Imagine an organisation that is highly decentralised and spread across not just locations inside one city, or a country, but with sites all over the world and on each site you have employees performing almost identical tasks. Sounds familiar? Almost every multinational company on this planet has a setup like this so it is not really anything unusual. What is interesting however is how we utilize todays computing technology to simplify certain operational activities on these sites so that we can make our lives (in my case as an IT professional) easier.
One of the common phrases used in every walk of life is “Work smarter and not harder”. How does that really translate in to our dispersed environment? There are number of IT related solutions that I can talk about in this space, but what I wanted is to use a real life example to show how we can save time and money by centralising certain services used in our environment, do things smarter so that we do not have to work harder.
Companies we work for have different services that they consume in order to achieve the end goal. In my case I work for a large manufacturing company with number of manufacturing sites across the world. Our manufacturing process is state of the art and it utilises appropriate automation technology to maintain the leading edge. One of the essential services our IT organisation within the manufacturing sites has to provide is training environment for all the employees that utilize automation systems. If we utilize traditional methods of training we would have to maintain some sort of training environment on all of our sites in order to ensure best training experience. But is that the smartest way of doing things? Traditional method calls for as many training environments as we have sites. We need to be able to manage and keep up to date all these systems in remote locations. If a new site is started we need to deploy more of these environments. Few years back we decided that this is not really a sustainable model if we want to keep this environment current and relevant for our end user. We need environment that can be easily accessible by all, easily updated whenever our live production process changes and let us not forget that we should try and maintain this environment with minimal effort, both in relation to capital cost and headcount. So the best solution was to consolidate this environment and centralise in one location.
Now we are getting to the technologies that enable us to build such environment. Having a well-connected international sites is obviously a must. As high speed links between the continents are readily available these days so this is not really a problem. Having said that we have built our environment sometime in 2006 and the speed at the time was much slower than what we have today. So it is worth looking at what kind of traffic one will be transferring to and from your centralised environment so that you do not create bottlenecks on your network infrastructure.
As I said, we have built the training environment which is really a scaled down version of our manufacturing environment. This environment does not have to cope with the demands of the real production so it can be built on much smaller systems. Look at what you have in real life and scale down. If you need to cater for tens of thousands of transactions every second in production – your scale down model may only have to cope with couple of hundred. As we figured out, we were able to replicate a scaled down version of the environment on a handful of servers that fitted in couple of server racks. While still in its infancy (remember original environment was built in 2006), we have embraced virtualization and used it heavily in this environment. Using this technology we are able to replicate environment with hundreds of systems that we would normally find on a manufacturing floor and fit them on to handful of virtual hosts. Another smart move on our part made possible by emerging technology.
At the end we got a centralised consolidate environment used for training purposes of thousands of technicians across the globe. Systems are all in one place, managed and maintained by small group of engineers, easily updated and kept in check in relations to any changes in real live environment. We can keep the environment relevant for our customers and cater for different training requirements by using easily updatable and configurable virtual machines that can be tailored as needed. In fact, we pretty much let our customers play with the environment and destroy the data with every class. Not a problem really since we just restore all our systems to the gold copy after every class. So making mistakes during the training is not a big deal.
With some creative class scheduling, we never exceed our capacity by overloading our environment. After all it is only a scaled down environment and can only cope with so many users. And that is the beauty of this type of system – it was designed and built with specific purpose and spec in mind and as long as it meets the requirement the system will remain relevant.
This was our approach to specific centralised consolidated environment few years back and we used this model to come up with number of other environments like this. Obviously not everything is suitable for centralisation so you need to make sure you understand the ins and outs of your system before you can decide to centralise.
Once again new technology is here to help – one cannot talk about centralisation without talking about the cloud. And while I am not trying to advertise cloud here, it is fact that this is the latest tool in centralising and consolidating specific service in one location. Or is it, it is a cloud after all, it can be anywhere. So why not use that option if it makes us work smarter and not harder…