Google LLC made even more product announcements on day two of its Cloud Next conference in San Francisco than on the opening day of the event — this time aimed chiefly at software developers.

Google is very clearly reaching out to the core enterprise customers that have flocked in greater numbers to its principal public-cloud rivals: Amazon Web Services Inc. and Microsoft Azure. All of the enterprise-focused announcements that it made on day one had their clear follow-on announcements on day two:

• The new Anthos multicloud integration fabric announced on day one will be enriched and extended by the new networking capabilities that are designed to help Google Cloud Platform customers build, scale and secure increasingly complex meshes. They’re also protected through the new identity and access management, security and trust, and mobile multifactor authentication infrastructures that the company rolled out.
• The new open-source big-data managed services launched on day one will take root in Google Cloud Platform alongside the new fully managed enterprise database subscription services and be supplemented by the new archival cold storage.
• The new Google Cloud Platform vertical apps rolled out on day one will have plenty of company amid the retail applications and AI-infused document, inventory and customer service applications — many from high-profile partners such as Deloitte, Accenture PLC, and Salesforce.com Inc. — that Google announced on day two.

But the most important theme of Cloud Next 2019 so far has been Google’s clear play for the hearts and minds of next-generation developers. On day one, the top announcement in that regard was Google’s Cloud Run, which shows that Google has made serverless app development a top priority in its outreach to these pivotal users.

Amid the many announcements on day two, Google sent a clear signal that it’s targeting a new breed of developers who rely on low-code, automated, embedded tooling for continuous integration and continuous deployment of cloud-native application code and AI models. Here are the most important day two announcements in that regard:

Automating low-code development of cloud-native apps

For developers of apps for the multicloud, Google’s chief announcement was Cloud Code, which is a set of plug-ins that supports low-code development of apps for deployment into public clouds and on-premises platforms.

This new low-code tool — which is distinct from the existing Google AppMaker tool for building G Suite productivity apps — has the following capabilities:

• Works with Visual Studio Code (in beta) and IntelliJ (in alpha);
• Helps Kubernetes developers to get started by providing an updated set of code templates that are preconfigured for debugging, build and deployment.
• Automates many steps in the process of developing, debugging, compiling and production of code for containerized Kubernetes applications;
• Extends the local code edit-compile-debug loop to target any remote Kubernetes environment, including Google Kubernetes Engine or GKE;
• Embeds Google’s command-line container tools Skaffold, Jib and Kubectl to provide programmers with continuous feedback on projects as they’re being built;
• Allows developers to set up profiles that define different deployment targets, including local development, shared development, test or production;
• Offers the flexibility to test and debug code on the developer’s workstation or in the cloud;
• Incorporates an embedded library manager that adds required dependencies to applications, enables Google APIs automatically on built-out apps and manages any required app secrets; and
• Integrates with existing DevOps tools and services such as Cloud Build and Stackdriver, so that all app configurations are managed as source code in a repo. App log files for any environment are viewable from directly within from an IDE that incorporates these plug-ins.

Democratizing development of sophisticated AI apps by business analysts

Google announced tooling that simplifies creation of cloud-native AI and other advanced analytics apps by business analysts, subject matter experts and other nontraditional developers. These new tools, most of which are in beta or alpha, help to automate and guide new developers in the following analytics-pipeline tasks:

• Data discovery and governance: Data Catalog helps organizations to quickly discover, manage, govern, and understand their data assets.  It is a fully managed cataloging system for capturing technical and business metadata. It provides a search-driven data-discovery interface. For security and data governance, it integrates with Google Cloud Data Loss Prevention to discover, catalog and redact sensitive data assets automatically, and also with Google Cloud Identity and Access Management to enforce data-access permissions.
• Data movement: BigQuery Data Transfer Service automates data movement from SaaS applications to Google BigQuery on a scheduled, managed basis. It enables business analysts to populate a data warehouse for downstream AI and analytics without writing a single line of code. In addition to Google’s first-party apps, it supports moving data from more than 100 popular software-as-a-service applications, including Salesforce, Marketo and Workday.
• Data ingest and integration: Cloud Data Fusion is a fully managed, cloud-native, no-code data integration service. It enables anyone to easily ingest and integrate data from various sources and transform that data, and blend or join it with other data sources, before using Google’s BigQuery to analyze it. It accelerates these integrations with a broad library of open-source transformations and more than 100 out-of-the-box connectors for a wide array of systems and data formats. It allows users to explore and manage all datasets and data pipelines in one console, and enables creation and management of data pipelines though visual drag-and-drop without any coding necessary.
• Predictive insight generation: BigQuery ML, which was announced last year, enables analysts to use familiar SQL to build and deploy AI models on massive datasets directly inside BigQuery. This week, Google announced that it has upgraded the product with more functionality to address additional business needs. Under the covers, it has added k-means clustering and matrix factorization for build customer segmentations and product recommendations. Users can now build and directly import TensorFlow deep neural network models through BigQuery ML. They can also apply machine learning to tabular data without writing a single line of code, using a new feature called AutoML Tables.
• Data pipeline creation: Cloud Dataflow SQL enables analysts to build their own Dataflow pipelines rather than rely on data engineers to handle this task. This tool uses familiar SQL that also automatically detects the need for batch or stream data processing. It uses the same SQL dialect used in BigQuery. This enables analysts to use Dataflow SQL from within the BigQuery UI, and then to join Cloud Pub/Sub streams with files or tables from across their data infrastructure and to directly query the merged data for real-time visualization and insights.

Fostering AI pipeline team productivity with robust team workbench

This week so far, Google’s principal announcement for AI developers has been the beta release of AI Platform.

AI Platform is an IDE for modeling, training and serving of containerized AI applications targeted at Google Cloud, other clouds or on-premises platforms. It provides a comprehensive, end-to-end environment for teams prepare, build, run and manage AI projects. It allows coders, data scientists and other specialists to collaborate, train models and scale AI pipeline workloads from within a common dashboard. It provides tools for developers to discover and build AI pipelines, notebooks and other project assets that can be run unmodified on premises or in Google Cloud.

AI Platform leverages Google’s Cloud AutoML, Cloud Machine Learning Engine, Kubeflow, and AI Hub to support automated DevOps workflow for AI-driven apps. Google announced that it’s working with various partners — including Accenture, Atos, Cisco Systems Inc., Gigster, Intel Corp., Nvidia Corp., Pluto 7, SpringML and UiPath Inc. — to build Kubeflow pipelines that work with AI Platform.

Google also announced new automation tools for creation of AI models from datasets with no coding necessary, for training and deployment of AI computer-vision models to edge devices and for AI-driven discovery and classification of video content.

And that’s a wrap for day two. For further information and insights on all this, check out interviews on theCUBE, SiliconANGLE’s mobile livestreaming studio covering all three days of the show.

Photo: Robert Hof/SiliconANGLE