Files
| File | Content | Size | Schema |
|---|---|---|---|
| vendors.json | Vendors, products, categories, deployment, licence model, ownership, capability levels and line items, each with source ids | 226 KB | schema |
| line_items.csv | One row per price: vendor, SKU, unit, commitment, price (empty when the vendor does not state it), source ids | 56 KB | schema |
| sources.csv | Every cited page: id, vendor, kind, URL, title and the date it was read | 118 KB | schema |
| costings.json | Which line items each vendor's estimate uses, with the quantity formulas, allowances, tiers and minimums | 34 KB | schema |
| profiles.json | The illustrative Small, Mid and Large workloads | 2 KB | schema |
| taxonomy.json | Definitions: categories, capabilities, levels, units, source kinds, confidence, workload variables | 7 KB | schema |
| MANIFEST.json | Dataset version, export date, counts and licence | 1 KB | schema |
| golden vectors | Every costing × workload × sampling scenario with its expected result, for checking an engine against |
Counts
- 34 vendors
- 634 line items
- 523 line items priced
- 1,006 capabilities
- 762 sources
- 30 costings
- 3 profiles
- 360 golden vectors
Rules the data follows
- Prices are vendor list prices. A price is a number only when the vendor states it: on its own pricing page, in its documentation, or in a marketplace listing it sells itself. A price that only a reseller or another third party states is null, and any total that needs it is a floor.
- Sources are the vendor's pricing pages, documentation and marketing, and independent third parties. Analyst research and one vendor's claims about another are not included.
- Workloads are illustrative, not measurements of any real system.
- Totals are list-price estimates for a standard deployment, not quotes.
The formula language in costings.json is arithmetic over the workload variables in taxonomy.json: numbers, names, + - * /, parentheses, and max, min and ceil.