| Server virtualisation uses a hypervisor to run separate virtual machines on shared physical hardware. Each virtual machine has its own operating system and applications. It can improve hardware use and make systems easier to move, but savings and recovery depend on the design. |
A server refresh raises the same question in any office with more than one box in the comms cupboard: replace each server like for like, consolidate them onto fewer machines, or move the workloads to a hosted service.
Server virtualisation is the technology behind the middle option. This guide explains what it changes, what it does not provide, and the costs and recovery decisions to work through before you commit to a design. It is written for owners and managers weighing up the decision, not for the team doing the build.
What is server virtualisation?
Server virtualisation runs several virtual machines, also called VMs or virtual servers, on one physical server. A layer of software called a hypervisor sits between the hardware and the virtual machines and shares the processor, memory, storage and network capacity between them.
What is a hypervisor?
A hypervisor gives each virtual machine its own operating environment while sharing the resources of the host. Isolation and resource controls help manage workloads, but capacity, network access and the host itself still need protection. If the physical host fails, every virtual machine relying on it may become unavailable.
Production servers use a bare-metal hypervisor, installed directly on the hardware. The hypervisor, its management accounts and the virtual networks between machines are part of the system to secure and update, alongside the virtual machines themselves.
What is a virtual machine?
A virtual machine is a software-based server with its own operating system, settings and applications. Because it is defined in software rather than tied to one box, it can be moved between suitably configured hosts and protected with backup tools designed for virtual machines.
These virtual machines share one physical host. Host failure can affect all three; this diagram does not show high availability or backup.
When does server consolidation make sense?
Consolidation suits workloads that are lightly used, run supported operating systems and have licensing that allows the move. Several single-purpose servers coming up for replacement at the same time is a common trigger, as are limited space, power or cooling.
It suits less well where an application needs specialist hardware, makes heavy and constant use of processor or storage, or comes with vendor restrictions on virtual platforms. Virtualising an unsupported operating system does not make it supported, so ageing systems still need a replacement plan.
Illustrative example. Consolidation can reduce hardware use. Total cost depends on licensing, capacity and recovery requirements.
Workload sizing is checked separately for each system. Peak demand, memory, storage performance, network traffic and growth decide the host capacity needed, and utilisation figures from different servers cannot simply be added together. Our IT assessment can review which of your current servers are realistic candidates.
What happens if the virtualisation host fails?
Consolidation puts several workloads on the same hardware. If that host fails, those systems may all stop together. Decide how long the business could tolerate that interruption, then plan recovery resources, backups and any availability features around that requirement.
Moving virtual machines, recovering them and keeping them available are different capabilities, and none of them follows automatically from creating virtual machines.
| Capability | What it provides |
|---|---|
| Virtualisation | Separate software-based servers on shared physical resources. |
| Live migration | Moves a running virtual machine between suitably configured hosts for planned work. |
| High availability | Restarts or fails over workloads through a designed resilient platform. Some interruption may occur. |
| Backup and restore | Recovers protected data or systems from completed backups, using available recovery resources. |
| Checkpoint or snapshot | A recovery point for certain changes. Not a substitute for independent backup protection. |
A second host or a cluster can reduce interruption from a hardware failure, but it changes the cost case. Not every small business needs that design; for some, a tested recovery process onto replacement hardware or a hosted platform is the better fit.
How are capacity and security managed?
Virtual machines share processor, memory, storage performance and network capacity, so a busy workload can affect others on the same host. Capacity is planned for peaks rather than averages, monitored after go-live and reviewed as the business grows. Microsoft documents storage quality of service controls for this kind of contention on shared storage.
Security works in layers. Isolation between virtual machines is one of them; the others are updating hosts and guests, restricting administrative access to the hypervisor and its management tools, controlling which virtual machines can communicate, and monitoring relevant activity.
What changes the total cost of server virtualisation?
Fewer physical boxes can reduce hardware, power, cooling and maintenance costs. Whether the project saves money overall depends on the full comparison, which should include:
- Hardware. The host or hosts, storage and any spare capacity needed when a host is unavailable.
- Licensing. Platform, guest operating system and application licences. Microsoft licensing guidance for virtual environments ties rights to licensed cores and virtual machine counts, so fewer servers does not automatically mean lower licence costs.
- Backup and recovery. Backup tooling, protected copies and recovery capacity for the new design.
- Migration and support. Planning, conversion, testing and ongoing management.
Compare that total against replacing each server like for like and against a hosted option over the same period, rather than assuming a payback.
Server virtualisation versus hosted cloud
Virtualisation is a way to run workloads; cloud describes how computing services are delivered. Virtual machines can run on your premises or on infrastructure run by a provider. Choosing virtual machines does not decide their location, the support arrangement or the backup coverage.
Keeping servers on site suits applications that need local performance or specialist hardware. Hosted servers remove the hardware from your building in exchange for a recurring service and a dependence on connectivity. Many businesses combine the two, which is covered in our guide to hybrid cloud.
Exodesk supplies server hardware and runs NZ hosting, so we have an interest in both options. We set out the costs and limits of each, including keeping your existing servers, and the choice stays with you.
What do migration and recovery require?
Converting a physical server into a virtual machine, moving a running virtual machine and recovering after a failed host are separate tasks. Microsoft describes live migration as moving running virtual machines between hosts, which needs a configured destination and suitable resources.
Recovery depends on usable backups, available capacity, application consistency, access and the recovery method. Copying active virtual machine files is not a complete backup plan, and Microsoft notes that a standard checkpoint is not a full backup. Configure and verify protection before cutover, not after the workloads are already moved.
Plan the move with a pilot, a change window, checks and a fallback, starting with lower-risk systems. Recovery targets sit in your disaster recovery plan, and the backups behind them can be run through our backup as a service.
Server virtualisation support from Exodesk
Exodesk can help compare retaining dedicated hardware, consolidating workloads and using hosted services, then deliver the option you choose. Our IT infrastructure team designs and upgrades servers, storage and hypervisors on site, and our cloud solutions cover hosted alternatives and cloud migration.
Exodesk has supported New Zealand businesses since 1989. Today, our Christchurch and Dunedin teams support businesses nationwide.
NEXT STEP
Considering a server refresh?
Exodesk can help compare retaining dedicated hardware, consolidating workloads and using hosted services. Agree compatibility, total cost, migration and recovery before choosing the setup.
Or explore Exodesk cloud solutions.
Frequently Asked Questions
What is server virtualisation?
It runs separate virtual machines on shared hardware. A hypervisor manages the resources, and each virtual machine has its own operating system and applications while relying on the underlying host.
What is a hypervisor?
It is the software that creates and manages virtual machines and their access to host resources. It supports isolation, but the platform, management accounts and virtual networks still need protection.
Can server virtualisation save money?
It can reduce hardware and running costs where workloads suit consolidation. Compare licences, migration, backup, spare capacity and support before assuming a net saving.
What happens if the physical host fails?
Virtual machines depending on that host may stop. Restarting them elsewhere needs the right availability setup or recovery process, with usable data and sufficient capacity.
Is server virtualisation the same as cloud?
No. Virtualisation is the technology, while cloud is a way of delivering computing services. Virtual machines can run on site or in the environment of a provider.
Does server virtualisation replace backups?
No. Backups still need application consistency, retention, protected copies and restore checks. A checkpoint or snapshot alone is not an independent backup arrangement.
How many virtual machines can one host run?
There is no useful universal number. Plan for workload peaks, memory, storage performance, network demand, growth and any capacity needed when another host is unavailable.
Is server virtualisation secure?
It can be secured, but isolation is only one layer. Protect and update hosts and guests, restrict administrative access, control network communication between virtual machines and monitor relevant activity.
Which workloads suit consolidation?
Suitable applications with supported operating systems and compatible licensing. Specialist hardware, heavy resource demands or vendor restrictions may justify a different design.
How do we get started with server virtualisation?
Identify workloads and dependencies, compare costs, check compatibility and agree migration and recovery requirements. Pilot the move with backups, a change window, checks and a fallback.
Contact us or call 0800 396 337 to talk about your next server change.
