Method & sources

Sources

Every answer is taken from the OEM's own website - product pages, datasheets, documentation, licensing guides, price lists and support or lifecycle policies - and links to the page it came from. Where an OEM does not publish something (most often a price), the tool says so and shows a labelled estimate with its basis. OEM data was last reviewed on 2026-09-19.

Scores

Most points carry a capability score from 0 to 3:

• 3 - native & strong. Built in, and a strength of the platform.
• 2 - supported. Available and documented.
• 1 - limited, add-on or third party. Possible, with restrictions, extra licences or another product.
• 0 - not available.

For lock-in a higher score means less lock-in; for cluster sizes and support response, a higher score means the more favourable answer. Descriptive points (a hypervisor name, a licence metric) are compared as text.

Ranking

Capability is the weighted average of the scores over the scored points you selected. Each use case you pick raises the weight of the points that matter to it (for example DR orchestration for disaster recovery, two-node support for edge). Use-case fit is each platform's stated fit for those use cases. Overall blends the two 60/40. Platforms that are end of sale are shown for reference and their overall score is halved so they are never recommended. The ranking sorts your selection; it is not a product rating.

Sizing

1. Physical cores needed = VMs × vCPU × (1 + growth) ÷ vCPU:core ratio.
2. Memory needed = VMs × RAM × (1 + growth).
3. Raw capacity needed = VMs × disk × (1 + growth) × the platform's data copies ÷ its assumed data reduction ÷ 75% fill.
4. Each node gives its cores and RAM less the platform's own reservation (controller VMs, services).
5. Nodes = the largest of the three, plus one for failover (N+1), then the platform's minimum and maximum cluster sizes and two-node rules apply.

Cost estimate

• Software = licensed units × the edition's price per unit per year × term. Units follow the OEM's metric: physical cores (with per-CPU minimums), sockets, nodes, TiB, VMs or clusters, and any minimum order.
• Capacity add-ons apply where a licence entitles a fixed amount of storage per core.
• Hardware = nodes × your price per node × a delivery factor (appliance 1.15×, disaggregated 1.1×, software-defined 1×), plus hardware support at your percentage per year.
• Windows Server guests, if chosen: all host cores licensed with Datacenter (16-core minimum per host) at $6,771 per 16 cores (Microsoft pricing ↗), unless the platform's edition already covers it.
• Services = your percentage of first-year software and hardware.

Everything is indicative. Discounts, regional price lists, promotions and multi-year terms move real prices a long way; ask the OEM or an authorised partner for a quotation.

Platforms and when they were reviewed

Accounts and privacy

HCI Hero uses your niyyasoft.com account - the same sign-in as the rest of the site, with email and password or Google. It never sees your password. Guests can use every comparison; saving and downloads need an account. Saved comparisons are stored against your account on niyyasoft.com and nowhere else.