An Exadata rack arrives configured to perform — and configured, by default, to use options that Enterprise Edition customers pay extra for. Exadata is still licensed by processor: physical cores × the 0.5 core factor, multiplied by the same option and pack prices as any other server. The difference is that a fully populated rack carries dozens of licensable processors and ships with features like Hybrid Columnar Compression and Smart Scan that pull in Advanced Compression and other options. This sub-guide of the Oracle Database options guide maps where engineered-system licences hide.
Exadata hardware is bought or leased separately, but the Oracle Database running on it is licensed exactly as it would be on any server: by processor, where processors = enabled physical cores × core factor (0.5 for the Intel Xeon cores in Exadata database servers). Database Enterprise Edition is mandatory, and most Exadata value comes from options — Partitioning, Advanced Compression (for Hybrid Columnar Compression), RAC, and often Multitenant — each licensed on the same processor count. Exadata also offers a capacity-on-demand model that lets you enable a subset of cores and licence only those, which is the single most important cost lever on the platform.
Because engineered systems pack many cores into one rack. A populated rack can present dozens of enabled cores per database server, and licensing every core is rarely necessary on day one. Capacity-on-demand (CoD) lets you enable cores in increments and licence only the enabled set — provided you stay within Oracle's minimum thresholds. The table illustrates the gap between a fully enabled rack and a CoD-managed one.
| Scenario (illustrative 2-server rack) | Enabled cores | Processors (×0.5) | EE licence at $47,500 |
|---|---|---|---|
| All cores enabled | 64 | 32 | $1,520,000 |
| Capacity-on-demand, half enabled | 32 | 16 | $760,000 |
| Capacity-on-demand, minimum start | 16 | 8 | $380,000 |
Those are Enterprise Edition figures alone. Add the options the platform encourages and the per-processor cost roughly doubles, which is why the enabled-core decision compounds across every line. The same core-factor mechanics govern non-engineered servers too — see Oracle processor licensing for the full method.
We model capacity-on-demand against options exposure before you sign.
Exadata's headline features map directly to chargeable options. Smart Scan and Storage Indexes are included with the platform, but the database features customers actually configure to exploit the rack are not. Hybrid Columnar Compression (HCC) is the clearest example: it is an Exadata storage feature, but using it on the database side activates Advanced Compression accounting in many configurations. Treat the rack as an accelerant for options exposure, not a shield from it.
| Exadata feature | Licence it touches | Per-processor list |
|---|---|---|
| Partitioned tables (near-universal on Exadata) | Partitioning | $11,500 |
| Hybrid Columnar Compression workflows | Advanced Compression | $11,500 |
| Clustered database across servers | RAC | $23,000 |
| Consolidation onto pluggable DBs | Multitenant | $17,500 |
| In-memory analytics on the rack | In-Memory | $23,000 |
| Performance tuning via AWR/Advisors | Diagnostic + Tuning Pack | $12,500 |
The detection mechanics are identical to any other database: these uses land in DBA_FEATURE_USAGE_STATISTICS, and Oracle's LMS scripts read them the same way. Our feature-usage audit guide covers the exact queries, and options & management packs lists every trigger.
Includes an engineered-systems annex: capacity-on-demand, HCC, and the options that ride the rack.
It changes the commercial wrapper, not the underlying exposure. Exadata Cloud@Customer (ExaCC) puts an Oracle-managed Exadata in your data centre on a subscription, with options included in the service metric rather than licensed separately. That sounds cleaner, but two traps recur: bring-your-own-licence (BYOL) positions that double-count entitlements you already own, and OCPU scaling that quietly increases the committed spend. Before moving to ExaCC, reconcile what you already own against what the subscription assumes — the same BYOL discipline we apply across Oracle audit defence engagements.
| Model | How options are paid | Main risk |
|---|---|---|
| On-prem Exadata (licence) | Per-processor licence + 22% support | Capacity-on-demand mis-sizing; silent options |
| ExaCC — BYOL | Bring existing licences; pay infra subscription | Double-counting owned entitlements |
| ExaCC — licence-included | Options bundled in OCPU rate | OCPU scaling inflates committed spend |
It has four parts. First, an enabled-core inventory per database server with the capacity-on-demand configuration documented, so the licensable processor count is defensible. Second, an option-by-option map of which rack features are in use and which licence each one pulls in. Third, the same DBA_FEATURE_USAGE_STATISTICS review you run on every database, scheduled quarterly. Fourth, for ExaCC, a reconciliation of owned entitlements against the subscription's assumptions. Get those four right and an Exadata audit becomes a confirmation exercise rather than a discovery one. Where the enabled-core count still outruns the real workload, our Oracle license cost reduction work resets the licensable baseline before the rack reaches renewal.
We map capacity-on-demand, options exposure, and Cloud@Customer economics so the licence matches the workload. $1.8B+ documented savings · 68% average audit-claim reduction · buyer-side only since 2016.
Weekly compliance intelligence for IT leaders.