Serverless—Vastly Improving the Economics of Building and Supporting Connected Devices
What does your cloud platform have to do with brand loyalty? I'll get to that... Let's start with some background on why serverless computing vastly improves the economics of building and supporting connected devices...
Why the Computing Paradigm of IoT Connected Devices is Different
The computing paradigm for the Internet of Things is unique within modern cloud computing architecture in that it is primarily stateless and event-driven, versus state-driven and multi-threaded. These unique qualities are important as they impact the chosen design patterns and architecture and can have a very consequential impact on the overall cost of the IoT solution.
Challenging the standard cloud-computing model for IoT means rethinking how cloud service providers compute charges for their services in an event-driven world as well as how your devices and applications consume those services.
IoT connected devices are mainly event driven, frequently only requiring server access for milliseconds at a time.
What is Serverless Computing?
Serverless computing is a pay-to-use, cloud services execution model in which a third-party cloud provider (such as AWS, GCP or MS Azure) dynamically manages the allocation of machine resources used by customers on an as-needed basis.
Serverless computing pricing models are a shift from early cloud-computing and virtualization service models in that pricing is based on the amount of resources consumed by an application rather than on a server or resource subscription model.
Where traditional cloud infrastructure offerings are referred to as ‘software-as-a-service’ or ‘infrastructure-as-a-service’, serverless computing offers the ‘Compute-as-a-service’ model.
Serverless Saves Money, and Has a Direct Impact on the Viability of Your Business Case for IoT Connected Devices
Serverless is not a new concept, but with fast-paced, cloud-connected, event-driven things appearing on the market, the serverless model offers a potentially vastly more cost-effective approach to delivering connected device services in a sustainable way.
As an example; let’s say you have a device that has processing it needs to do, that takes place over 100 milliseconds. The cloud service will be responsible for that processing. In this example, the cloud server powers up for that length of time, runs the processing, and then powers down.
You only get charged for the 100 milliseconds of use; the costly day-to-day maintenance of servers disappears. It's a smart idea.
While serverless may benefit any industry with high peak processing requirements, it is particularly suitable for the event-driven nature of IoT.
Focus on the Product, Not the Infrastructure
In a serverless environment, businesses no longer have to worry about maintaining boxes, calculating how to optimize them, or whether you need to invest in more/less server technology. The serverless provider covers it all.
Serverless Gives You a Competitive Advantage
Device developers are operating in the most globally-competitive environment in history, against well-established competitors and scrappy upstart innovators alike. The ability of move your focus from infrastructure to the end product isn’t a nice-to-have benefit in today’s connected device market, it’s a must have.
Dedicating time, resources, and dollars to planning, building and maintaining the infrastructure puts you well behind, with long-term cost burdens that competitors leveraging serverless-based solutions don’t have.
Serverless evens the playing field from the start of your connected device endeavor, which means you can focus all resources on what you really care about; the product.
What About Cost? Are You Really Saving That Much on Infrastructure?
Over the lifetime of a connected product the answer is a resounding, “Yes.” The traditional model for cloud computing offers customers cloud-based and even virtualized server infrastructure that they maintained and ran 24/7. The cost was typically flat but always constant; even if the applications were quiet or even not operational, customers still pay for this slow/downtime.
In addition, the real as well as the opportunity cost of spending resources and brainpower on server maintenance and dev-ops instead of on product development/management are direct and indirect costs that simply don’t exist in the serverless model. Resources are dedicated wholly to designing delightful product experiences and solving real customer problems.
Is There a Downside to Serverless?
Serverless can have drawbacks. A potential downside to serverless is losing visibility and control over the hardware. With the right service level agreement in place, this becomes a non-issue. At Yonomi we believe in focusing on your core business — making great products. The ability to know exactly how the cloud service is managing each event should not be critical or even a concern. Yonomi’s ThinCloud addresses this need, alleviates the concerns and is used by Schlage, as one example, for all the reasons listed.
Great—I Understand Serverless...But What Does Any Of This Have To Do With Brand Loyalty?
No matter the industry, product or service, brands that succeed focus on customer experience with their products and services. Downstream challenges that impact that experience are crucial to brand loyalty and customer satisfaction. Serverless as a compute services paradigm is a critical backend component for any connected device developer to leverage as it refines delivering delightful customer experiences.
Brands that maintain customer loyalty are brands that deliver those experiences consistently and at a sustainable cost over time. For brands that win, serverless may prove to be the unsung hero in the economics of connected devices market.
So Why Not Invest in Server Technology That’s Fit-For-Purpose?
That’s exactly what we’ve done at Yonomi. We’ve invested in serverless technology because it fits the bill for connected device support.
Interested in learning more? Download our Guide to Building a Connected Device.