The use of serverless computing, e.g., AWS Lambda, Microsoft Functions, and Google Cloud Functions, is not widespread in the insurance industry despite the recent, rapid adoption of cloud in other industries. Serverless computing is a promising solution for many short-duration and event-driven tasks, though it has limitations as a strategic platform for insurers.
Organizations that utilize traditional cloud services (e.g., IaaS) spend less time maintaining physical infrastructure. Serverless computing is the next step in this process. Serverless can increase the bandwidth of developers, allowing them to focus on writing code rather than maintaining an organization’s infrastructure.
Cloud providers apply usage-based price models for serverless. They typically charge by the request and the executed code’s elapsed time. Serverless often leverages function as a service (FaaS), where code executes in response to an event. Organizations often find the most benefit when utilizing serverless within a microservice architecture, where discrete functions can be triggered.
Serverless is most suitable for short-term, self-contained tasks. Traditional IaaS provisioning can manage applications that are used consistently. Serverless can also support use cases like real-time data transformation, ingestion of data from IoT sensors, or the creation of scalable chatbots during demand spikes. Employing serverless prevents insurers from provisioning (and paying for) too many servers while remaining prepared for periods of high demand.
Serverless is not a cure-all, however. There several considerations that carriers should evaluate before deploying a serverless model. For instance, serverless can increase the complexity of an environment by making it challenging to test code. This challenge can be exaggerated when serverless codes execute in response to an event, as it is difficult to test all the scenarios. Carriers should also weigh the threat of vendor lock-in. A carrier’s serverless code will remain vendor-specific and will only integrate with other specific cloud provider services.
The low barrier to entry for serverless computing makes it a good fit for midsize carriers that are seeking to increase their speed of deployment and expand their developers’ bandwidth, as well as for any carrier with discrete, intermittent needs for powerful, high-volume computing tasks. But other approaches to cloud are generally more appropriate for most other applications in insurance.
For more, see Novarica’s research on Cloud Computing Platforms in Insurance.
Add new comment