Daisy-chaining APIs makes serverless sense

Daisy-chaining APIs makes serverless sense
Credit: Thinkstock

Serverless infrastructure is set to take off, but to take advantage of this, businesses need to have their services and data accessible via APIs

Enterprise and startups are moving to cloud-based infrastructure to create an API-enabled value chain for new products and workflows. With this has emerged the idea of serverless infrastructure: running functions and applications completely in cloud-based servers where the hosting provider handles all of the sysadmin requirements.

This is a new frontier in how business is leveraging cloud, and it is set to explode, especially amongst system integrators and consultants, as more data from sensors and machines are incorporated and as traditional businesses move even more of their IT infrastructure to the cloud.

But to take advantage of this opportunity, businesses need to have their services and data accessible via application programming interfaces (API), which for most businesses is still just an emerging trend.

A tweet last week by Alex MacCaw, the CEO of contact intelligence startup Clearbit, exemplifies a current trend in application and enterprise workflow development:

In MacCaw’s case, he is suggesting that users could build a whole business based on Clearbit’s new Reveal API. The Reveal API is able to identify the specific company of a website visitor based off their IP address, meaning a webpage could load from the get-go with customized content for each web visitor. It is a clever tool that can create real value from a rich data source.

Creating commercial products from APIs

But the subtext of the tweet is even more interesting. MacCaw is saying that a savvy startup could create a whole business just by stitching together APIs. Daisy-chaining APIs together is often done to automate workflows and sync data across various Software-as-a-Service (SaaS) products. But more recently, startups are creating commercial products that are basically a series of APIs where code functions are then applied to transform the data as it is funneled along the API value chain.

Bustle is doing that with its new media tech company aimed at millennial women and parents. A Cloud Guru is doing that with its online training classroom. BuildFax is doing that with commercial property products. Lithium is doing it with its new Reach product, which uses its social management software Klout to create a product entirely made of APIs end to end.

That’s how growth startups are doing it. Enterprise are also beginning to act just as quickly, albeit in a slightly different channel. Companies are making their data accessible via an API so that their partners—systems integrators and consultants—are able to build custom dashboards and mobile applications right in front of their customers.

Workflows and prototyping

Enterprise innovation labs are using the approach to build and test prototypes and to speed up their agility to enter and compete in new markets. They are also putting a user-friendly interface on front of their internal workflows, turning them into enterprise applications so that their workforce are more easily able to leverage value previously locked away in a company’s data storage and in siloed cost centers.

Businesses of all sizes are finding benefits in meeting customer needs more quickly, and they are reducing inefficiencies by automating the internal data movement from one system to another.

Some are also looking to partners to create new products. Melissa Jurkoic, Product Strategist, Services Platform at Amadeus Hospitality, has been leading the introduction of APIs at the hospitality enterprise, which employs nearly 1,000 people. “So the dream is that someone will leverage three APIs and create a new product we haven’t dreamt of,” says Jurkoic, who will be sharing her experiences on driving an enterprise-wide API strategy at next month’s API Strategy and Practice in Boston.

Cost benefits of serverless

There are two other financial benefits that are a huge driver for businesses taking up this cloud and API opportunity. Products can be managed completely in a serverless environment. Of course, there are servers to store the data, carry out the compute functions and display the results in some form of application. But the business is not managing those servers themselves.

First, costs are incurred only when the end user seeks a value from the data. When a system integrator creates a dashboard to show a client how a mix of their data and their company partner’s data intelligence come together, it is only when the cloud-based server is pulling together the data, making calculations and displaying the results that any compute engine power is used, often charged in cents on the dollar.

Second, there are no costs incurred in managing server operations. The cloud server takes care of autoscaling and managing traffic loads if an application increases in usage. If the website a startup builds using Clearbit’s Reveal API suddenly goes viral, more instances are created to meet demand at a global scale. Media juggernaut Bustle is already managing its tech company in this way: It has zero DevOps on call to respond if any of its content suddenly goes viral. The company's serverless environment scales as needed without any sysadmin involvement. They are alerted when traffic peaks, but those alerts are routed to editors and writers to engage with the audience around the story, not to backend engineers struggling to manage infrastructure demands.

The dominant cloud platforms all now offer a serverless product to help businesses take up this advantage: Amazon, Google, Microsoft and IBM each has a solution. But often that requires businesses to host their data on the platform as well or to build their APIs within the platform they have chosen. Businesses wanting to get started with their own data sets need to use products that can create, host and manage APIs in the cloud.

API serverless products

My research into API serverless products (sponsored by Restlet) identified three trends that are driving interest in this type of API serverless architecture:

  1. Businesses are increasingly moving to cloud and hybrid environments to manage their data and infrastructure.

  2. They are doing this at the same time as realizing they need to break down their legacy code base into composable components—microservices—that can be connected by APIs.

  3. As they reorient, they are seeing a need to restructure themselves operationally to bring together technical and business sides of the organization in new team environments.

API serverless architecture map

Serverless API architecture products are becoming more sophisticated in order to help businesses move towards this paradigm. CA’s Live API Creator recently added new features for visualization tooling and links to API gateways. Restlet’s API Spark lets users create APIs from any cloud-based data source, even converting data from a Google Sheet into an API. And new products continue to enter the market, such as Webshell’s recent Materia product release, which lets users create a serverless API and then build a whole application on top of it.

The next cloud frontier

This has always been the promise of APIs and cloud: to be able to build new value chains and create products without a bare metal server dragging down a business’ agility and vision. An explosion of bots, IoT big data analytics, programmable business models, startups spun up around daisy-chaining products, and a new professional class of non-technical systems integrators: those are the next frontiers that serverless makes possible. But those are all ideas to explore in a future post. For now, the message is clear: serverless is here.

This article is published as part of the IDG Contributor Network. Want to Join?

Must read: Hidden Cause of Slow Internet and how to fix it
View Comments
Join the discussion
Be the first to comment on this article. Our Commenting Policies