Mulesoft’s Paul Crerand looks at the perfect marriage of API-led connectivity with DevOps.
In our technology-driven world, organisations cannot afford to stand still. They need to deliver software at high velocity without sacrificing stability, which is why DevOps has become such a popular approach.
As the name suggests, DevOps is an attempt to combine software development and IT operations in order to effectively get the best of both worlds: the rapid iteration favoured by the former and the stability, reliability and security favoured by the latter. Ultimately, DevOps helps to ensure that IT can more effectively support the needs of the business.
But it isn’t just DevOps practices that improve the bottom line for companies. Alongside the move to DevOps, companies are also recognizing the need for API strategies to accelerate innovation.
According to MuleSoft’s 2017 Connectivity Benchmark Report, 94 per cent of IT leaders say they can deliver new products and services to market faster with APIs. Together, DevOps and APIs deliver greater business value than what they can deliver individually.
When Dev met Ops
The DevOps movement has exploded over the past few years. Gartner pegged 2016 as the year when it finally broke through into the mainstream, employed by 25 per cent of Global 2000 organisations. Furthermore, of the 50 enterprises ECS Digital surveyed earlier this year, only 11 per cent said they aren’t planning to adopt DevOps.
Modern, “fail fast”, cloud-centric organisations powered by automated and agile development processes like Amazon are leading the DevOps charge. For those more reliant on legacy technology and operating in highly regulated industries, such as financial services, adoption of this relatively new approach has been slower.
After all, DevOps requires a significant cultural change. However, many are starting the journey by adopting new tools to strip out inefficiencies from the development lifecycle. It’s all about accelerating time-to-market to meet customer demand, whilst reducing failure rates and inefficiency.
An unexpected bump
There are challenges associated with this new approach to software development and delivery. For example, if organisations become too focussed on the production model, it can become all-too-easy to keep on pushing out software assets without thinking about whether they offer real value to the business.
Or worse, those assets live in a silo and aren’t discoverable and reusable by the rest of the business. Just because it is much easier to produce in this environment, it doesn’t mean the code production treadmill should keep on turning unfettered.
Organisations therefore need to balance the highly efficient DevOps production model with an efficient consumption and reuse model via APIs. Consider a DevOps project designed to expose customer data from a legacy platform onto a mobile app.
With a well designed API, that legacy data can be discovered inside the organisation by other teams so that they too can use the information for other projects. The approach is faster, more efficient and cost-effective. Additionally, the more discoverable the assets, the faster they become ready for reuse, potentially to audiences that might otherwise not have known about them, thus driving greater agility and innovation.
DevOps and APIs – the perfect partners
With this API-led connectivity approach, every asset becomes a managed API, discoverable through self-service with no loss of control. Organisations as diverse as Spotify and Siemens are already embracing this approach to become more agile, efficient and innovative.
Like Amazon, Spotify has been at the vanguard of the DevOps movement, benefitting from a slick, seamless pipeline of software production. However, it also found that multiple departments were producing duplicate assets and that these assets were difficult for those outside their specific silo to take advantage of.
Using APIs to build an application network, Spotify has been able to expose the capabilities being moved through the pipeline to third-party partners and internally to speed up time-to-market and create new business opportunities.
Siemens, meanwhile, has been able to expose assets managed by its data management team via APIs, eliminating bottlenecks and ensuring partners and internal teams can self-serve to drive greater business agility.
It did so in part thanks to a cross-functional team spanning central IT, line of business and mobile teams that ensures all assets created through the DevOps pipeline are broadly consumed and fully utilised. The results speak for themselves: Siemens has halved the time taken to get a first release of a new capability out to other teams in the business working on related projects.
The marriage of API-led connectivity with DevOps will not necessarily be easy, requiring as it does a change in mindset to one where assets are produced with the intent that they will be consumed by others in the business. As such, central IT needs to change culturally to become an enabler of reusable, self-service consumption. However, as the likes of Spotify and Siemens can attest, the rewards speak for themselves.