There’s little need to question whether virtualization is one of the biggest IT trends today. Gartner continues to report that more than half of all server workloads are virtualized. By 2020, virtualization will account for 98 percent of IT growth spend. This increased virtualization spend will provide more efficient, cost-effective ways to reduce downtime, gain IT productivity, increase hardware savings and so much more.
Whether your organization has already invested in cloud computing or is about to move forward with a first-time migration, it might be critical for you to stop and consider the extent of the role of hypervisor systems in your environment.
Both VMware and Microsoft Hyper-V virtual platforms have become fairly comparable in a number of respects, each one represents a different set of advantages and disadvantages. We explore those here by directly comparing Hyper-V and VMware vSphere.
Both VMware and Microsoft Hyper-V implement server virtualization with a Type 1 VMM architecture. The products we are considering here represent two of the design subcategories in this architecture: Monolithic Hypervisor and Microkernelized Hypervisor.
A Look at VMware vSphere
VMware offers an array of solutions for virtual systems, but we focus only on vSphere in this article. vSphere uses a Monolithic Hypervisor Design, in which the device drivers are an integral part of the Hypervisor Layer.
A big vSphere advantage is its independence since an operating system isn’t necessary to control all of the virtualization components. Also, no security patches are necessary for the Controlling Layer components. Out of the box, vSphere offers many governance capabilities and organization can leverage transparent page sharing.
One disadvantage is the steep learning curve (for many users). But perhaps the worst is the fact that you can’t run vSphere on unsupported hardware—though VMware does provide a list of compatible hardware. Also, more initialization time is necessary because device drivers initialize in the hypervisor layer of the architecture. Any corrupt code in this layer can cause the initialization to slow down, or even cause the server to hang or crash.
- No operating system is necessary for controlling the management components
- No security patches necessary for Controlling Layer components
- Excellent vendor support
- Out-of-the-box governance feature set
- Available AWS apps
- Incompatibility with hardware that VMware doesn't support
- Complex device drivers will slow the initialization time
- Steep learning curve
- Corrupt external code may slow initialization or hang a server
- Trial software missing some functionality
How Does Hyper-V Compare?
Microsoft Hyper-V is capable of helping many companies build a private cloud, virtualize workloads, and scale public cloud services. Hyper-V is an integral part of Windows Server, but it is also installable separately as a standalone Hyper-V Server.
Hyper-V is built on a microkernalized design, so the device drivers run and operate independently in the Controlling Layer. These components are built into the Controlling Layer:
- Storage migration
- NTFS and SMB file systems
- Live/quick Migration
- Hyper-V Replica
The Hypervisor Layer is independent and contains both the Network Stack and Storage Stack. Similar to the monolithic design of VMware, the Hardware Layer includes the Physical Network and Storage Devices.
For most administrators who've worked with Microsoft products in the past, Hyper-V has an easier learning curve. Perhaps the best advantage is in how little effort is necessary to manage device drivers since new devices can be added without drivers—which means that a wide range of devices can be used with Hyper-V. Any necessary device drivers are installed directly into the operating system that is running in the Controlling Layer. These drivers are then accessible by virtual machines to access the hardware.
It takes Hyper-V just a few minutes to install and deploy new server roles in addition to the main virtualization role. Also, since the Microsoft hypervisor code is only 600 kB in size, initialization takes less time.
Unlike vSphere, there is no risk of corrupt code injection into the Hypervisor Layer, since Microsoft doesn't expose any APIs for that layer. In many respects, maintenance is much easier, since there is no downtime. Backups are also quick and easy, and you can perform live migrations faster.
- Minimal device driver management
- A wide range of compatible devices
- New server roles are easy to install
- Shorter initialization time
- High resilience to corrupt external code
- Zero downtime to perform maintenance or apply security updates
- Readily scalable services
- A crash of the primary OS will crash all VMs
- OS must be installed in order for the Hypervisor Layer to operate
- Frequent OS and security updates translate into more overhead
- Lack of support for service templates
It's important that you consider all the basic benefits of virtualization before exploring and comparing these two products. Both Hyper-V and vSphere have acquired a broad following, and many organizations have come to depend on these solutions. As businesses begin to virtualize networks, abstracting virtual machines and disks becomes common. Abstracting and managing this layer with software-defined networking (SDN) offers a multitude of benefits, including dynamic resource allocation, streamlined deployment, greater scalability and more. Acronis Cloud Manager is the unified solution that offers robust management and security for the private cloud and its virtualized infrastructure.
With Acronis Cloud Manager, admins can deploy, manage and optimize their virtual datacenter. This platform covers everything, from Hyper-V hosts and clusters to virtual network appliances. It offers capabilities similar to the Microsoft System Center at a fraction of the cost. Acronis Cloud Security can be configured as a virtual router and managed by a network controller. Your infrastructure will be protected by a virtual firewall, agentless antivirus, intrusion detection, network anomaly detection, deep packet inspection and network analytics with granular user access control.
Editor's Note: This post was originally published in January 2018 and has been completely revamped and updated for accuracy and comprehensiveness.