Why all our VPS services include High Availability protection against hardware failure
30 Apr 2026, by Micron21
When you're looking to evaluate the options between different Virtual Private Server (VPS) providers, it can be surprisingly hard to know what to actually compare. In a lot of cases, there's very little information provided about the details that you'd actually need in order to make an objective comparison. For example, one CPU core is not equivalent to one CPU core on an entirely different architecture. The same is true for the speed and generation of RAM, as well as the type and brand of SSD storage. Unfortunately this kind of information is still often displayed in the abstract, making out that this inclusion of such a quoted vCPU count or RAM allocation somehow will provide you with a meaningful measure of the performance you'd expect to get.
Unfortunately though, this isn't just true and happening on the performance side - it's also happening on the reliability side. This is because it's often impossible to tell how reliable your VPS service actually truly is, outside of whatever uptime information the provider has chosen to share with you. What hardware is your VM (Virtual Machine) actually running on? How old is the host, the CPU, the RAM, and the storage itself? What protection exists for the data stored on the system? Is it protected via a RAID (Redundant Array of Independent Disks - which is the practice of combining multiple physical drives into a single logical unit so that data is duplicated and protected against drive failure), or is there no protection at all? And another important question that we'll be covering in this article... is the VM hosted on a single server, or on a cluster designed to protect you against all hardware-related downtime?
That's why, in this month’s article we'll be focusing on our VPS offering. We'll be covering the added lengths we go into ensuring that your services remain online, even when things go wrong at the underlying hardware level. Primary amongst these measures, is the addition of High Availability (HA). Specifically, that this protection is not just for our high-end services either - that it is included as standard for all of our VPS services, starting right from our base level smallest and cheapest plans.
What protections exist with most standard VPS plans?
Before we can talk about what protections we've put in place, it's important to first take a look at what's standard in the industry.
The typical VPS configuration with most providers looks something like this:
- A single physical server, also known as the host, running a hypervisor such as KVM, Xen, VMware, or Hyper-V.
- Multiple customer VMs running side-by-side on that single host, each isolated from the others.
- Local storage attached directly to that host, often configured in some form of RAID for drive-level redundancy.
- Possibly a backup process that runs on a regular schedule, copying data to a separate location for recovery purposes.
This is a perfectly serviceable configuration, and it's what powers the majority of the budget VPS market. However, in the name of trying to provide services as cheaply as possible, many providers will cut corners on the redundancy that genuinely keeps you online. They don't run multiple servers for redundancy, they may skimp on the RAID configuration and choose options that aren't as reliable as they could be (for example, RAID 0 striping with no redundancy, or RAID 5 instead of RAID 10), and they may even utilise older or cheaper hardware to push their costs down even further.
The problem with this approach is straightforward. Those measures help the provider hit a lower price point, however that comes at the cost of reliability. If the host's CPU fails, every VM running on that server is offline. If the RAM fails, the same applies. If the storage fails in a way that exceeds what the RAID is able to recover from, you're now facing an outage that could last hours or days while data is restored from backup. And throughout all of this, you have very little visibility or control over how quickly things get resolved.
Worse still, even when the hardware is fine, simple maintenance such as replacing a faulty stick of RAM or upgrading a host's firmware will often require taking the whole server offline. If your VPS happens to live on that host, your service goes down with it.
What we've chosen to do differently - High-Availability across the board
Rather than accepting these limitations as the cost of doing business, we've taken a different approach with our VPS services. Every single one of our VPS services - which, as long-time readers will know, run on our mCloud platform - includes High-Availability protection as standard.
So what does that actually mean? High-Availability, often abbreviated to HA, refers to the practice of designing systems so that they remain online even when individual components fail. It does this through redundancy. Critical components are duplicated, and when a failure is detected, traffic and workloads automatically shift across to the remaining healthy components without manual intervention. We've covered this concept in more depth in our The Importance of High Availability for Mission Critical Workloads article from a while back.
In our case, mCloud delivers HA through two complementary platforms - mCompute for the compute side, and mSAN for storage. Both are built on industry-leading open-source technologies, namely OpenStack and Ceph respectively, which we chose specifically because they were designed from the ground up with this kind of resilience in mind.
On the compute side, mCompute runs as a cluster of physical hosts rather than a single server. If one of those hosts experiences a hardware failure, such as a CPU fault, a RAM failure, or even a complete power loss to that node, the affected VMs are automatically restarted on a different host within the cluster. There's no manual intervention required, and the time taken to recover is usually measured in seconds to a couple of minutes rather than hours.
On the storage side, mSAN is built on Ceph and uses what's known as 3N replication. This means that every block of data written to your VM's storage is automatically at replicated at least three times across separate physical storage nodes. If any one storage node fails, your data is still available on the other copies, and Ceph automatically begins re-replicating the data to a new node to restore the full level of redundancy. From the perspective of your VM, the failure is essentially invisible.
Bringing it all back to the customer experience, this is what HA actually delivers. With other providers, a CPU, RAM, or storage failure on a host can result in extended downtime for your workloads, often with no concrete guarantees on how long it will be before your systems are back online. With Micron21, those same kinds of failures don't affect the availability of your services. Your workloads automatically rebalance across the cluster of servers, getting back online almost immediately.
It must be said, this approach does mean that our pricing isn't quite as low as it could be. Building and maintaining a clustered HA platform costs significantly more than running individual standalone hosts, and that cost has to be reflected somewhere. We may potentially miss out on the customers who would prefer the absolute lowest price above all else. However, we believe the trade-off is a worthwhile one. Reliability is one of those things that's easy to undervalue right up until the moment you need it, at which point the cost of an outage often dwarfs whatever you saved by going with the cheapest option.
Another benefit - Geographic High-Availability via a single click
Not only do all of our VPS services come with HA included as standard, but one of the more unique features of our mCloud platform is the ability to extend that protection across multiple geographic locations with no reconfiguration required from our customers. It's simply an option that you can select when signing up. Once enabled, your service is provisioned to span across three geographically separated locations rather than just one.
This feature is known as Geographic High-Availability (GeoHA), and it goes a step beyond standard HA. Standard HA, as covered above, protects you against hardware failure within a single data centre. GeoHA, on the other hand, also protects you against issues that could affect an entire site, such as a localised power event, a connectivity issue affecting the data centre's upstream providers, a natural disaster, or even physical damage to the building itself.
Behind the scenes, this is made possible by our mSAN Geo Diverse NVMe storage cluster, which spans three of our Melbourne data centre locations. Data written to your VM is replicated in real-time across all three sites via a 100 Gbps dark fibre backbone, ensuring that a complete site loss anywhere along that chain doesn't result in data loss or extended downtime.
Traditionally, this kind of multi-site protection has been the reserve of enterprise customers willing to invest very serious money. Setting it up yourself usually requires duplicated systems running in parallel across multiple sites, dedicated load balancers to direct traffic between them, replication tooling, monitoring, and the engineering hours required to design and operate it all. Those costs add up quickly, and for many organisations the price tag has historically been prohibitive.
With Micron21, the additional cost typically works out to be only around 25–35% on top of the cost of the VPS itself. Given the level of reliability and peace of mind that you gain in return - particularly for production workloads, customer-facing services, or anything that's revenue-generating - that's usually a very easy decision to make.
Bringing it all together
When you're shopping around between VPS providers, it's worth looking past the surface-level specifications. A vCPU count and a RAM allocation tell you almost nothing about how reliable a service will be, or what happens when the hardware underneath your VM eventually fails. Hardware always fails eventually - it's not a question of if, but when.
Most standard VPS services, in trying to hit the lowest possible price point, run on single hosts with locally attached storage and minimal redundancy. That's perfectly fine for some workloads, particularly development environments, low-impact non-production services, or anything where occasional downtime is acceptable. However, for anything that genuinely matters to your business, that level of protection simply isn't enough.
That's why we've chosen to include High-Availability protection across our entire VPS range, from our smallest plans right through to our largest. Every VPS runs on our clustered mCloud platform, with mCompute providing automatic compute failover and mSAN providing 3N replicated storage. For workloads that need the highest possible level of resilience, GeoHA can be enabled with a single click at sign-up, extending that same protection across three separate Melbourne data centres for a relatively modest additional cost.
Reliability isn't a feature that should be reserved for the most expensive plans. We believe that everyone deserves to have their services stay online, regardless of what they're paying.
Have any questions about our VPS services or High-Availability?
Our VPS services are the next step above our entry-level Shared Web Hosting plans - if you need added control over your hosting environment, or just want the peace-of-mind that comes from VPS services protected by HA, this is where to look.
If you have any questions about our VPS services, our mCloud platform, or how Geographic High-Availability could fit into your environment, let us know! We're always happy to have a no-obligation chat about your requirements and recommend the right configuration for your workload.
You can call us on 1300 769 972 (Option #1) or reach us via email at sales@micron21.com.
