Immutable Infrastructure

Changing server configurations is a common task that can be wrought with problems. You take an existing configuration and alter the configuration to fit new requirements. This may be an acceptable solution for smaller companies with only a few servers or components. But what happens when companies grow and add more components? This situation describes mutable infrastructures, which is the traditional scenario that has existed for many years.

Immutable infrastructures can solve these problems as new servers and configurations replace the previous images. One reason why immutable infrastructures are becoming popular is provisioning new servers has become simple due to the cloud. Before cloud architectures, provisioning new servers would require the purchase of physical servers or components, and that could take weeks or months. With the cloud, IT can request new servers with the flip of a software switch.

Mutable infrastructures often create multiple configurations and are managed on a piecemeal basis for various reasons. The network may go down often on one server or may be slower than others and miss important, critical file feeds. IT may choose to add a fix to that one server. Now, the configurations are not in sync. Each time a new release is required, the out-of-sync configuration becomes part of the exceptions that need processing. The more exceptions there are, the more likely errors will occur, and tracking the history of these errors, is difficult, at best.

Mutable infrastructures lack stability. When you can change the server at will, it increases the risk of mismanaged configuration or configuration drift. It’s impossible to avoid this drift when you need to manage changes to configurations on multiple servers.

When working with immutable infrastructures, however, the configuration never changes. When changes are needed, a new implementation replaces the old one. Each server that requires the change will receive the new configuration. The old ones can either be deleted or versioned. If there are problems with a new configuration, every server that received the update can be rolled back.

In some ways, mutable vs. immutable infrastructure can be considered synonymous with physical hardware vs. virtual servers. While it’s possible to use mutable infrastructures with virtual machines, it is becoming less commonplace due to the advantages of immutable infrastructures.

What the benefits of immutable infrastructures?

Implementing this scenario will lead to reduced calls for support. If a configuration instance is causing a problem, support can delete the instance. It’s often easier to find where the problems occur.

Instances of immutable infrastructures can also be used as a litmus test of sorts. IT staff can check new instances against the stable version, and if it passes the test, and whitelist the instances. This also helps when factoring in security measures as the instances have already been tested.

Configuration drift is also eliminated, as was previously discussed. There are no multiple configurations of servers. All the instances requiring updates will contain the same configuration.

Immutable infrastructures are meant to work hand-in-hand with cloud technologies. Virtualization is the key to allowing this implementation to occur.

In traditional mutable infrastructures, what are the chances that your IT staff would experiment with different configurations to see which ones work best? The likely answer is that they wouldn’t dare touch any configuration unless there was an actual business reason to change anything. They would be too afraid to tamper with configurations that are currently working. It’s the adage, if it isn’t broke, don’t fix it.

Daring configuration changes are possible with immutable infrastructures because no aspect of the original infrastructure is touched. IT can try out as many images as they like, and if they find a better implementation, they can use that image as the new configuration. This experimentation benefits the organization by always searching for better configurations that can improve speed and stability.

Immutable infrastructures are tracked on the version rather than on a specific configuration implementation. This makes tracking changes much more efficient.

Another measure that warrants comparison is a server’s state. When you configure multiple servers with different configurations, each of these servers is potentially in a different state. This can cause unintended consequences that can be difficult to track down. Immutable infrastructures maintain the same state for the associated versions.

Disadvantages of immutable infrastructures include applying security updates in real-time. Also, suppose immutable infrastructures are against security policies. In that case, you’ll need to sell the security team on the benefits of the immutable philosophy. That is not always easy to do as security teams also stick to the adage, if it isn’t broke, don’t fix it.

You may be convinced after reading this that immutable infrastructures are right for your situation. But you’re concerned that the skills of your IT staff are not up to the task. Your team may be the best at handling traditional data center tasks. Cloud technologies are entirely different. It’s a valid concern, but one where our services at Eplexity can bridge the gap. We work to help augment your team. You won’t need to train your staff because we already have expertise in cloud architectures and immutable infrastructures.

Imagine having a proper immutable configuration set up for your servers and not wondering if you did it right. Our experts work alongside your current staff, and we ensure that the new infrastructures are correct. We’ll apply proper standards, and best practices and have you up and running quickly. Contact us today to have our experts implement your immutable infrastructures

Recent Posts