Docker - A revolution for IoT world.

Today, June 2nd, 2016, I will share a brilliant practical capability of Docker in IOT on the device side. 

I first heard this name long ago while building a Yocto image (An embedded Linux version ) in a virtual environment. Still, I can bet that we will have to adapt to this in the future for many reasons, which I will explain below.

To understand Docker, please go through the below-mentioned videos.

Difference between docker and virtual machine - 

Docker and Container -

So, as per the above tutorial, Docker is one of the standards for the Container technologies introduced by Linux a long time ago. A few initial container-based standards are groups and namespace/userspace.  Now, the objective of the container is to save the rework of fulfilling the environmental dependencies of different applications on different machines.  For example, you created an application on a Ubuntu environment, but you want to run that over a server hosted in a Windows environment. Now without a container, you have to either create a virtual machine with ubuntu over it or fulfil the dependencies over the Windows by installing dependent packages.

With a container, you need to run a docker engine over a different machine and, on top of it, place your docker application's container. That container is self-sufficient to provide all the dependencies required for the application.  Now the question is, with Docker, I can save the workload of the cloud's dev ops guys by just providing a self-sufficient application environment in the container, but how would this be beneficial over the device side? Docker runs over the OS and provides a self-sufficient environment for applications.

Over the device side, where we place different sensor collaborations and gateway stacks, we used to do much work writing different JNI interfaces for reading sensors' data or MARRA /UPM changes for adapting different sensors.  These things are tightly bound with OS capabilities, for example, for one of the esteemed clients I had written a gateway stack with all necessary changes based on a generic yocto image,  Now after some time as per market demand the client requested to move the same stack on much more advance embedded Linux flavor (Wind River) for higher security grounds. Now, if I was far-sighted at the beginning of my development, then I must not have made the mistake of developing a complete stack over a yocto environment and fulfilling the required dependencies, i would have worked only on one single docker image and placed the same image on all the future requested embedded OS .  Well, that's the reason Docker is beautiful and future !!