Mon, 22 May 2017
We discuss Austen’s early bet on serverless computing from the first time he saw AWS Lambda. Serverless, even in the early days, has many benefits. It is microservice-based, event-driven, requires no administration, and has a compelling “pay-per-execution” pricing model.
Serverless was launched as an application framework. The problem with serverless computing today is that if you want to build a sophisticated system on this type of service, you’re dealing with lots of independent units of deployment. One application is a combination of many Lambda functions. Dealing with this all-together, not to mention the event-driven computing, can be chaotic. Serverless offers a simple file that can define a serverless application. The framework provisions all the infrastructure for you and the app is up in seconds.
We are still in the early phases of serverless computing and the trend is still yet to be defined. It’s impressive how fast the cloud providers are moving with serverless computing and building new features around it. Adoption from enterprises has also been fast. The challenges of serverless computing are that there are a lot of changes at once for an organization to adopt it, and this often requires cultural shifts as well. Serverless computing requires a new way of thinking for enterprises, which is a challenge. But the for enterprises who embrace it, the gains are worth it.