By Jon Nordmark
When Amazon created its virtual assistant Alexa, the retail giant didn’t do all of it in-house. In part, it identified three relatively unknown startups. One had no venture capital funding, a second one had raised a few million dollars from unknown investors, and a third had raised venture. Each had developed voice technologies which could be repurposed for Amazon’s pending personal digital assistant and other voice initiatives on the drawing boards at Amazon. The result: a product that’s driving a revolution in voice technology.
Like Amazon’s Echo/Alexa, great products often involve multiple technologies — multiple point solutions — that are woven or fused together. The Amazon Go stores are famous for their “sensor fusion” technologies. That kind of modular approach to innovative software development is made easier now, thanks to a new breed of “microservices” popping up. Microservices allow large companies to drag-n-drop various software APIs together to get one fast, nimble solution.
By 2022, 90% of all new apps will feature microservices architectures that improve the ability to design, update, debug and leverage third-party code. And about 35% of all those production apps will be cloud-native, according to research firm IDC. This growth will lead to new, internally-generated apps called “hyper agile apps.” They’re modular, distributed, continuously updated and they leverage cloud-native technologies such as containers and serverless computing.
Amazon already operates that way— far beyond its Alexa example. As Dave Gray, points out in his book, Connected Company, you can go to any page on Amazon.com, and in the background it may have 300 custom microservices, bundled together to optimize the shopping experience. Each of those 300 microservices, described as independent but loosely coupled pods by Gray , are all replaceable by new, better-performing services.
These technical pods are owned by small self-contained teams, which also operate like like pods — or microservices. McKinsey writes that organizations of the future will not operate as hierarchical machines; rather, they will need to be agile, like a living organism which adapts quickly to changing competitive landscapes.
Microservices are not necessarily a completely new approach to software engineering. They cobble together successful and proven concepts, such as agile software development, continuous delivery (CD), service-oriented architectures and API-first design. But by moving to modular from monolithic, you can dramatically speed up deployment cycles for new software, foster innovation and ownership, and improve maintainability and scalability of software applications.
You get systems that are highly decentralized in terms of data management, development, deployment and operations. You get flexibility: pods can be changed, upgraded and replaced without affecting other components. Each component, meanwhile, is designed for a set of capabilities and certain levels of complexity. As a whole, microservice architectures allow for enormous customization and the freedom to choose the best tools for each problem.
Most retailers, most media companies, can’t invest $30 billion in R&D as Amazon does today. Correct? In fact, adding a single software developer can be a big decision for some $100 million companies. But, in this digital-age, lack of capital (or lack of human resources) shouldn’t be an obstacle to innovating faster and more effectively.
My company, Iterate.ai has bet big on microservices to help companies of all sizes innovate faster and at an affordable cost. We built an entire technology workflow around this kind of modular architecture.
Our microservice tool — Interplay — allows companies to prototype new digital technologies 6 to 15 times faster by using an architecture composed of pre-wrapped APIs from independent startups, APIs from enterprise applications and a lot of independently written AI modules. Any of those independent nodes can be drag-and-dropped onto a digital workbench so that they can immediately work together — as a working prototype. It’s like an enterprise version of Scratch which kids use to learn software development. The result: chatbots that include natural langage processed (NLP) language vocabularies, customized to individual brands, that produce inbound and outbound automated text capabilities. Video aggregation tools. NLP-driven FAQs for websites. Automated sales assistants for specialty products. Voice bots driven by custom NLP languages. To build these, Interplay uses APIs that are pre-wrapped and connected together — placed in sequences like game pieces on a board.
Two of Iterate.ai’s large customers even discovered that Interplay’s prototypes can be ported into production, removing the need to rely heavily on IT resources to integrate inventive new technologies.
Companies taking advantage of Iterate’s Interplay are moving quickly toward IDC’s 9th prediction — that new tools and platforms, plus agile methods, will facilitate lots of code reuse and allow a new breed of less-technical, creative developers to emerge — and that this will create an explosion of digital innovation
And their proliferation underscores how innovation is changing. No longer does one company create linear tech stacks hosted on-premise. Often, the best solutions involve the cloud and a combination of enterprise and startup technologies. Often, great ideas hide out inside small, unknown startups—and not just one firm, but often the secret is found by combining multiple point solutions from startups.
Just as companies are have evolved to become agile organizations that embrace quick changes and action-based leadership, we predict that technology architecture will follow suit — they’ll become more modular, and startup technologies will become more drag-n-drop.
Stay ahead of trends with insights from iterate.ai experts and advisors