microservice discovery through a managed service registry on AWS (generally available); and a cloud-native microservice mesh on AWS (public beta).
• Discovery of relevant AI model training runs: The company announced the Amazon SageMaker Search capability, which allows AI developers to easily search for the most relevant training runs with respect to their current modeling experiments. It supports streamlined assessment of the most relevant model training runs from hundreds and even thousands of Amazon SageMaker model training jobs.
• Integration of SageMaker modeling notebooks with Git for improved data science team collaboration: AWS said that SageMaker users can now associate GitHub, AWS CodeCommit and any self-hosted Git repository with SageMaker notebook instances. This enables users to collaborate easily and securely and ensure version-control with Jupyter Notebooks from within SageMaker.

AWS is executing brilliantly on many levels. Nevertheless, it seemed a bit lacking at this year’s re:Invent in several respects. Going forward, Wikibon would like to see the company address the following issues, all of which are pivotal to its retaining its substantial leadership position in public cloud:

• Fostering a vibrant partner ecosystem within AWS’ cloud services marketplace: That was a significant focus in this year’s AWS Marketplace announcements. AWS indeed has a substantial global partner ecosystem, many of which are selling through Marketplace. But this year’s re:Invent featured an AWS new-product blitz so overwhelming that it eclipsed any attention that innovative partners otherwise might have gained at the event.
• Empowering customers to build composable cloud applications that span AWS’ cloud service portfolio: The development tools AWS rolled out during re:Invent 2018 addressed the usual programming silos — containerized, serverless, AI and others — but didn’t seem to compose into a broader multicloud or even hybrid-cloud-focused development portfolio. As the multicloud development and management tooling market expands, we expect that AWS will ramp up its partnerships and organic development of tools for users who’ve decided not to put all their IT eggs in the Seattle-based company’s cloud basket.
• Delivering AI-powered conversational user interfaces in AWS’ cloud: AWS has a powerful portfolio of AI services for natural language processing, IoT analytics, mobility and so forth, but re:Invent 2018 seemed to overlook the growing cloud developer focus on chatbots and other embedded conversational user interfaces. Yes, Alexa was omnipresent in keynote presentations and the like, but that market-dominating AI-powered conversational technology was conspicuously absent from the product announcements, partnerships and roadmaps.
• Sustaining a truly open AI cloud-native development ecosystem: Although AWS has built SageMaker into a category-dominating data-science toolchain solution, what was missing from re:Invent 2018 was any clear discussion of how it plans to evolve the service to incorporate emerging industry open-source projects such Kubeflow that are standardizing the containerization of the end-to-end data-science DevOps workflow.

One final note: In his keynote, Vogels did a masterful job of responding, if indirectly, to recent comments by Oracle Corp. Chairman and Chief Technology Officer Larry Ellison to the effect that AWS’ cloud databases are not mature enough for enterprise-grade deployments. Vogels presented a very well-structured discussion of how AWS’ Aurora cloud relational database has been engineered from the start to support Amazon’s most demanding hyperscaling requirements with superior performance and availability. Check out this recent article of mine for a dissection of that controversy.

Also, check out what AWS executives, partners and customers had to say this week on theCube at re:Invent 2018, with more interviews slated to post in the coming days.

Photo: AWS