Tech made simple: Horizontal Elastic Scaling

In the digitally driven world of today, I think we can all agree that the latest software tools are only growing in power and influence? Specialist software, such as those used for data analytics or video editing, is now easy to come by and widely adopted by the mass market. The problem with this is the more complex they become, and the more data they process, the more computing power you need to run them. So on your device, you run a program and in all likelihood, you’ll soon need to open another, and another, and soon you end up with multiple programmes on one device! 

There comes a point when you realise your device isn’t powerful enough to handle it all, so you get a bigger and better one. Expensive.

But, did you notice that there are large chunks of the day in which your device sits idle? 

Here lies the opportunity for shared platforms and computer usage. 

Imagine if you and your colleagues could share the collective power of all your devices whilst only using one?

Your colleagues probably won’t be running power-hungry programmes all the time, so if you could share the processing power, you would get the performance you needed without needing to upgrade your device all the time. When you need more power, you could all group together to buy an extra device to share in the pool, so the costs are shared and everyone benefits.

This is how we do computing now. We use a cluster of connected computers, but when we run a programme, it works like it’s on a single device. If we need more power we add more computers, if we need less power we remove them. This is called horizontal elastic scaling.

It means that the computers are working at full capacity, running different user programmes and maximising their power. Instead of lots of users operating their own devices, we have fewer computers that are working hard, and fewer computers mean lower cost! From the user perspective, there is no difference in performance!

At Starcount we design our applications and services in this way, utilising technologies such as Docker to package up our programmes into standalone containers that can run on any device. These containers can then be deployed in cluster technologies like Kubernetes, Docker Swarm, or Nomad.

And that is how today’s technology is becoming more efficient and greater efficiency means fewer dolphins die.

Written by Starcount, Senior Data Engineer