With a CAGR of 32%¹, the serverless computing market is rapidly picking up steam, and for good reason. Traditional computing architectures require a painstaking amount of hardware management. Whether it’s being managed onsite or in the cloud, provisioning, scaling and maintaining the hardware on which one’s software is deployed is such a tedious job that scaling organizations often have no choice but to employ entire teams of developers to tackle it.
Serverless computing assumes that burden by selecting, deploying to and managing hardware for organizations. Developers upload location agnostic software services, which are deployed automatically to the appropriate computers when there’s demand for them. No picking out computers, no guessing at scale and no paying to run software when it’s not being used.
The ‘servers’ still exist, but developers don’t have to work directly with them.
Serverless will significantly lower the barrier to entry when it comes to making software available over the internet.
But not all serverless platforms are created equally.
Cloud companies, for example, have been hard at work delivering serverless platforms that deploy software automatically to computers in their respective data centers. But if the promise of serverless is that any service can run on virtually any computer, why use a platform that only provides access to computers in a small handful of data centers?
Cloud-only serverless infrastructure makes it easier to get software into the cloud, but it doesn’t solve some familiar cloud shortcomings:
- Geography: while powerful, cloud data centers are incredibly sparse when it comes to geographical footprint. One’s distance to a given data center can significantly increase latency, a problem for technologies that need near-instant access to processing power like IoT, VR AR and more.
- Complexity: cloud companies have established an unbelievable business by building tools with developers in mind, which has produced highly technical platforms that can be opaque to non-technical business owners. Not only do platform-specific cloud learning curves create vendor lock in, they often require businesses to maintain enormous teams of developers.
- Centralization: datacenters are a single point of failure that lack resiliency. They’re also a single point of a control, providing a handful of cloud vendors a near-monopolizing influence on how and where data is sent and stored over the internet.
At ActiveAether, we see serverless’ additional abstraction between software and the hardware on which it’s running as the perfect opportunity to dramatically increase the quantity and type of hardware included in our computing infrastructures.
ActiveAether’s serverless platform can deploy software to any computer in the world that’s running the ActiveAether Service Agent. That includes computers in the cloud, at the edge and everywhere in between. The result is a more decentralized, resilient and competitive serverless platform that addresses the shortcomings of a cloud-only architecture:
- Diverse Geography: Cloud serverless platforms only include computers in the cloud, while ActiveAether can include computers right at the edge and even onsite. It solves an important latency hurdle, and creates an infrastructure without single points of failure. For enterprises that maintain a private cloud, ActiveAether makes onsite serverless possible.
- Simplicity: ActiveAether’s platform was built to deploy software for businesses, allowing users to upload code and broadly specify for needs like lowest cost, lowest latency or green-only compute, without having to understand how deployment to resources that meet those descriptions actually occurs. For developers that means more time spent building core functionality, for business owners, it means removing some of the mystery (and resource drain) of DevOps.
- Decentralization: Infinitely more computers in the market means a lot more competition in price. It also means that no single hardware owner will maintain a monopolizing influence on the way software across the globe is developed, deployed and managed. Local businesses can buy compute from peers and neighbors instead of solely sending cash outward to cloud providers, circulating profits through local economies.
Serverless in the cloud is a huge step forward for organizations and developers in its own right, but as we progress toward a future where any software service can run on any computer, cloud datacenters will be just one part of a much broader spectrum of computing infrastructure.
‘Serverless anywhere’ is the future, and you can check it out for yourself here!
¹ “$7.72 Billion Function-as-a-Service Market 2017 — Global Forecast to 2021: Increasing shift from DevOps to serverless computing to drive the overall Function-as-a-Service market — Research and Markets", Research and Markets, February 2017)