You can glean virtualization cost savings with the right storage management plan. For most environments, managing storage is one of the more challenging aspects of technology. When
For virtualization storage management, there are a few main areas where virtualization cost savings can be realized. These include a reduced port count, thin provisioning and deduplication benefits.
Storage management savings: Reducing the port count
Reducing the port count is an important distinction for potential virtualization cost savings. Generally speaking, it is expensive to connect to a Fibre Channel SAN. These costs are primarily driven by host bus adapter interfaces and switch ports. Server consolidation allows the number of SAN endpoints to be reduced by having only the virtualization hosts connected to the fabric.
For most SAN switching equipment, the number of ports in use is a licensing option. The reduced port costs may not aid the current connectivity situation, but it can accommodate new systems.
Effectively every organization will incur an increase in total OS count during the migration to a virtualized data center. By consolidating the storage connectivity to a limited number of virtualization hosts, the incremental switch and port costs for net new servers can be contained to the host inventory.
But not every organization can make this argument for reducing a port count in their storage management plan. For server-based workloads that are already SAN-attached in the physical server world, consolidating these systems to virtualized environments makes sense. But for organizations that are new to shared storage as part of the initial investment in virtualization, this is a new cost. So, there may be no virtualization cost savings with storage management in these cases.
Thin provisioning reduces virtualization storage requirements
Thin provisioning provides more virtualization cost savings. It can reduce the virtualization storage requirements in two ways: virtualization-based thin provisioning or storage system-based thin provisioning. When VMware's vSphere added thin provisioning support to VM disk files, all major hypervisors now support thin provisioning on the virtualization engine.
Thin provisioning from the hypervisor creates a VM disk file that consumes only the disk footprint from the actual contents of the guest operating system. Storage system-based thin provisioning for virtualization generally creates a thin provisioned logical unit number (LUN) to the host systems. In this case, the disk cost to the greater storage system is what is actually used on the LUN.
In both cases, storage over-allocation is a risk that can become an issue. The general consensus is to let the storage system manage thin provisioning at the LUN level. Without thin provisioning, storage allocations can quickly allocate disk resources at factors of 30% or more disk consumption.
Virtualization storage savings with deduplication
The last storage management strategy that can realize virtualization cost savings is deduplication, which saves the storage system from writing the same patterns many times.
Virtualization introduces a cornerstone opportunity because the VMs will -- in most cases -- be on shared storage, and there can be a lot of similar components to each VM. The clear candidate for disk savings is the operating system component of the VM. The storage characteristics of like operating systems will have many similarities, making a great opportunity for deduplication efficiencies and more virtualization cost savings.
Unlike thin provisioning, deduplication is a storage system feature, and hypervisors generally can't assist in this space. But exceptions exist. VMware's linked clone functionality for VMware View and Lab Manager are popular examples. Linked Clones is a base VM that is launched many times for a large number of similar systems. The disk cost results in only the differential configuration between each subsequent VM.
Adequate planning for virtualization technologies is critical to maximize the potential savings through storage management. Although specific configurations and requirements will greatly determine the virtualization cost savings, rolling in each of these features if it makes sense can reduce the overall virtualization storage costs for consolidated environments.
About the expert
Rick Vanover (firstname.lastname@example.org) has the vExpert, VCP, MCITP, MCTS and MCSA certifications. He is an IT Infrastructure Manager for Alliance Data in Columbus, Ohio and is an IT veteran that specializes in virtualization, server hardware, operating system support and technology management. Follow Rick on Twitter @RickVanover.
This was first published in January 2011