API Predictions for 2014
Join the DZone community and get the full member experience.Join For Free
2013 has been an impressive year for API technology both in terms adoption, technology and investment. The number of APIs listed in the ProgrammableWeb directory crossed 10,000 for the first time, there were investments and acquisitions, as well as multiple new conferences on APIs sprang up during the year including our own API Strategy and Practice Co-organised with API Evangelist. There were also new initiatives like API Commons that try to address some of the longer-term industry challenges.
At 3scale the number of live customer APIs we power has already more than doubled over the year and so has API traffic.
So after a great 2013, we asked ourselves what next year might have in store – here are our API Predictions 2014! (These predictions were also kindly published at Business to Community.)
(Predictably) More API Growth: an unsurprising first prediction given the many new API rollouts in 2014, but all indications are that the number of new APIs will continue to accelerate. The number of public APIs listed in well know directories has just reached 10,000 and there are many more private and semi-private APIs outnumbers the public APIs. These private APIs include a wide assortment of growing categories including mobile backends, SAAS integration APIs, heavyweight B2B integrations being replaced by APIs or even APIs for in home electronic hardware devices such as the Philips Hue lightbulb.
In 2014 we expect these trends to continue with the number of public and private APIs climbing to between 100,000 and 200,000 – the lower bound of which we may already be close to if estimates that private APIs outnumber public by as much as 9:1 are correct. We expect the growth to take two forms – continued new API rollouts (as an indicator 3scales sees 100’s of new signups per months and 10’s of new API launches per month) and also a gradual increase in visibility for many private APIs as companies open up to progressively wider audiences.
Broadening Mainstream Adoption: the second and perhaps more important prediction with regards to growth is a broadening of the types of company launching APIs – these now include Sportswear Companies, Retailers, High Tech Equipment Manufacturers, Construction Companies, City Governments, Charities and almost every economic sector we can think of. We expect this trend to continue and deepen. In a study for our recent ebook we identified five key use cases for APIs from mobile enablement to driving business transactions, and these five types of value seems to be transversal – applying to almost every business sector.
API Copyright will be a key debate: with the re-opening of the Oracle v’s Google Java copyright case, the specter of copyright enforcement over API specifications has returned and the case could have wide ranging implications if overturned – implying that specific interfaces patterns could be copyright protected and re-use prevented. While in some cases protection may be warranted, general copyright for APIs would likely have a deep chilling effect on innovation – effectively locking up common and critical interface patterns and barring usage by others. Activities like API Commons may help to build up a body of re-usable patterns – but efforts such as this are still in their infancy – the copyright debate will be an important feature of 2014.
Service description technologies will come sharply into focus: one of the key emerging challenges for REST APIs is that API Providers rarely, if ever, publish locatable machine parsable meta-data about their APIs that can be found and processed. While there are some emerging formats such as WADL and Swagger available, as well as new ones such as Blueprint and RAML emerging, they are at best partial solutions and not widely adopted. This lack of automated descriptions hampers everything from the generation of stub libraries to API discovery – for example, an API search engine is unlikely to emerge until publishing such electronic descriptions becomes the norm.
As a result we expect more industry and best practice debate around which formats to adopt and conventions on how to publish / discover them. Without this, it is likely to become impossible to maintain public directories of APIs such as ProgrammableWeb.
More API Aggregators will emerge: as the number of APIs grows, middle layer initiatives such as Segment.IO, Zypr (*) and others are emerging to combine APIs in a specific segment into a single addressable uniform API. An aggregator of this type creates a single point of integration for multiple backend APIs. In some cases, aggregators may be in a precarious position w.r.t the APIs they integrate, but in others they will be welcomed with open arms – reducing complexity for developers and cutting direct support costs.
As the number of APIs increases the number of areas where aggregators will emerge will only grow. Usage of APIs is rising even faster than API creation which leads to the next prediction…
API Developer Tools will be a big growth area: many of the vendor tools for APIs to date from companies such as ourselves, Mashery, Apigee and others have focused on enabling API Owners to provide those APIs to other. This is necessary technology and now seeing wide adoption. However, up until now there has been little tooling for the consumption side of APIs – namely for developers to debug, monitor and track API usage for their applications. 2013 has already seen a number of new companies emerge to meet this need including Runscope and API Science, there is likely more to come here including a new 3scale offering.
Innovation for developers using APIs is critical since there is still a large barrier and plenty of frustration for developers using even well documented APIs to achieve their goals.
As a result, we expect that SPA architectures will start to become the norm for new business/functionality style Apps in 2014 and slowly make an appearance in E-commerce/Retail style applications.
Hypermedia APIs will see more adoption: Hypermedia style APIs provide the means to increase the flexibility of an API by encoding permitted actions within the responses returned by API calls (see Richardson, Amundsen and Ruby’s useful book) – allowing the potential for APIs to dynamically change the API calls they make to adapt when things change. There are still significant tradeoffs in using Hypermedia style APIs versus static fixed URL APIs, however there are now a few emerging larger deployments including the new Public Media Platform APIs and Amazon’s recently released Appstream API.
Adoption will take time, but we expect 2014 to see more high profile use cases and more tooling to help accelerate usage.
Civic Data APIs will have a breakout year: in trying to predict which categories of APIs will grow the fastest, the category we kept returning to is civic APIs. Global cities such as New York, Chicago, San Francisco, Amsterdam, Helsinki have already released increasing numbers of data sets and now APIs over time – providing infrastructure data to citizens who in turn can produce their own Apps, Visualizations and Mashups. We see this trend accelerating even more rapidly as it becomes prohibitively expensive for governments to provide all the Web and mobile applications demanded by their citizens – opening data provides a channel for innovation and opportunities for citizens to help themselves. Another reason for acceleration is the increasing effort through initiatives such as open311, CitySDK and CityProtocol towards standardizing some of these interfaces across cities which should make adoption easier and at the same time broaden the market for Apps which use such data.
So there we have it – an exciting year in prospect and with surprises also no doubt! We’re looking forward to seeing which predictions come through and what unexpected new happenings. What are your 2014 API Predictions?
[Update: in the original article we incorrectly listed uShip as an aggregator which, while they have an awesome API you should check out, they are not - apologies. Predicting they have an excellent year though!]
Published at DZone with permission of Steven Willmott, DZone MVB. See the original article here.
Opinions expressed by DZone contributors are their own.