This content is part of the Essential Guide: Virtualization to cloud: Planning, executing a private cloud migration

Selecting the right virtual machine for your private cloud

So you’re ready to get into private cloud? Pace yourself. Choosing to cloud-enable the wrong VM could negatively affect performance.

Industry conversations have moved beyond understanding what cloud computing is to determining how best to deploy it in an enterprise IT environment. This evolution parallels the past debate of virtualization candidacy -- knowing how to determine if a physical server would make a good virtual server.

The idea of cloud candidacy, which means choosing the best candidates to cloud-enable, is just as important to reach the most return on your cloud investment. Here’s a checklist of how to determine the best candidate for your private cloud.

  1. What are the network requirements for the cloud workload? Much like server virtualization considerations of processing and memory requirements, the amount of network resources a workload or application consumes within your private cloud is critically important.

    A workload with heavy network resource requirements that link back to other servers within the LAN might not make the best cloud candidate. Placing this workload into a private cloud could create bandwidth and latency issues, affecting the overall cloud performance. Measure network utilization among servers residing in the cloud and any other servers located in the local data center.

  2. What other services are interconnected to the private cloud? Take into account any communication between a cloud workload that interacts with other workloads in the network. For example, almost any Windows-based IT service leans on Active Directory for authentication, as do desktop operating systems. Domain controllers must process large amounts of traffic, so completely moving them into the cloud is probably a bad idea.

    Alternatively, services that have little interaction with others in the local data center are less likely to be affected by their geographic relocation. This will also have little effect on latency. Looking for virtual machines (VMs) with few interconnections is a good way to find cloud-ready resources.

  3. Does an obvious separation (i.e., “hard line”) exist between workloads residing in the private cloud and other workloads? Isolating services can be a good dividing line between good cloud candidates and bad ones. Well-isolated services that make excellent cloud candidates, for example, are those services that reside within your DMZ. By design, the DMZ has “hard lines” surrounding its services, specifically in firewall rules that separate it from the Internet and an internal LAN.  That separation means that you could easily move all contents from a DMZ to the cloud. The very nature of a DMZ draws a clear demarcation for what might make a good private cloud candidate. Combine the isolation with virtual firewall rules from different cloud services, and DMZ services have the potential to be successful in the cloud.

  4. Looking for virtual machines with few interconnections is a good way to find cloud-ready resources.
    Is the cloud service replication-friendly? Some IT services are already designed with replication in mind; communication was designed to support the types of latency and bandwidth you would experience in LAN-to-cloud communications. These services could include offsite backup copies, failover servers or any service with protocols that are prepared for network conditions that don’t meet LAN speeds.

  5. How much pain will you experience if data that resides in your private cloud becomes compromised? This final check is really a direct result of early concerns surrounding cloud security and data ownership. Security, ownership and regulatory issues surrounding data must be a consideration when choosing to place a VM in the cloud. Some VMs and certain IT resources work with data that, if lost or compromised, wouldn’t incur significant pain for the business. 

    Pay attention to the types of data a potential cloud-enabled virtual machine could process. If the loss or exposure of that data is greater than the value gained in a cloud migration, then it might be best to keep that VM in the data center.

Choosing the right virtual machines for a private cloud requires you to analyze resource use of each VM and service. You also need to examine the risks associated with moving these resources to the cloud. With the right due diligence, finding the best resources and services to deploy in the private cloud should be relatively easy.


Greg Shields, Microsoft MVP, is a partner at Concentrated Technology. Get more of Greg's Jack-of-all-trades tips and tricks at

Dig Deeper on Managing cloud infrastructure