Methodology

How a capability gets onto this site, what a gap marker does and does not claim, which versions each comparison was measured against, and why there is no price row. If you think something here is wrong, this page is also how to say so.

What this site is

CompareHypervisors maps enterprise virtualization platforms to each other one product at a time. For every capability it names the actual product on each side, the edition or add-on that carries it, and the cases where one side has no equivalent at all. It is not a scorecard: there is no total, no winner and no star rating, because the whole argument of the site is that which rows matter is a property of what you run and not of the platforms.

There is no vendor affiliation or review arrangement behind any page, no affiliate links, and nothing on the site is for sale. Analytics count visits and which comparisons get opened, and you can opt out at any time.

How a capability gets on the list

A row has to pass three tests. It must be something a buyer would put in a requirements document rather than a feature a vendor invented for a slide. It must be answerable for every platform on the site — including by saying that a platform has no answer. And it must be checkable against something the vendor publishes, not against opinion.

The list is deliberately weighted toward the boring end of the job: licensing terms, lifecycle, data protection, management and the things people discover after signing. It also carries rows where a commercial suite has nothing to show, such as source code availability or running in production at no cost, because a taxonomy assembled only from what the market leaders ship would decide the argument before the evidence arrived.

The same list is applied to every pairing. Rows are not added or dropped to suit a platform, and a platform is not excused a row because that row is awkward for it.

What a gap marker means

A dashed, unfilled cell reading no direct equivalent is a narrow claim: that the platform ships nothing of its own that answers the row. It is not a claim that the need cannot be met. Nearly every gap on this site has a route around it — a third party, an open-source component, a cloud service, or a change in how the workload is built — and the detail panel for the row says which. What the marker tells you is that the route is yours to find, buy, integrate and support.

A capability the vendor ships as a technology preview, early access or limited availability in the scoped version is not a gap. It gets a product and a small dashed preview chip with a half-filled circle after the name, in the map and in the detail panel, and the tier line says what the status means: usually that the vendor does not yet support it in production. The row counts as an equivalent; the marker tells you how far to trust it.

The opposite failure matters just as much. A row with a product on both sides is not a declaration of parity, and several pairings on this site are directionally right and materially uneven: a basic file share against a product that displaces a NAS is still one row. Where that happens the detail panel says so in as many words. Read the row, not the shape of the table.

Gaps are counted, never estimated. The totals quoted in each comparison's analysis are derived from the data on the page rather than typed into the prose, so they cannot drift out of step with the rows when the list changes.

Versions, and the date on every page

Each comparison is measured against named releases, and the versions are printed on the card and again in the analysis panel: what was current at the time the research was done, not whatever is newest now. A capability that arrives in a later release is not in the table until the comparison is re-researched.

Every comparison carries the month it was researched, shown as a pill on its card and again under its title. That date is the honest expiry notice for the page. This industry re-prices and re-bundles faster than any static comparison can track, so a comparison several months old should be treated as a starting point for questions to put to a vendor rather than as a current quotation. When a dataset changes, its date changes with it.

Sources

Rows are written from vendor documentation, licensing guides, datasheets, release notes and knowledge-base articles, and every comparison lists its sources at the foot of the page. Primary sources are preferred to secondary ones in every case. Where the only available statement is a vendor employee's post on a vendor forum, that is used and labelled as what it is.

Sometimes a vendor's own documents contradict each other. Where that happens the page says so and quotes both rather than picking the reading that makes a tidier sentence.

Why there is no price row

This is the most common question about the site, and the omission is deliberate.

List prices for enterprise virtualization bear almost no relationship to what anyone pays. Discounting runs deep and varies by volume, term, region, partner, renewal timing and what else is on the paper; a published figure would be accurate for approximately nobody, and a comparison built on two such figures would be worse than no comparison at all. Quoting them would make the site look more decisive and be less true.

What the licensing rows do instead is describe the structure of the cost, which is stable enough to be worth writing down and is usually what actually decides the budget: what the unit of counting is — core, socket, node, VM or user — and what minimum applies to it; which capabilities are withheld at lower tiers and therefore become a purchase rather than a setting; what sits outside the top bundle as a separate SKU; whether anything perpetual remains; and what a platform costs to run without a subscription at all. Two platforms can carry similar list prices and produce very different invoices for the same rack, and it is the structure that explains why.

Where a vendor publishes a figure openly and it is genuinely comparable — a per-socket subscription tier, for instance — that figure appears in the row. The difference is between citing a published number and manufacturing a total.

Corrections

Errors on a site like this are inevitable: products get renamed, bundles get re-cut, and a documented limit quietly changes in a point release. Corrections are welcome and get made, and a corrected dataset gets a new research date.

If something here is wrong, write to corrections@comparehypervisors.io with the comparison, the row and ideally a link to the vendor document that settles it. Claims that come with a primary source get handled quickly. Claims that do not are still read.

Reuse

The comparison data on this site is published under the Creative Commons Attribution 4.0 licence. Quote it, copy rows into a document, build on it, use it commercially — all of that is fine, on one condition: credit CompareHypervisors and link to the comparison the rows came from. Each comparison's structured data carries the same licence, so a machine reading the page gets the same answer.

Product names and logos remain the property of their vendors. The licence covers this site's own work: the capability list, the pairings, the gap markers and the analysis.

Back to the comparisons