Practical tips for migrating virtual machines to Compute Engine
This article contains guidelines and practical tips for migrating computer production tasks to the Google Cloud Platform.
When you migrate a production task to the cloud, you must ask yourself how you deal with the various components of your infrastructure. A different method is used for moving data than for moving databases or computer resources.
Many customers who are considering a migration to the cloud do so primarily to save costs. With discounts for long-term use on virtual machines (VMs) from Google Compute Engine, the costs can indeed be considerably lower than when managing hardware or virtual machines in a conventional data center. Agent in Melbourne for Migrations from another cloud to the Cloud Platform provides the same price benefits.
Many more customers are doing it for flexibility. In the cloud, they can create virtual machines almost immediately and do not have to wait until all resources have been collected and delivered. With VMs in the cloud, companies can quickly run new apps, experiment with them, and disable them if necessary. This ability to experiment quickly, perform fail-fasts if necessary and to discover in a short time what works and what does not work is a huge advantage and leads to lower innovation costs. Departments in your company don’t have to worry about buying and then using the infrastructure for a little experiment. Even customers switching from other cloud providers see the benefits of a fast, global network and short start-up times for VMs.
Finally, many customers can consolidate their overhead. Usually, data centers need many different suppliers, each with their own relationship, billing model and contracts. By switching to the cloud, you can reduce this overhead considerably. Your employees no longer have to run a data center and can fully concentrate on what helps your company move forward.
Because you need computer resources, or Compute , for most technical production tasks, this article discusses the tasks that you must perform to migrate VMs, with recommended practices. Compute, however, is so essential for most production tasks that we also need to address other issues that make apps work, such as databases, messages, and analysis.
The migration process does not consist of just one huge step. The recommended steps are described in the following sections.
Calculate the costs
The first step, before you move VMs, is to calculate the costs of the move. This means that you investigate how much you pay for what you are currently performing in your data centers. These are not only the costs of the physical machines, but also the costs of the network equipment, power supply, cooling, personnel and the leasing of the data centers.
Once you have calculated these costs, you must create a cost model for the cloud-based on the following considerations:
- Which operating system will you use? Do you need a license for this?
- What types of virtual machines will you use? These come in different sizes, and the costs depend on the size. You must have an idea of the performance characteristics of your apps.
- What other services are required for your apps in addition to virtual machines, and what do they cost?
- Do you need licensed software for these apps? How much will that cost and is the software available in the cloud?
These are all costs that you have to consider before you even move one VM. The Cloud Platform sales team can help you with this.
If you already have a cloud provider, these calculations may differ. For example, you do not have to take into account the costs of leasing your own data center. But the principle remains the same: before you migrate, you need to understand the different billing models of your current provider and of Google. For example, if you migrate from Amazon Web Services (AWS), you can view a price comparison for virtual machines in the Cloud Platform blog .
Determine which items you want to migrate
When you know how much migration will cost, you can start looking at what you want to migrate. In modern companies there are many different types of apps, from apps for customers and back office apps to development tools and experimental apps. It is not wise to migrate all these apps at the same time and in the same way.
We recommend that you divide the apps into three broad buckets:
- Apps that are easy to migrate. These have fewer dependencies, are newer, are written internally, so do not need a license and are better suited to scaling and other typical features of the cloud.
- Apps that are difficult to migrate. These have more dependencies, are less suitable for scaling or have complex licensing requirements.
- Apps that cannot be migrated. Some apps may not be eligible for migration because they run on specialized or older hardware, have to stay in your data center to meet business or legal requirements, or are not allowed to run in the cloud due to complex licensing requirements.
These are just examples of the types of apps that fall into these three buckets. Your apps may have many more factors that determine distribution. The Cloud Platform sales team can help you with this.
These considerations are all important whether you are migrating from a data center or from another cloud provider.
After this step, choose the apps that you want to migrate first. We strongly recommend that you first migrate a small number of apps. This not only provides a model for future migrations but also helps define your migration processes.