Google is deep into the application development ecosystem, if for no other reason than that it created many open-source components that have since become ubiquitous.
Google’s DNA is all over modern applications. It would be difficult to find modern web apps that aren’t written for the Chrome browser, mobile apps that don’t work with the Android operating system or AI apps that don’t have hooks into the TensorFlow library.
Over the past year, Wikibon has seen Google make important announcements to deepen its application development portfolio. It continues to build out its core mobile and Web application tooling and has leveraged its deep AI assets for a steady stream of feature enhancements throughout its application portfolio.
In mobility, the top Google story of this past year was a significant “AI-first” refresh of its Android mobile OS and supporting mobile development tooling. Key announcements in this regard included:
- Android P, the next major version of its mobile OS, which uses machine learning to learn the user’s habits automatically and predictively adapt settings for battery life, screen brightness and application suggestions.
- Google Assistant, the next-generation AI-enabled mobile digital assistant embedded in Android P, which can respond to questions with multiple subjects; continue conversations without having to constantly repeat the trigger phrase “Hey Google”; and, via the new Google Duplex feature built and trained in TensorFlow Extended, carry on natural phone conversations, based on its ability to understand complex sentences, fast speech, long remarks and speaker intent.
- Flutter, a new cross-platform mobile app software development framework for building apps that can run on both Android and Apple’s iOS platform without any changes to the underlying code, thereby enabling developers to build “native” mobile apps for both mobile operating systems more quickly than they can use traditional platforms. It later enhanced the SDK with mobile app bundles and in-app payment support.
- Cloud Firestore, a new serverless, NoSQL database for mobile, web and “internet of things” applications that integrates with Google Cloud Platform and Firebase, which is Google’s mobile development platform.
- ML Kit for Firebase, an SDK that supports low-code application development of machine learning apps for Android and iOS mobile platforms. It enables machine learning developers to build apps with just a few lines of code and supporting application programming interfaces for TensorFlow Lite, Google Cloud Vision and Android Neural Networks, providing prebuilt support for text recognition, face detection, barcode scanning, image labeling and landmark recognition.
- Maps, a mobile app that provides AI-personalized real-time recommendations in Android and iOS and a new “visual positioning system” feature that displays an augmented reality view of place names, street names and directions in the user’s smartphone camera view.
In web apps, Google continues to enhance its G Suite productivity suite with look and feel enhancements, an improved document API, new document sharing capabilities, expanded conferencing capabilities, tightened suite security and an AI-powered grammar checker. It also released a new tool that allows developers to speed webpage image downloads to the Chrome browser.
Just as important as these client front ends are Google’s rich back-end capabilities in Kubernetes, Istio, Knative and other cloud-native application platforms. In recent articles, Wikibon covered Google’s activities and roadmaps in the cloud and in embedding AI everywhere, as well as providing the tooling and APIs to make this consumable to developers.
Looking back at theCUBE interviews that took place a year ago on at Next ’18, Google’s priorities for enriching its application development ecosystem are clear:
- AI drives the intelligent conversational interfaces at the heart of modern cloud apps: Dan Aharon, product manager for Google Cloud AI: “[Google] Dialogflow is a platform for building conversational applications and interfaces. So it could be chatbots, it could be voicebots, and it started from the acquisition of API.AI, that we did a year and a half ago, and it’s been gaining a lot of momentum since then. [There are] two big things, one is cloud, the other is machine learning and AI … advanced speech recognition, natural language understanding, speech synthesis. With cloud, there’s now a lot more processing that’s done centrally and there’s more availability of data that could be used to train models and that feeds well into machine learning, and with machine learning we can do stuff that was much harder to do before machine learning existed. What makes Dialogflow special is you could use it to build stuff very, very easily, so I showed last year at Google Cloud Next how you build a bot for an imaginary Google Hardware store.We built the whole thing in 15 minutes and deployed it on a messaging platform and it was done, and it’s so quick and easy anyone can do it now.”
- Serverless will unlock developer productivity: Diane Greene, former chief executive of Google Cloud: “My daughter is a computer science major and she just taught at coding camp this summer, and they started from kindergarten and went up. It was amazing to hear what those kids were doing. I think a lot of applications are almost going to be like assembling Legos. You have all these APIs you can put in, you have all these open-source libraries, you have serverless, so you just plop it in these little containers and everything is taken care of for you.”
- Cloud-native abstractions are becoming progressively more consumable by developers: Jennifer Lin, director of product management at Google Cloud: “When Kubernetes was donated to the open-source community, there were some things that needed to be defined, such that the abstractions could be very clean outside of a Google environment. But that framework, obviously, held up very well and hence the growth with Kubernetes. Istio, I think similarly, models a lot of the way that we’ve done service management with the service mesh within Google. [Embedded in these open-source frameworks is] a lot of operational domain knowledge on best practices and how to essentially enable automation at a much more granular level of applications.”
- App development simplicity won’t be lost as clouds grow more complex: Aparna Sinha, group product manager of Kubernetes, and Chen Goldberg, director of engineering of Google Cloud: “What we’ve seen in the industry is that it’s only become too easy to create microservices, or services overall. But we still want to move fast, so with the industry today, how can you make sure that you have the right security policies? How do you manage those services at scale? And what Istio does really, in one sentence to explain it, it decouples the service development from the service operations. Developers are free, they don’t need to take care of monitoring, audit, logging, network traffic for example, but instead the operation team has really sophisticated tool to manage all of that on behalf of the developers in a consistent way.”
- Serverless developer simplicity will leverage Knative: Oren Teich, Google Cloud director of product management: “Serverless is foundational to what we’re doing at Google. Knative is infrastructure. It’s the building blocks for serverless, and so the key is, there’s lots of lots of companies today that are building serverless products. We’re building them. Red Hat’s building them, IBM, everyone’s building serverless products. And what we want to make sure is that customers have the ability to seamlessly move between them, that they can take advantage of serverless, without being feeling trapped. OpenStack is using Knative moving forward. The whole idea is, this is an underpinning to give you common control plane APIs, as well as execution APIs, as well as a reference implementation that everyone can build off of the shared pieces together. Where serverless meets Kubernetes. And let’s be clear, Kubernetes is the orchestration layer for people who care, and so for people who care, now you have a serverless layer on top. And for people who don’t care, you come to a cloud vendor, and we’re going to hide all of that for you. But you can still have the exact same control APIs. So, if you think about it, you’re writing things, I have 20 functions and I write scripts to manage all of those, you can now have the same APIs to manage them everywhere.”
- DevOps automation will drive development team productivity: Melody Meckfessel, vice president of cloud engineering at Google Cloud: “We’re very excited to announce Cloud Build, which is a fully managed continuous integration and delivery platform. It lets developers build and test their applications in the cloud at any scale, and it’s based on a lot of the lessons learned that we had within Google, iterating over the last two decades with developer and operator tools. Google does some crazy scale internally, and we’re really excited to bring that automation and scale out to our customers. What we’ve found is that there’s a tremendous amount of value of automating away, you said toil, the things that developers want to do. So some of the research, industry research that we’ve done, developers want to write code, they want to do design, they want to work on requirements. They don’t want to take care of the plumbing and the pipeline of how their build, test and release happens. What makes developers happy? One thing is to give them automation so that they can focus on code.”
- API manageability is critical to efficient cloud-native microservices DevOps: Ed Anuff, director of product management at Google Cloud: “APIs are how software talks to software. And what we announced this week at the show with Kubernetes and Istio are new ways for people to build software and deploy it, in new distributed fashions. What we announced was that Apigee is making it now possible for you to have all the tools that we’ve given you for managing your APIs, for, you know, getting your mobile apps to talk to your cloud services and all that, now is also going to apply to these new microservices that you’re building. We’ve been using Apigee for about four years now, and so over the time I think we were have 200 plus internal APIs, so we’ve over that time learned how to operationalize that piece of it. Over the last couple of years we’ve really been focused on the microservice layers. Writing cloud-native applications, essentially. And that layer, and now with the Apigee hook into Istio, we’re going to have a much better way to manage it.”
As Wikibon looks ahead to Next ’19, Google needs to double down its application development tooling, platforms and strategies in several areas where it is at risk of falling behind other cloud powerhouses:
- Focus on low-code development tooling: Google will come up short with developers until it rolls out a competitive low-code development workbench. Wikibon recommends that Google build out App Maker, which is currently included with G Suite’s Business and Enterprise editions, into a more full-featured tool for web, mobile, edge and robotic process automation apps. Google should broaden App Maker’s functional reach so that it can support templatized development of diverse cloud-native applications for a wide range of business-vertical and mass-market requirements. The vendor should also integrate App Maker with Cloud AutoML in order to democratize embedding of cloud-native AI microservices into all apps.
- Bring AI directly into the Web/mobile app development workbench: Google is hiding one of its most differentiating Web development assets under the proverbial bushel basket. This past year, it rolled out js, which is a browser-based machine learning framework for JavaScript developers. TensorFlow.js is an interactive framework for developing client-side AI applications where the data remains entirely in the browser. To get a jump on its chief cloud and AI rivals, Google should incorporate TensorFlow.js directly into its low-code workbench to enable AI model building and training entirely in the browser. It should also converge TensorFlow Lite, TensorFlow for Swift and Android Things software development kits into the tool to support deployment of trained AI models to a wide range of mobile and edge platforms.
- Jumpstart an edge-AI dev workbench for autonomous robotic platforms: Google will fall further off the radar of developers building AI-infused apps for autonomous edge-AI scenarios until it offers a robotics development platform equivalent to those of Amazon Web Services Inc. and Microsoft Corp. To kickstart a robust workbench for these challenges, Wikibon recommends that Google leverage TRFL, a new library of building blocks for developing robotics reinforcement-learning agents in TensorFlow that was recently open-sourced by its DeepMind business unit. Just as important, Google should enhance Cloud AutoML to provide a reinforcement learning AI-toolchain cloud service to match AWS’ SageMaker RL.
To push forward your cloud application development initiatives, you should attend Google Cloud Next 2019, which is taking place April 9-11 in San Francisco. TheCUBE will be doing live interviews with Google executives, developers, partners and customers during the conference.
