Oracle Autonomous Database is the managed database service on Oracle Cloud Infrastructure. The pricing model is consumption based and is built from OCPU consumption, storage consumption, and tier selection. The model is fundamentally different from the perpetual processor model that applied to on premises Oracle Database. The cost can be lower or higher than the equivalent on premises deployment depending on the utilisation pattern and the negotiation outcome. This article describes the pricing model, the cost drivers, the negotiation levers, and the buyer assessment pattern.
The pricing components.
The Autonomous Database pricing has three primary components. The OCPU component is the compute consumption measured in OCPU hours. The storage component is the data and backup consumption measured in storage hours. The tier component sets the feature set and influences the per OCPU rate. The three components combine into the monthly invoice.
The compute consumption is the dominant cost driver in most deployments. The storage consumption is a smaller cost driver in most deployments but can become significant in data heavy workloads. The tier component is a configuration choice that affects every OCPU hour in the deployment. Buyers should understand the three components separately before assessing the total cost.
The OCPU consumption.
The OCPU is the Oracle compute unit on OCI. One OCPU is approximately equivalent to one physical core of a recent Intel Xeon processor. The Autonomous Database service provisions OCPUs in increments that depend on the tier. The provisioning can be fixed at a defined OCPU count or can be elastic with auto scaling between a defined minimum and a defined maximum.
The auto scaling feature is the principal value driver in the Autonomous Database service. The feature allows the OCPU consumption to track the workload demand and reduces the over provisioning that is common in fixed allocation deployments. The auto scaling feature also produces a billing pattern that is difficult to forecast without operational data. Buyers should test the auto scaling behaviour on the actual workload before committing to a long term contract.
The storage consumption.
The storage consumption is billed per terabyte per month and covers the data storage, the backup storage, and the long term backup retention. The data storage is the live database storage. The backup storage is the automated backup retention typically at sixty days. The long term backup retention is the optional retention beyond sixty days at a different rate.
The storage consumption is often overlooked in the initial cost estimate. The actual storage consumption can be substantially higher than the database size because of the backup retention and the redo log retention. Buyers should request a worked storage estimate from Oracle that includes the backup retention and should reconcile the estimate against the actual storage consumption on the existing on premises database.
The tier selection.
The Autonomous Database service is offered in two principal workload tiers. The Autonomous Transaction Processing tier is positioned for transactional workloads. The Autonomous Data Warehouse tier is positioned for analytical workloads. The two tiers carry different feature sets and different per OCPU rates. The tier selection is therefore a substantive commercial decision rather than a technical preference.
The tier selection should be driven by the workload characteristics and tested on representative data. A workload that has a mix of transactional and analytical characteristics may not fit cleanly into either tier. The Oracle account team will recommend the tier that produces the higher revenue. The buyer position should be derived from the actual workload measurement rather than from the Oracle recommendation.
The BYOL option.
The Bring Your Own License option allows buyers to apply existing Oracle Database licences against the Autonomous Database service. The BYOL rate is approximately one third of the License Included rate. The BYOL option produces a substantial cost reduction for buyers with existing perpetual licences and active support coverage.
The BYOL option also extends the value of the existing perpetual licence portfolio into the cloud period rather than stranding the portfolio at the on premises retirement. The BYOL option is therefore an important consideration in any cloud migration assessment for an Oracle estate. See the cloud data egress note and the BYOL deal type page for the related pattern.
The commitment discount.
The standard Autonomous Database pricing is consumption based with no commitment. Oracle offers commitment based pricing through the Universal Credits program which provides a discount in exchange for a multi year commitment with a defined minimum consumption. The discount is typically in the range of fifteen to thirty percent depending on the commitment size and term.
The commitment is structured as a draw down arrangement against a prepaid balance. The unused balance at the end of the term is forfeited rather than refunded. The commitment therefore needs to be sized below the realistic consumption forecast rather than at the optimistic consumption forecast. Buyers should also negotiate the right to redirect unused balance to other OCI services if the actual consumption pattern differs from the forecast.
Engaging an independent advisor.
The Autonomous Database commercial benefits from external assessment. An independent advisor can compare the Oracle proposal against the on premises baseline, against the BYOL economics, and against alternative cloud database options. The independent assessment supports the negotiation position and produces a defensible business case for the cloud decision.
For the wider cluster see Cloud Negotiation. For the service see Cloud Migration Advisory. For the deal structure see OCI Universal Credits. For the Oracle product see Oracle Database. For the full research read the Oracle Negotiation Playbook.
A worked example.
A North American healthcare buyer was assessing the migration of an Oracle Database workload to the Autonomous Database service. The Oracle proposal sized the service at sixteen OCPU License Included with twenty four terabytes of storage at an annual run rate of approximately five hundred thousand dollars.
An independent advisor reviewed the buyer existing licence portfolio and identified sufficient perpetual licences with active support to cover the BYOL rate at the proposed size. The corrected commercial position was BYOL at sixteen OCPU with the same storage at an annual run rate of approximately two hundred and twenty thousand dollars. The Universal Credits commitment was also restructured to a one year term rather than the three year term in the original proposal to reduce the forfeiture risk on unused balance. The total saving against the original proposal was approximately three hundred thousand dollars per year over the term.