Five Rules of Automating a virtual data center

Send to a friend

Send this article to a friend.

So we're all tired about talking about cloud. In fact, it’s gotten to the point where we’re tired about talking about how people are getting tired about talking about cloud! The reality is that for the most part, organizations have already passed the point of no return, and are now looking for real ways to realize the benefits that the virtualization paradigm delivers. As it turns out, some of the most immediate benefits in the virtual datacenter are achieved through automation.

 

According to Wikipedia, "Automation is the use of machines, control systems and information technologies to optimize productivity in the production of goods and delivery of services. The correct incentive for applying automation is to increase productivity, and/or quality beyond that possible with current human labor levels so as to realize economies of scale, and/or realize predictable quality levels."

 

I really like the second part of this definition, because it reminds me that most of us are interested in technology and are truly amazed at the possibilities it presents - so much so that we often forget the most practical incentive of applying automation: because it saves organizations time and money.

 

More importantly, it can make my life as an IT administrator much easier by automating repetitive tasks and ensuring that complex tasks are performed consistently. As a result, potential problems are reduced and I am able to extend this functionality to much less sophisticated users while still maintaining control.

 

Of course automation is not new. Those of us experienced with test automation have been through some of this before, and can provide simple rules about automation that are noteworthy. Here are some of the more obvious, or in some cases, not so obvious examples:

 

1. Automation is only practical when you have a repeatable process. We've all seen it before - while implementing a new process or tool, an organization doesn't have clearly defined process requirements. There are requisite tweaks and tunings that need to occur. Sometimes, you don't know exactly how things should work and time must be invested to unravel the kinks before automating – to avoid continual changes and tuning efforts. In the testing world, a rule of thumb is a test must be consistently repeatable and used more than five times before it makes sense to consider automating it.

 

2. Pick the area of most return. Sure, it is cool to fantasize about an entirely automated process that provisions VMs in a production environment, but if I'm only adding on average 4-6 VMs a quarter to these environments, it hardly seems worth the effort. Focusing instead on more dynamic environments such as non-production testing/QA, development, and lab environments where we may be creating or destroying 30-40 VMs or more a quarter is more likely to increase overall productivity. Non-production environments also have the added benefit of providing a lower risk option to refine the process before deploying it further.

 

3. Ensure your automation suite has a good ROI. Evaluate ALL of the costs. When costing an automation solution, it is easy to simply consider the licensing and maintenance costs and forget about the effort required to initially configure the product, train users, and reconfigure it when the requirements of the organization change. These activities add significantly to the cost and can lead to productivity and morale issues as staff members become increasingly frustrated with needlessly complex tools. Products without a good ROI impede the goal of increasing productivity.

 

4. Evaluate your automation requirements. Do you need a mini-van or a Porsche? Sure it's great that the solution that you're planning to implement has the power and features to automate a Rube Goldberg device, but will your organization make use of it? Will those features and associated complexities deliver value, or will you simply be paying for something that will never be used?

 

5. 100% automation is not always possible. Sometimes simply automating the process, instead of full automation of the entire end-to-end solution can provide just as much benefit. Focus on where the benefit lies and avoid automating what is not necessary. This will still translate to faster provisioning times and overall better service to end users.

 

Automation in the datacenter is a game changer, and for the first time, organizations are standardizing enough through virtualization to take advantage of its benefits. Making sure your organization knows where to start will allow you to capitalize on your virtualization investment!

Matthew Dufresne is a senior virtualization and automation specialist with Embotics – a leading software vendor for V-Commander, providing data centre automation, reporting and streamlining.  Matt has worked with organizations on a global scale and assists with the resources at IDS Systems to facilitate data centre optimization.  Please visit our web site for more details:  www.idssystems.com  You can contact Matt by sending a note to Embotics.Matt@idssystems.com.

  • 1
  • 2
  • 3
  • 4
  • 5

Thanks for voting!

Top of page

Comments

Comments