Serverless

Serverless is a paradigm in digital service delivery in which the operator of a service or piece of software has no visibility or interaction with the servers and underlying architecture that implements the software’s functionality. While the software still runs on compute infrastructure and servers, the underlying hardware and software configuration is full managed, with patching, fault tolerance, and scalability hidden away and seamless to the user.

As such, the billing mechanism is pay per use of the software service, not cents per hour of a server that runs the service. Don’t use it, then nothing to pay. Use a Serverless service and the cost will reflect that duration of use. Contrast that with a virtual or physical server and you pay for it in cost units over some period of time, or up front for a physical server, whether you use it or not.

With AWS, the concept of Serverless covers a range of types of services: storage, compute, application integration, analytics and machine learning services.

Scaling with Serverless architectures is seamless; the granularity of the compute function, instead of the compute server, is controlled by managing the concurrency of the service. We set design an upper limit on concurrency, and the provider operates up to that level.

Gone is the need to manage servers and operating systems; patching, rebooting, downtime; this becomes the cloud provider responsibility, at no additional cost. All we concentrate on, is implementing business logic in a programming language version and updating the language major-version over time (eg, Python 2 to 3, Java 8 to 11).

Modis has a history in the Cloud of skating to where the puck is heading (as AWS says) and being ready to deliver technical solutions to customers that are well-architected, secure, and scalable. When AWS services launch (and sometimes before), our technical teams start understanding and interpreting the value they may have to our customers. For example, with the introduction of AWS Lambda, the Modis team knew it would be significant for a number of customer workloads. When the Sydney launch of Lambda occurred, Modis had already completed our Proof of Concept in other regions, with deployment templates and patterns ready to go.

In early 2016, when AWS Lambda launched in Australia, Modis stood up a production service for one of our customers, to be used by staff at a federal government agency. With a monthly cost measured in single-digit dollars, that service has been secure and cost effective for the agency.

These days, we run Serverless analytics pipelines, web services, integration components, and much more using Serverless as the underlying capability.