JavaScript Applications with Node.js, React, React Native and MongoDB: Design, code, test, deploy and manage in Amazon AWS by Bush Eric

Author:Bush, Eric [Bush, Eric]
Language: eng
Format: mobi
Publisher: Blue Sky Productions Inc.
Published: 2018-07-05T16:00:00+00:00

You have completely autonomous teams (even geographically separated) that don’t know about the other services teams and each is creating and deploying services independently.

Each service gets deployed on the number of machines required to handle the load it needs to support.

You would monitor the transactions and loading needs of each service and deploy them out to your cluster as needed. The point of a sophisticated PaaS solution for a Microservices architecture is that it would handle this distribution and shuffling of services to balance them across machines.

Microservices frameworks

If you have split up your Node.js application into several smaller applications, each needing their own “npm start” command to run, then you will want to consider using Docker for each of these discreet Node.js applications. To do that, you will want to use something like Amazon ECS, EKS, or Fargate. There is also OpenShift Online, Kubernetes, Docker Swarm, Cloud Foundry, or NGINX Unit (Application Server and Service Mesh). These frameworks manage the deployment of your Docker containers to a cluster of machines.

Really sophisticated Microservices architectures may even employ some type of service discovery registry. There are frameworks such as Scenica that are available to do this. You need to justify that you really need this complexity and overhead.

There is a Node.js package named Hydra that you can download from NPM that might be of benefit to you is you are creating a more complex implementation. Hydra gives you things like service discovery, distributed messaging, message load balancing, logging, presence, and health monitoring.

Queues and worker processes

A common approach in backend services is to take requests that require longer running processing and accept that request immediately and return success, but actually just queue it up. Imagine you had a photo storage service that has an HTTP Post endpoint API that accepts images. You might want to create thumbnail images and do some image processing to do facial recognition and add meta data to identify people.

To keep your service scalable, you could have your web service accept the image, copy it to S3 and then place an entry into MongoDB or AWS SQS to be looked at later. Then you could have separate EC2 machines or AWS Lambda functions that scale and independently read the work requests and do the processing. This is how to manage the work and keep the web service very responsive.

Download

Copyright Disclaimer:
This site does not store any files on its server. We only index and link to content provided by other sites. Please contact the content providers to delete copyright contents if any and email us, we'll remove relevant links or contents immediately.