Oracle OCI MySQL HeatWave: Database-as-a-Service Pricing & Compliance Guide 2026

MySQL HeatWave Architecture: Components and Tiers

Oracle MySQL HeatWave on OCI is structured as three distinct layers, each with separate pricing and capability scope. Understanding the architectural separation is essential for accurate cost modelling.

The first layer is the MySQL Database Service (MDS) — a fully managed MySQL 8.x instance (Standalone, HA, or Read Replicas) providing standard OLTP database capabilities. MDS is priced per OCPU and storage, with no separate Oracle MySQL license fee for the managed service — the MySQL license cost is bundled into the OCI consumption pricing. The second layer is the HeatWave Cluster — an in-memory analytics accelerator that runs as a separate cluster of nodes attached to the MySQL DB system, enabling real-time analytics queries (OLAP) directly against transactional MySQL data without ETL or data movement. The HeatWave cluster is an optional add-on charged per node. The third layer includes HeatWave ML (machine learning inside MySQL) and HeatWave GenAI (natural language database queries, document analysis), which add further capabilities at additional cost per OCPU.

The commercial architecture is deliberately modular — Oracle allows customers to start with a basic MDS deployment and add HeatWave capabilities incrementally. The risk is that each layer addition increases the monthly OCI bill in ways that are not always apparent in the initial architecture review. A MySQL HeatWave deployment designed to replace both a transactional database and a separate analytics platform can cost 3–5× the base MDS cost once the HeatWave cluster and ML/GenAI layers are added.

Pricing Model: OCPU-Based DBaaS Consumption

MySQL Database Service (MDS) on OCI is priced on an hourly OCPU consumption model, with separate charges for provisioned storage. Unlike Oracle Database licensing — where the license cost is separate from infrastructure — MySQL HeatWave bundles the MySQL license into the OCI service price. This means enterprises paying for MySQL HeatWave on OCI do not separately pay Oracle MySQL Enterprise Edition commercial license fees for the managed service instances.

Storage pricing is separate from compute and accrues for all provisioned storage — including hot standby nodes in HA configurations. An HA MySQL HeatWave deployment stores data on three nodes (the primary plus two standby replicas), so effective storage cost is 3× the provisioned data volume. This multiplier is frequently missed in initial cost estimates.

Oracle OCI Universal Credits can be applied against MySQL HeatWave consumption, which is one reason Oracle encourages enterprises to commit to Universal Credits for their overall OCI usage — the flexibility allows MySQL HeatWave consumption to draw from a single committed credit pool alongside other OCI services. For enterprises negotiating OCI Universal Credits commitments, MySQL HeatWave cost should be included in the total OCI consumption model. Our Cloud Advisory team builds these models independently before enterprises enter negotiations with Oracle.

HeatWave Cluster: Analytics Accelerator Pricing

The HeatWave Cluster is the defining differentiation of Oracle MySQL HeatWave — it enables analytics queries to run directly against MySQL data stored in memory across a cluster of dedicated HeatWave nodes, without requiring data movement to a separate analytics database or data warehouse. Oracle's benchmarking positions HeatWave as significantly faster than MySQL alone for analytical queries and competitive with or superior to dedicated cloud analytics services for many workload profiles.

HeatWave cluster nodes are priced separately from the base MySQL DB system. Each HeatWave node has a fixed memory allocation (the HeatWave shape's in-memory capacity), and the number of nodes required depends on the total data volume to be loaded into HeatWave memory. Oracle provides a HeatWave Estimator tool that analyses a MySQL schema and estimates the required number of HeatWave nodes for a specific dataset.

A typical production HeatWave deployment for a 1TB analytics dataset requires 2–4 HeatWave nodes (depending on compression ratios). At $3.20/hour per node, a 2-node HeatWave cluster costs approximately $6.40/hour or ~$4,600/month at continuous operation — a material cost addition to the base MDS expense. For workloads that use HeatWave analytics only during business hours, stopping the HeatWave cluster outside peak hours reduces cost — HeatWave nodes can be started and stopped without losing the MySQL DB system, but HeatWave data must be reloaded on restart.

Cost Comparison: OCI HeatWave vs AWS RDS MySQL vs Aurora MySQL

Oracle's marketing positions MySQL HeatWave as a cost-effective replacement for a MySQL database plus a separate analytics service (such as Amazon Redshift or Snowflake). The cost comparison depends heavily on whether the HeatWave analytics capability is genuinely a workload requirement — if the workload is purely OLTP, HeatWave's cost premium over AWS RDS MySQL is not justified by capability.

The headline finding: at the base compute layer (OLTP only), OCI MySQL HeatWave and AWS RDS MySQL are price-competitive. HeatWave's cost advantage is specific to workloads that genuinely use the HeatWave analytics acceleration — replacing a separate analytics database subscription with HeatWave can represent net cost reduction if the analytics database (Redshift, Snowflake, BigQuery) costs more than the HeatWave cluster add-on. For pure OLTP MySQL workloads, HeatWave offers no cost advantage over AWS RDS MySQL or Google Cloud SQL.

Oracle MySQL Licensing: When On-Premise Obligations Apply

MySQL HeatWave on OCI is a managed service — Oracle bundles MySQL license costs into the consumption pricing, and enterprises do not separately license MySQL for managed OCI deployments. The Oracle MySQL licensing question becomes relevant in two scenarios: organizations with on-premise MySQL Enterprise Edition deployments, and organizations considering BYOL MySQL on OCI compute (self-managed MySQL on OCI VMs rather than using the managed HeatWave service).

Oracle's MySQL Community Edition (open-source) carries no license obligations for any deployment — on-premise, cloud, or OCI. MySQL Community is GPL-licensed and does not require Oracle commercial license fees. MySQL Enterprise Edition (commercial) requires Oracle license fees for on-premise deployment at approximately $2,000–$5,000 per server/year depending on subscription tier, but is bundled into MySQL HeatWave managed service pricing on OCI.

For enterprises running MySQL Community Edition on-premise and considering OCI MySQL HeatWave as a cloud migration target: the managed service model does not create any new Oracle MySQL license obligations. You pay OCI consumption rates, which include Oracle's MySQL license cost for the managed service, and your on-premise MySQL Community deployments remain license-free.

For enterprises with existing MySQL Enterprise Edition subscriptions on-premise: Oracle does not offer a BYOL model for MySQL Enterprise Edition on OCI managed services (unlike Oracle Database BYOL). If you have existing MySQL EE subscriptions, they cannot be applied to reduce OCI MySQL HeatWave consumption costs in the same way that Oracle Database BYOL reduces OCI compute costs. The MySQL EE subscription covers on-premise deployment only. This is a meaningful commercial consideration for enterprises paying MySQL Enterprise subscriptions who are planning OCI migrations — the two cost streams do not offset each other.

HeatWave ML and HeatWave GenAI: Additional Cost Considerations

Oracle has extended MySQL HeatWave with in-database machine learning (HeatWave ML) and generative AI capabilities (HeatWave GenAI). Both features require the HeatWave Cluster to be active and carry additional per-OCPU charges above the base HeatWave node cost.

HeatWave ML enables training and running ML models inside MySQL HeatWave without data movement to external ML platforms. Oracle's positioning is that this eliminates the cost and latency of extracting data to separate ML infrastructure. For organizations with moderate ML requirements (classification, forecasting, anomaly detection on structured MySQL data), HeatWave ML can reduce the need for separate ML platform subscriptions — but the cost comparison must be made against the actual ML platform alternative (AWS SageMaker, Google Vertex AI) at similar scale.

HeatWave GenAI enables natural language SQL queries (using vector embeddings and in-database retrieval-augmented generation) and document processing capabilities. As with HeatWave ML, the cost impact depends on usage volume and the alternative architecture being displaced. For enterprises evaluating HeatWave GenAI as a replacement for a separate LLM API integration with their MySQL data, the in-database approach eliminates data egress costs and reduces latency — but the HeatWave compute cost for GenAI workloads should be modelled against the actual inference volumes expected.

Cost Optimization Strategies for MySQL HeatWave

Enterprises deploying MySQL HeatWave on OCI have several proven approaches to reduce total service cost without sacrificing database capability or availability.

• Stop/start HeatWave clusters for scheduled workloads: If analytics queries are batch-oriented (e.g., overnight reporting runs), stop the HeatWave cluster outside analytics windows and restart before the batch begins. Account for data reload time in the schedule. This can reduce HeatWave node hours by 60–70% for batch-only analytics patterns.
• Right-size the MySQL DB system shape: Start with the minimum shape required for OLTP workload performance and scale up only when CPU or memory utilization consistently reaches thresholds. OCI MySQL HeatWave supports online shape changes without service interruption.
• Use OCI Reserved Capacity for base MDS instances: Committing to 1-year or 3-year reserved capacity for MySQL DB system shapes reduces hourly rates by 36–60% versus on-demand pricing. For production MySQL instances with stable workloads, reserved capacity delivers significant monthly savings.
• Evaluate HA vs. Read Replicas: High Availability configurations provide synchronous multi-node replication but cost approximately 3× the base MDS price (three nodes active). For workloads where eventual consistency is acceptable, a primary instance with Read Replicas may deliver similar read scalability at lower cost than HA.
• Include MySQL HeatWave in OCI Universal Credits negotiations: Universal Credits committed spend can be applied against MySQL HeatWave consumption. Negotiating a Universal Credits commitment that includes MySQL HeatWave alongside Oracle Database and OCI compute creates leverage for improved overall OCI discount rates. Our OCI pricing negotiation guide covers this in detail.