Application-centricity is everywhere, at every level. There are concerted moves across the technology industry focused on making system-level infrastructure services more intelligent, more widely orchestrated and essentially more automated. In simple terms, we need better backends if we want to try and achieve better frontends. The future of efficient enterprise software application development needs to be an essentially holistic play i.e. from correctly provisioned and configured lower backend infrastructure services all the way through to the front end user interface.

Substrate Checkmate

The fact that cloud-native applications (in so many instances) fail to be wholly integrated with the lower substrate services that underpin them is enough to upset any self-respecting cloud architect. Why then, in the age of the cloud, do cloud services sometimes appear to suffer from a disconnect to the operational DNA of the apps they are supposed to serve? Why is the core lower tier of computing not as tightly woven into the fabric of the tier above and below it?

Google Cloud wants to answer these questions and provide a way forward. The company has detailed what it now calls a new application-centric approach that puts the application at the center of developers’ cloud experiences. The company’s latest platform extensions enable software developers to plan, operate and manage infrastructure at the application layer with full business context, instead of getting siloed information from individual services that applications run on.

Focused on what he calls “every step” of the software application development lifecycle is Brad Calder in his role as VP and GM for Google Cloud. The company is putting applications at the center of the cloud-native developer experience by abstracting away the infrastructure complexities of the traditional cloud model. Software engineers will now be able to design, observe, secure and optimize their applications at the application level itself, rather than having to perform these procedures at the infrastructure level.

It’s applicaion-centricity, but from the ground up this time, basically.

Force Feeding Foundational Factors

The Google Cloud engineering team has focused on inefficiencies that come about as a result of the traditional resource-centric cloud model, which introduces an inherent degree of complexity to application design, deployment and management. This reality ends up forcing developers to translate business requirements into lower-level infrastructure details.

In other words, let’s say a business team wants a new big data analytics application to extract and analyze customer behavior patterns from a supermarket retail point of sale system. With a responsibility to shoulder the resource management side of the resulting application, the development project leader will need an appreciation for the number of customers served, the scale of products to be processed, the speed of transactions (some could be real-time, some could be slower batch data loads) and perhaps even cyclical seasonal changes. None of which is really focused on what the business wanted i.e. a customer analytics tool. This inevitably obscures the application’s overall purpose and performance goals.

“The traditional resource-centric cloud model unnecessarily complicates application design, deployment, and management. Developers spend too much time translating business requirements into complex infrastructure details, losing sight of the application’s core purpose and performance goals… and making it difficult to optimize and troubleshoot. Our application-centric approach solves this by putting applications at the center of the customer's cloud experience. Developers can then focus on building, securing and managing applications that deliver business value, instead of focusing on infrastructure complexities,” stated Calder.

The new Google Cloud Platform service is known as Application Design Center and is designed to help cloud platform administrators and developers streamline the design and deployment of cloud applications so that they are secure and aligned with best practices. In addition to application programming interfaces and a “gcloud” command line interface (the hard coding line that bona fide software programmers use to enter language syntax and actually build applications), Application Design Center provides a visual canvas-style approach to designing and modifying application templates. It also enables software engineers to configure application templates for deployment, view infrastructure as code in-line and collaborate with teammates on designs.

Central Code Command Center

“The goal for creating Cloud Hub, is to create a place where developers and operators have at their fingertips the vital insights for their applications, resources and cloud environment. Cloud Hub provides that unifying powerful view into a deployment’s health and troubleshooting, optimization, maintenance, quotas, reservations and support cases. It allows users to leverage these insights to initiate contextual assistance through Gemini Cloud Assist,” enthused Google Cloud’s Calder.

Google Cloud has twinned news of this new service with more information detailing its work in AI-assisted application development and management. With Gemini Code Assist and Cloud Assist the company says it is focused on software engineers to accelerate the application development and streamline your cloud operations. Gemini Code Assist now adds features that include AI agents that can translate natural language requests into multi-step, multi-file solutions. Other new tools that make it easy to connect Code Assist to external services, third-party partners, or even other agents

Google Cloud’s Sundar Pichai has stated that some 25% of Google’s code is written by AI and the organization makes note of several customers who are following suit. But when agents are doing work in the background, Calder insists that it’s important to be able to understand which tasks are underway and what they’ve completed.

“Now that developers have agents working on their behalf, it’s essential they have clear visibility into ongoing and completed tasks, along with a direct way to interact with the agents for those tasks. That’s why we created the Code Assist Kanban board. The Kanban board provides a real-time display of all tasks that Code Assist is actively working on, alongside a record of completed items. In addition, it offers developers a direct channel to engage with our agents, providing feedback, clarifying requirements, and initiating further actions as needed. To take full advantage of Gemini Code Assist in a developer’s integrated development environment, engineers also need instant access to external knowledge. Gemini Code Assist tools deliver just that, where we provide pre-built connections within the chat to Google apps and essential partner tools like Atlassian, Sentry, Snyk and more,” said Calder.

Infrastructure Initiatives Initially

Google Cloud’s focus on application-centricity that stems from initial initiatives aligned to address infrastructure-level considerations is very much of-the-moment in terms of the current approach to platform engineering (and indeed infrastructure-as-code) that is playing out across the cloud-native enterprise software space.

The company clearly has the breadth and scope to build this level of services and its pedigree in core datacenter services and ability to see how applications are wrangling with the neurons that connect them to the backend should enable many of these tools to find a home in contemporary developer environments. What matters next is how well Google Cloud listens to programmer pain points and finesses its next moves in this space… and perhaps also listen to what the AI agents are getting done best and looking for where they are struggling.

After all, developers and robots both need love too, right?