Web back-end development trends and scenarios
What you can expect in 2021–2022
In the backend, we’re already seeing far-reaching changes in the way servers are used and managed. In the coming years, we’ll also have a lot more options in databases, and edge devices will bring part of computing back to local environments.
The term “serverless” has many meanings. The point of the trend is that developing services will no longer require managing servers.
AWS Lambda, Azure Functions and similar services where code is run directly in the cloud are becoming more mature, and in many cases they can offer cost optimizations which may make them easier to adopt in projects compared to other new technological trends.
Support for integrating serverless functions into the rest of a service’s architecture is also becoming more mature and production-ready.
Running a server does not have to mean running it on physical hardware. What’s more, it doesn't even have to run on virtual hardware anymore. Today, more often than not, it means running in containers within distributed microservice architectures. Technologies such as Kubernetes, Docker and Fargate – usually provided as managed services by cloud providers – have been growing both in popularity and adoption, becoming the new normal.
The days of choosing which particular flavor of SQL server you were going to use are gone, and not a day too soon. Traditional SQL databases are still going strong, but at the same time, options for specialized databases are increasing.
Cloud-native database options for big data already exist – take BigQuery by Google for example – and AWS has also offered DynamoDB as a simple serverless database for a while now. Even traditional SQL databases are going serverless with cloud providers offering fully managed SQL servers based on MySQL and PostgreSQL.
The need for data and lack of ubiquitous high speed internet is starting to result in computing moving back from the cloud to local environments.
Increasing computing capabilities in edge devices make it feasible to analyse raw data locally and only send the results to the cloud.
Machine learning in the browser also means that the users themselves can use their devices for edge computing instead of relying on cloud services.
Apple’s new Macs based on the M1 chip are already demonstrating the incredible benefits of a custom system on a chip (SoC) with an integrated GPU, memory and neural engine, allowing complex machine learning workloads on budget consumer devices.