In the last few years at Scale Computing, we have expanded our appliance offerings to continue providing the most computing resources at the best cost. The total cost of ownership of our solutions is an important factor in how we design hyperconverged and edge solutions and in this blog post, I want to talk about some of our newest options that give you flexibility and help manage costs.
Single-Node Configuration
For some Scale Computing appliances, where the storage configuration permits, running a single appliance rather than a cluster of three or more can be a cost-effective solution for disaster recovery, backup, or even primary computing at a small site. The ability to run Scale Computing HyperCore on a single appliance is not new. It has been around for a couple of years now but some recent appliance updates have expanded the appliance options for single-node configuration.
The most recent appliance edition, the HC3250DF, is an all-NVMe storage appliance that is well suited to running as a single node and delivering high performance for workloads and applications. The HC3250DF can also be configured with some higher-end CPUs and up to 1.5TB of RAM and that can allow for a good number of virtual desktops for a VDI solution. This makes the HC3250DF a good candidate for running both applications and VDI at a remote office or a small business, as examples, to support up to a few dozen employees.
Another appliance of note for single-node configuration is the HE100 Series. The HE100 is a micro-appliance that is designed specifically to support the computing needs of small remote locations such as branch retail locations or even a micro-business that is getting IT services provided by a Managed Service Provider.
The HE100 is not just small in size, designed to run on top of a desk or shelf, but it comes with a very affordable price tag. So small that clustering it for high availability may make more sense, but we know that for some businesses and organizations, every penny counts in the budget. This is why we’ve enabled single-node on the HE100 where even a modest cost can become very large especially when multiplied across dozens or hundreds of sites.
How Scale Computing is able to offer such small yet powerful appliances like the HE100 is a whole other topic but you can find out more about that and more ways SC//HyperCore differs from other HCI and edge computing solutions in our The Scale Computing Platform Difference white paper.
You may be wondering if there are any special considerations for deploying a single-node appliance. As I had just mentioned above, appliances must be clustered for high availability. If there is only one appliance and that appliance fails, your system is down, which would require workloads to be spun up on a secondary system. That second system could be local, remote, or in the cloud. Commonly, we see single nodes being deployed at branch locations and replicated back to the data center. If there is no central location, a local replication system or the cloud may be used.
We would always prefer that Scale Computing systems are configured in a highly available cluster as that provides the greatest system uptime and the least impact on business and revenue when a failure occurs. And that is why the next topic is our new storage-only mode designed to lower the cost of clustering.
Storage-Only Mode
With our latest SC//HyperCore release in version 8.8, we introduced a new storage-only mode for our appliances. This mode is exclusively for use in clustering and can put one or more appliance nodes in a cluster into a mode where they cannot run VMs but instead only contribute storage to the cluster.
But, you may be wondering, what is the point if you have to buy the node anyway? Well, there are two points. First, a cluster node in storage-only mode does not need software licensing to run VMs. For example, if you would normally buy a Windows Datacenter edition license on each node of the cluster to run VMs, you would need one fewer license for each node that was set to storage-only mode. With some of our smaller cluster configurations, licensing can be a substantial part of the total cost and storage-only mode can help alleviate some of that cost.
Secondly, if you are planning to place an appliance into storage-only mode, the resource requirements for that appliance can be much lower. Fewer CPUs, less-expensive CPUs, and less RAM are required if that cluster node is only contributing storage. And even if that node is now only contributing storage, it may be that the storage contribution can be smaller as well. That all means that a cluster could be designed, for example, to have two larger appliances as the nodes that can run VMs to provide high availability for the VMs, and a third smaller, less expensive node that contributes storage but primarily acts as a cluster witness and fulfills the third node requirement for the cluster.
Check out more information about the real costs and capabilities of different cluster configurations in our On-premises Edge Computing Infrastructure for the Real World white paper.
Storage-only mode is just one more option Scale Computing provides to help keep costs down where a full-fledged three(or more)-node cluster may not be necessary logistically or financially. We want to encourage you to talk to some of our experts to find out how a solution from Scale Computing can help you lower and control the total cost of ownership of your IT infrastructure solutions. Email us at info@scalecomputing.com or call us at 877-722-5359 for more info on our appliances, single-node configuration, storage-only mode, or any other way we might be able to help lower your IT costs.