Storage has traditionally been one of the most expensive and complex infrastructure components within the data center. For most of the history of virtualization, however, it only appeared to become more and more expensive, largely due to the explosion in the growth of data and ever-increasing demands for higher storage performance by a similarly increasing number of applications. A primary aim of vSAN is to give service providers a storage solution that can contribute significantly to controlling these growing costs.

vSAN also fundamentally changes the management of storage, so that it becomes far more VM-centric. As a result, this helps to yield a reduction in OpEx cost impact, versus the upfront storage acquisition model. We estimate the reduction, based on our service provider industry experience, to be $3-$5 per gigabyte, with the ongoing costs of operating and maintaining a typical mid-range storage system at a run rate of between $5 and $8 per gigabyte, annually. These estimated operational costs consist of a number of elements, including:

The vast majority of these costs, typically estimated at around three quarters or as high as $6 per gigabyte, revolve around the ongoing management of storage.

With vSAN, VMware integrates the storage platform into the same core vSphere compute infrastructure. Therefore, storage and compute can be deployed and scaled as required, in a far more iterative and operationally simplified manor. On top of this single datastore storage pool, provisioning and management tasks occur at the VM-level, with each VM’s storage characteristics, including availability, being maintained at the VM-level. This VM-centric approach enables vSphere operational teams to deal with storage within the same skillset and in the same operational procedures as other vSphere administrative tasks.

As a result, vSAN has a potentially significant impact on storage acquisition and operations, which must be addressed by service providers who are considering making the move towards operating a software-defined storage model. For this paper, the aim is to focus on providing a hands-on analysis of the most time-intensive storage tasks within service provider data centers, where vSAN is having a significant impact.

There are a number of operational tasks that are not being assessed as part of this example. For instance, the routine replacement of failed drives, storage fabric operations, and off-site backup procedures. Even though these operations are not being addressed here, they do make up a significant portion of the overall operational cost of maintaining a storage platform.

The following is a summary of the estimates on which the TCO example is based and defines what portion of the annual costs of ownership for storage are made up by the time and effort involved in specific tasks. This provides a clear business case for the VMware hyper-converged model, making it a clear choice for adoption by VMware Cloud Providers.

The following table outlines the frequency of occurrence of the operational tasks we are measuring as part of the example TCO assessment. The aim here is to demonstrate how these activities contribute to the ongoing time and effort, and therefore the operational costs of storage.

Table 1: Estimated operational overhead for scenarios when using vSAN versus a medium-sized traditional array-based storage system with 1000 VMs.

 

As addressed later in this paper, a deeper analysis of each of the three common activities being highlighted above demonstrates the total impact of vSAN on these specific operations leads to a 33 percent reduction in the total storage management OpEx. However, these are examples only, which differ significantly across environments, and there are other variables, such as administrator skill level, hardware, software tools, or automation that might already be in place with the service provider. All of these factors have a measurable impact on a total cost of ownership analysis, and must be considered.