Data center virtualization is the practice of abstracting physical compute, storage, and network resources so they can be pooled, partitioned, and presented as logical resources rather than dedicated hardware. Instead of provisioning a physical server, storage array, or network switch for each individual workload, virtualization allows many workloads to share the same underlying hardware while operating as if each had dedicated resources of its own.
Data center virtualization is broader than any single technology. It encompasses server virtualization (running multiple virtual machines on physical servers), storage virtualization (pooling physical storage into logical volumes), and network virtualization (abstracting network resources from the underlying physical switching and routing). When an organization virtualizes all three layers and manages them through unified, policy-driven automation, the result is what's known as a software-defined data center (SDDC).
The Three Layers of Data Center Virtualization
- Compute virtualization. A hypervisor abstracts physical server hardware, allowing multiple virtual machines (VMs) to run independently on a single host. This is the most mature and widely adopted form of data center virtualization.
- Storage virtualization. Physical storage from multiple devices is combined into a single logical pool that can be managed, allocated, and scaled from a central console, independent of the underlying hardware.
- Network virtualization. Network resources are abstracted from the underlying physical infrastructure, allowing logical networks, segments, and other virtual networking resources to be created, configured, and managed independently of the physical network topology.
Why Data Centers Virtualize
Organizations pursue data center virtualization for several operational and financial reasons:
- Improved hardware utilization. Physical servers running a single dedicated application often use a fraction of their available capacity. Virtualization allows that unused capacity to be shared across multiple workloads.
- Reduced physical footprint. Consolidating workloads onto fewer physical machines reduces the number of servers, and the associated space, power, and cooling required to support them.
- Faster provisioning. Virtual machines and virtual networks can be created, resized, or decommissioned in minutes, compared to the lead time required to procure and install physical hardware.
- Improved disaster recovery. Virtual machines can be replicated, snapshotted, and migrated between physical hosts, which can simplify and accelerate certain disaster recovery and failover scenarios.
- Foundation for hybrid and private cloud. Virtualization is a foundational technology for private cloud and hybrid cloud architectures, since pooled, software-defined resources are what make on-demand self-service and elastic scaling possible on-premises.
What Are the Challenges of Data Center Virtualization?
Virtualization introduces its own management complexity alongside its benefits:
- VM sprawl. Because virtual machines are so easy to create, organizations can accumulate large numbers of underutilized or forgotten VMs that continue to consume compute, storage, and licensing resources.
- Physical-to-virtual visibility gaps. As workloads move between physical hosts, it becomes harder to track which physical server, rack, and location is actually running a given virtual workload at any point in time.
- Capacity planning complexity. Virtualized environments make it harder to predict physical infrastructure needs, since the relationship between VM count and underlying hardware requirements isn't always linear.
- Licensing and compliance. Software licensing tied to physical cores or sockets can become difficult to track as VMs move across hosts.
Track Virtualized Infrastructure with DCIM Software
Virtualization abstracts workloads from physical hardware, but the physical hardware still has to be tracked, powered, cooled, and maintained. Data Center Infrastructure Management (DCIM) software bridges that gap by connecting virtual infrastructure to the physical assets running it.
DCIM software supports virtualized environments by:
- Mapping VMs to physical hosts. An out-of-the-box connector for VMware automatically syncs virtual machine data with the physical servers hosting them, so teams know exactly which workloads and customers would be affected by planned maintenance.
- Supporting more efficient physical planning. By knowing which physical servers are hosting which VMs, teams can make more informed decisions about workload consolidation, hardware refreshes, and potential decommissioning of underutilized infrastructure.
- Maintaining a single source of truth. DCIM software tracks the physical assets, power, and connectivity supporting virtualized environments, providing a centralized view of the physical infrastructure supporting virtualized environments.
- Informing capacity planning. Correlating physical rack space, power, and infrastructure capacity with information about virtualized resources helps teams plan hardware refreshes and expansions more accurately.
Want to see how Sunbird's world-leading DCIM software helps you manage the physical infrastructure behind your virtualized environment? Get your free test drive now.