5 Things I noticed when comparing Scale Computing to VMware

cody.ault · 2025-06-10T17:50:57+00:00

VM networking was much simpler. In VMware, there is a lot of planning when managing the network. You create virtual switches per host and assign properties like MTU size, VLAN, etc and pair those with physical NICs. You setup VMkernel NICs with specific roles to control host traffic. With Scale computing, you set a native VLAN on the physical switches and use standard VLAN Trunking to the host. You assign an IP on the management network which uses the default VLAN and then tag VMs on their virtual NICs directly, skipping all the complex concepts of a vSwitch. A separate Linux bridge with different physical NICs or a subinterface with a VLAN is used for the cluster backplane used for storage and inter-node cluster communication depending on how many NICs the host has. If a VM needs to be on multiple networks, you can assign one NIC with a certain VLAN, and another NIC with a different VLAN. You can also choose not to tag at the port and let the guest OS setup VLAN tagging, similar to assign VLAN 4095 on a vSwitch in VMware. Scale only supports their hyperconverged storage. This simplifies things quite a bit compared to other KVM solutions. With the prescriptive approach to storage, you get a simple and easy to setup platform without the hassle of worrying about choices such as qcow2 vs RAW disk types, snapshot capabilities of the storage, changed block tracking and other decisions that might impact performance and features sets. Due to all the work put into their storage architecture, you have a very efficient platform that automatically works to take advantage of the storage presented to it from the nodes. It accepts a combination of SSD/HDD/NVME and will tier blocks to an appropriate level of storage. If you’re moving from VMware vSAN to Scale Computing HyperCore, the experience and features are comparable without an additional license to use the storage. You can get to the management web interface for the cluster by going to the IP address of any node in the cluster. You’ll see all the VMs and hosts in the cluster and there’s options to connect to a remote cluster to setup native replication of VMs. If you want to see all the clusters in a single UI, there is a SaaS based interface named Fleet Manager. This is particularly helpful due to the edge computing use cases for Scale Computing such as retail locations. There isn’t an on-premises equivalent to vCenter to manage multiple clusters. Like most Linux based hypervisors using VirtIO storage and network controllers, Windows will need to have the drivers installed for them to see the network and storage. This is typically done on an initial installation and is pretty simple. Most people will leverage a template with these drivers already loaded which has been anonymized with sysprep so the need to install Windows from scratch on a regular basis isn’t necessary. There are options to pick other drivers when a VM is created which are compatible across a wide array of operating systems, but they tend to have lower performance as a trade-off as they are emulated instead of paravirtual. Like VMware, it is recommended to install the Scale tools on guest systems to provide a mechanism for the hypervisor to communicate to the guest OS. Like VMware, Scale Computing is a proprietary system that has taken all of the complexity out of deployment. It is based on open-source technology, but you don’t need to be a Linux expert to install and configure Scale Computing. There are some basic command-line tools to setup/initialize a cluster that ask you to provide a couple IP addresses and license keys, but afterwards, most management is done through the web interface. There’s no need to install Linux first and then a slew of commands to install packages and dependencies and modify config files before deploying or complex installation scripts that require modifications like some open-source offerings. It is a very easy, tailored out-of-the-box experience and quick to get setup and running VMs.

VM networking was much simpler. In VMware, there is a lot of planning when managing the network. You create virtual switches per host and assign properties like MTU size, VLAN, etc and pair those with physical NICs. You setup VMkernel NICs with specific roles to control host traffic. With Scale computing, you set a native VLAN on the physical switches and use standard VLAN Trunking to the host. You assign an IP on the management network which uses the default VLAN and then tag VMs on their virtual NICs directly, skipping all the complex concepts of a vSwitch. A separate Linux bridge with different physical NICs or a subinterface with a VLAN is used for the cluster backplane used for storage and inter-node cluster communication depending on how many NICs the host has. If a VM needs to be on multiple networks, you can assign one NIC with a certain VLAN, and another NIC with a different VLAN. You can also choose not to tag at the port and let the guest OS setup VLAN tagging, similar to assign VLAN 4095 on a vSwitch in VMware.
Scale only supports their hyperconverged storage. This simplifies things quite a bit compared to other KVM solutions. With the prescriptive approach to storage, you get a simple and easy to setup platform without the hassle of worrying about choices such as qcow2 vs RAW disk types, snapshot capabilities of the storage, changed block tracking and other decisions that might impact performance and features sets. Due to all the work put into their storage architecture, you have a very efficient platform that automatically works to take advantage of the storage presented to it from the nodes. It accepts a combination of SSD/HDD/NVME and will tier blocks to an appropriate level of storage. If you’re moving from VMware vSAN to Scale Computing HyperCore, the experience and features are comparable without an additional license to use the storage.
You can get to the management web interface for the cluster by going to the IP address of any node in the cluster. You’ll see all the VMs and hosts in the cluster and there’s options to connect to a remote cluster to setup native replication of VMs. If you want to see all the clusters in a single UI, there is a SaaS based interface named Fleet Manager. This is particularly helpful due to the edge computing use cases for Scale Computing such as retail locations. There isn’t an on-premises equivalent to vCenter to manage multiple clusters.
Like most Linux based hypervisors using VirtIO storage and network controllers, Windows will need to have the drivers installed for them to see the network and storage. This is typically done on an initial installation and is pretty simple. Most people will leverage a template with these drivers already loaded which has been anonymized with sysprep so the need to install Windows from scratch on a regular basis isn’t necessary. There are options to pick other drivers when a VM is created which are compatible across a wide array of operating systems, but they tend to have lower performance as a trade-off as they are emulated instead of paravirtual. Like VMware, it is recommended to install the Scale tools on guest systems to provide a mechanism for the hypervisor to communicate to the guest OS.
Like VMware, Scale Computing is a proprietary system that has taken all of the complexity out of deployment. It is based on open-source technology, but you don’t need to be a Linux expert to install and configure Scale Computing. There are some basic command-line tools to setup/initialize a cluster that ask you to provide a couple IP addresses and license keys, but afterwards, most management is done through the web interface. There’s no need to install Linux first and then a slew of commands to install packages and dependencies and modify config files before deploying or complex installation scripts that require modifications like some open-source offerings. It is a very easy, tailored out-of-the-box experience and quick to get setup and running VMs.

Page 1 / 1

It is interesting how the virtualization landscape is changing. I am sure we will see much more of this to come.

Thanks for this post Cody. I have yet to dive into Scale as I’m looking at another h/v at the moment, but I do plan to. This will give me a barometer from which to gauge at least. Thanks for sharing!

Very cool. I enjoyed their booth at VeeamON and had a demo recently myself.. It seems like a great product and I really want to demo it when i can.

Thanks for sharing this information. As someone who has typically worked with VMware and Hyper-V environments, it’s always great to know some of the differences between platforms when it comes to usability.

Actually, one of the first user groups I ever attended in my tech career was sponsored by Scale Computing in Indianapolis. It’s nice to see them once again and also that Veeam supports data protection for these environments now.

Comment

Sign up

Login to the community

Scanning file for viruses.

This file cannot be downloaded