Fastly API Solves Real-Time Cache Needs for Location-Aware, Context-Specific Applications

Mark Boyd
Aug. 01 2014, 11:51AM EDT

Content delivery network Fastly has an API that's designed to solve real-time scalability needs for managing contextual, personal data and content. Several of its newest features are still ahead of demand, as content publishers, including app makers, are not yet advanced enough to take advantage of everything Fastly can offer, CEO Artur Bergman told ProgrammableWeb.

“Fastly is a content delivery network tailored for content where an API is a large part of what we do,” Bergman says.

fastly

“Some of the key features we have are real-time purging, and by that, we mean 150 milliseconds. What that enables is the ability to cache things that weren’t cached before. You can do very advanced things at the edge to change what you are caching,” he says.

Growing Demand: Caching Geofenced Data

Caching API-retrieved data on the client side is still an underutilized strategy to manage contextual information flow in mobile applications. For large volumes of data — for example, customized maps with specific information matched to where a user is located — caching can be an important strategy to reduce API calls and associated costs, and it's an essential tactic when looking to scale API usage in applications.

“We have quite a few clients who do things like using geofencing in their cache to be able to give a cached response in a particular area so it greatly reducing their costs,” says Bergman.

In fact, demand for caching this type of data is one of the fastest-growing uses for the Fastly API, according to Bergman. “You can cache API calls at the edge, and when the data is relevant to the user’s search, then we can instantly purge. For example, you can tag responses and then purge those responses when one of them changes.”

Overall, Fastly calls this combined service of providing cached data and instant purging “API acceleration.” It is a suite of services that is being taken up by many location-aware apps, including Foursquare, Shazam and Yammer, as well as leading content platforms like GitHub. As part of the API-acceleration service, app and content providers can also do analysis on who is using the API in real time.

Meaningful Hyperlocal Services

Over the past six months or so, more content and application providers are building out approaches that can make use of contextual data in meaningful ways. For too long, many would describe the endgame of location-specific, personalized context data as meaning that when you pass a shop you like, you would be sent a discount coupon for a product you might be interested in.

However, a growing understanding of the hyperlocal context has been creating more personally meaningful ways to use location-aware data in applications and web services. For example, AT&T is partnering with Sabre to test using location-specific data in airports so that passengers can be informed of gate changes and can be shepherded from wherever in a terminal they are whiling away their time onto the plane in a more efficient fashion. In part, this aims to reduce the significant costs and environmental impacts that occur when planes depart late.

Hyperlocal health services like Toothpick in the U.K. are using context-awareness data to help website visitors identify local dentists, and are even providing real-time public transport data so that patients are confident in getting to appointments, seeing both a rise in dental bookings and a lower cancellation and nonattendance rate among patients.

There is still a huge amount of opportunity in using energy, logistics and package delivery, events management, transport and traffic, and, of course, weather data in creating a new generation of hyperlocal content now that more businesses are thinking outside the box of contextual advertising. This is at the core of why Fastly is seeing a rise in demand for using its geofencing cached data and instant purge capabilities.

Supply Before Demand: Real-Time Data Updates

But while content and application providers are catching up on the potential of using hyperlocal data in this way, the promises of providing real-time data are still beyond many businesses. The technology is there (PubNub and Twilio have multiple use cases demonstrating the potential of real-time data flow). But for many content providers not at the leading edge, how to make use of real-time technologies to improve user experiences is still beyond their thinking.

This appears to be the case with some of Fastly’s real-time cache capabilities. Bergman explains:

There is another approach as well, which we have seen some of the more real-time companies use. Even if you only cache for five seconds, it means that your origin servers only need one request every five seconds. If the data changes on a rapid basis, you can use short-lived projects so that end users always get a very fast response.

It’s a technique that people are starting to use, as real-time APIs are realizing they can cache, there is definitely growth there. We are trying to educate and develop features that make it easier for customers to do this. Where we see it primarily is with news organizations where content changes every 15 minutes.

We also support a feature called request collapsing. So no matter how many people are requesting, we only make one API request, we make just one API call and then just release those calls to all their users who are waiting for that call.

People are catching up with that, but there definitely are customers that do that.

While these real-time features are just beyond current demand, Bergman sees media, e-commerce, “social media long-tail, user-generated content” and mobile video as the verticals most prepared to make use of these newer capabilities.

Caching = Cost Reduction

At the end of the day, it may not be these real-time capabilities that are the key factor in why customers want to take up using Fastly initially, but the cost reductions that such technologies are enabling. By using the Fastly API and API-acceleration services, content and application providers are building more scalable API architecture that is financially viable. “We have examples of customers who have been able to shut down large amounts of infrastructure once they have started caching with Fastly,” confirms Bergman.

The Fastly API allows content publishers and application developers to access all of Fastly’s API-acceleration services via a REST interface. Fastly is also committed to open source projects, providing its services for free to OpenStreetMap, Video.js, W3C and the Debian Project, among others.

Mark Boyd is a ProgrammableWeb writer covering breaking news, API business strategies and models, open data, and smart cities. I can be contacted via email, on Twitter, or on Google+.

Comments