The idea of lightweight middleware platforms has been a powerful idea in enterprise software for decades. A few years ago, Forrester’s analyst Michael Facemire published a brilliant piece in which he proclaimed that mobile backend as a service(mBaas)could become the next enterprise lightweight middleware. Even though the prediction didn’t turn out correct, Facemire’s ideas expressed in the article remain incredibly valid. A few year after the mBaaS movement, a new technology is seriously positioned to power lightweight middleware capabilities in the enterprise.
Serverless computing has become a relevant capability of modern platform as a service(PaaS) stacks. Brought to mainstream adoption by technologies such as AWS Lambda and followed by the releases of competitive stacks such as Google Functions, Bluemix OpenWhisk or Azure Functions, serverles patterns are one of the most popular architecture styles in modern software applications.
From an enterprise perspective, many of the most common scenarios for serverless architectures can be expressed as traditional middleware patterns that execute specific integration logic in response to events. From that standpoint, serverles stacks are already playing some middleware role in the enterprise. Differently from previous lightweight middleware pretenders, there are many characteristics of serverless technologies that can help this trend to succeed as a lightweight middleware stack in the enterprise.
Technology Perspective: 5 Reasons Why Serverless Can Become the Next Lightweight Enterprise Middleware
Technologically, serverless architectures can, many times, result on a more efficient model to implement enterprise middleware patterns by leveraging some of the following characteristics that contrast with the tradititional heavy enterprise middleware technologies.
— Developer Friendliness: Serverless stacks are vert veryearn by mainstream developers and rarely require integration specialists.
— Event-Driven: Serverless architectures are event-driven in nature which makes it easily adaptable to different enterprise middleware scenarios.
— Extensible: Extensibility (or the lack thereof)is one of the main challenges of traditional middleware solutions. Serverless stacks are highly extensible by definition [they are based on defining new atomic functions that execute via events] which enables the implementation of complex middleware scenascenarios.Multi-Language: Serverless stacks typically support atomic functions written on several different languages. From an enterprise perspective, this capability can streamline the developer adoption of serverless technologies.
— Hybrid: Serverless architectures can be deployed on-premise and cloud infrastructures which facilitates their adaptability to different enterprise middleware scenarios.
Market Perspective: 5 Reasons Why Serverless Architectures can Become the Next Lightweight Enterprise Middleware
Complementing the previous section, there are several market elements that favor the adoption of serverless stacks in enterprise middleware scenarios:
— Driven by Cloud Incumbents: Differently from some of previous attempts to capture the enterprise lightweight middleware space, serverless stacks are mostly driven by cloud incumincumbents as Google, Amazon, Microsoft or IBM which adds strong credibility to the space.
— Pricing: The pricing of serverless platform is a race to the bottom and this is a factor that can really influence its adoption in the enterprise.
— IOT: The evolution of IOT middleware scenarios is catalyzing the adoption of serverless stacks in the enterprise.
— Lack of Innovation in Traditional Middleware Technologies: Recent years have seen very small, incremental innovation in traditional middleware plaplatformsich have kept enterprises looking for innovation. Serverless technologies are certainly a fresh and modern approach in a space that hasn’t seen much movement in the last few years.
— Talent Availability & Cost of Ownership: Serverless architectures are easily understood by mainstream developers. This is an attractive factor in the enterprise spspeciallyompared with tethexpensive cost of traditional integration engineering talent. Also the cost of ownership and infrastructure required to run serverless solutions is almost neglectable.