What is server consolidation?

Server consolidation is the process of reducing the number of physical servers in a data center by combining workloads. If a data center underutilizes server hardware, then admins can configure servers to host multiple workloads. This means servers can better utilize hardware, which reduces the total number of servers required to run workloads.

What are the benefits of server consolidation?

Benefits of server consolidation include reduced hardware maintenance costs and the overall data center footprint, which is beneficial for organizations that lease space in a colocation facility.

Admins should consider several factors during the early phases of their server consolidation planning process. For example, admins should outline specific goals, determine whether one consolidation method is better suited to achieve those goals, devise a plan to deal with possible security and compliance violations and consider whether the cost savings justify the total expense of the server consolidation project.

Types of server consolidation

There are two ways admins can consolidate servers. First, admins can run multiple workloads on a server’s OS. The other option is to use server virtualization. The idea here is to convert physical servers into VMs. As virtualization hosts can accommodate multiple VMs, the net result will be fewer servers running in the data center.

Server consolidation architecture

Before deciding on one of the two server consolidation methods, consider the advantages and disadvantages associated with the two architectures.

Migrate workloads to a server’s OS

The first server consolidation method involves combining multiple workloads onto a single server. The main advantages associated with this technique are that it’s relatively easy to implement and it’s less expensive than using server virtualization. Only a single OS license is required and there is no need to purchase and learn how to use virtualization software.

Although this type of consolidation is simpler and less expensive than using server virtualization, there are some disadvantages to consider:

• Potential interference with other workloads running on the server. Because virtualization software isn’t being used, there is nothing to keep a workload’s resource utilization in check. If a workload consumes excessive hardware resources, other workloads running on the server will suffer.
• Security implications. Because multiple workloads share a common OS, it means that if a single application has a security vulnerability, then the application can act as a point of entry for attackers, who could then compromise all of the other workloads running on the server.
• Server outage. If the server were to suffer an outage, then all of the workloads running on that server would incur downtime as a result. As such, the failure of a single server could cause a major outage in the absence of a high-availability platform.

Use server virtualization

The second method of consolidating servers is through virtualization. When admins use virtualization, the physical server takes on the role of a virtualization host.

There are numerous advantages to using virtualization technology for workload consolidation:

• Isolated workloads. Applications run inside VMs with their own OS, which helps isolate workloads from one another. A security breach within one virtual server typically doesn’t affect another.
• Fault tolerance for VMs. Fault tolerance means that if a problem were to occur on a virtualization host, then all of the VMs running on that host can automatically fail over to a different host, thereby preventing downtime.
• Automation for monitoring sufficient hardware resources. Similarly, automation can be used to monitor VMs to make sure that they are receiving the required hardware resources. If a VM needs additional hardware resources to deal with an activity spike, then the underlying virtualization technology can automatically handle the resource allocation. If sufficient hardware resources aren’t available, then the virtualization software can even be configured to automatically migrate the VM to a different host where the required resources are available.

However, virtualization isn’t without its disadvantages. The main disadvantages are that virtualization is often expensive and complex to implement. Not only is there the software license cost to consider, but you might need to purchase additional hardware, such as a storage array that can accommodate your VMs.

What to consider before consolidating servers

In most cases, virtualization is the preferred tool for server consolidation. Even so, there are several factors admins should consider during the virtualization planning process.

Data center capacity

One of the first factors is data center capacity. Admins must consider the number of physical servers required to host the anticipated number of VMs. It’s worth noting, however, that VMs aren’t all created equally. Some can be larger than others, based on workload requirements, and it’s important to make sure that virtualization hosts can provide the required capacity.

It’s also important to design the IT infrastructure with more capacity than is actually needed. This extra capacity is useful for accommodating future workloads. More importantly however, it is needed in case one or more virtualization hosts were to fail. VMs will only be able to fail over to another host if the destination host has enough capacity available to handle its existing workload, plus the workload from the failed host.

Virtualization architecture

Another key step is to decide which virtualization architecture to use, such as Hyper-V or VMware. Admins must then decide how many server nodes to include in a cluster and choose a storage architecture.

Public cloud hosting

It’s also a good idea to consider the public cloud as an option for hosting VMs. Cloud providers offer virtualization platforms as a managed service. This means that organizations can focus their server management efforts on the VMs themselves without having to worry about the underlying service-level infrastructure. The cloud computing provider handles low-level server management and optimization on the customer’s behalf.

Using the cloud for VM hosting can be a good option since it’s less complex than hosting VMs on premises. This approach also reduces energy consumption within an organization’s data center because servers are running in the cloud rather than running on site.

There are some disadvantages to the public cloud. For example, cloud providers don’t give their customers full control over the virtualization environment. Those needing granular control might be better off with an on-premises platform. Depending on the use case, running virtual servers in the cloud might be more costly than hosting those VMs on premises, though the opposite can also be true.

Server consolidation costs

As admins progress through this process, they should keep an eye on the project’s cost. Hypervisor licensing is one of the first costs to consider, but admins might also require supplementary management and monitoring tools, or capacity planning and management tools. In addition, admins might require a support contract depending on which virtualization platform they choose.

Migration methods

Finally, admins should consider how to migrate their workloads to a virtualized system. Options vary by workload, but the migration process should be seamless and avoid service interruptions.

How to consolidate servers

The server consolidation process is going to vary widely depending on the consolidation method being used and the types of workloads that are being consolidated. The consolidation process is also going to vary based on whether workloads will continue running on-premises or if those workloads are being migrated to the cloud.

There are two main consolidation methods that tend to be used, though there are countless variations.

1. One technique involves using software to assist in the consolidation process. For example, cloud providers often offer tools to assist in the process of migrating a server’s workload to a cloud-based VM. There are also software companies that offer database migration tools or physical-to-virtual conversion tools that can automate the VM provisioning process.
2. Another approach that’s often used involves using disaster recovery software. This method involves backing up a physical server’s workload and then restoring the backup to the physical or VM to which the workload is being migrated. A more complex variation of this technique involves using disaster recovery software to perform a workload failover to the destination host, thereby enabling the workload to be migrated without incurring downtime.

Brien Posey is a 15-time Microsoft MVP with two decades of IT experience. He has served as a lead network engineer for the U.S. Department of Defense and as a network administrator for some of the largest insurance companies in America.