‘Reactive’ was a term that pulsed through the sector in 2015, and as we enter 2016, the movement for modern responsive modern systems is edging from the bleeding edge into the mainstream. In a survey by Reactive specialists Typesafe (full results available here), out of a pool of 3060 tech professionals from an array of industries, 43% stated that they believed the “Going Reactive” trend was either gaining momentum or something they needed to look into. A further 40% perceived it as “something we needed yesterday.” Only 5% expressed dismay as the resurgence of the term. On the whole, 80% expect enterprise adoption by 2018.
Before we go any further, a quick clarification on what we’re talking about when we talk about “Reactive.” Whilst the term has been framed in a variety of ways across enterprise land, for a system to be deemed truly Reactive, there are four tenets it needs to incorporate: Message-driven, Resilient, Elastic, Responsive – a quadrant of priorities that will be familiar to many in 2016, even if your company is not explicitly “doing” Reactive (yet).
Doing it at scale
As application developers become increasingly savvy about the need for responsiveness in their systems, the notion of achieving optimum scalability has become a primary driver for many in the enterprise. Overall, 31% of those surveyed placed scalability as their primary driver for taking the plunge into Reactive, with a further 22% citing resilience, and 17% stating that for them, Reactive was about modernisation.
These drivers also coincide with another key trend in 2015 – the indomitable microservices movement. Typesafe note a definite correlation between the adoption of loosely coupled architectures and how far enterprises have progressed in their microservices “journey.” 33% of those questioned are using microservices-based systems, and out of these, 50% are building and deploying Reactive systems.
The age of real-time
With Mobile, Web and (IoT) systems thirsty for real-time data flowing on tap in their Reactive systems, as we’ve seen in other vendor reports over the past few months, there’s been a shift from traditional batch processing systems to ultra zippy pure-stream based architectures. Software like Apache Spark is responding to this need, and helping to drive the future of the web. In this survey, 22% of respondents reported that they were working with Spark. Out of these, 28% were already building and deploying Reactive systems, and a further 21% were engaged in researching and prototyping.
Respondents who embrace microservices and fast-data also hold strong opinions around the tools and technologies utilised to keep their systems ticking over. There’s a strong skew among these adopters for Akka, Apache Cassandra, Apache Kafka, Apache Spark, Play Framework, Amazon EC2, Apache Mesos, and of course Docker.
Whilst the year is still young, it’s safe to say that you’ll be hearing a whole lot more about Reactive in the months to come. If you’re still in the research phase, in this interview with Typesafe’s Jonas Boner, filmed at ScalaDays Amsterdam 2015, we expand on what constitutes a Reactive system – and where ideas diverge in the market.