Assumptions & Notes

  • Pricing assumptions: ECPUs, BYOL, ATP, list price. No discounts/commits included.
  • Portability is treated as a single criterion (not split). Within Serverless, joining/leaving an Elastic Pool is click-based; moving to Dedicated Exadata is typically a migration exercise.
  • Scores are relative (1–5); importance (0–100) can be set independently per criterion to model stakeholder priorities.
  • For a concrete cost estimate, use the official OCI Cost Estimator.

When to Choose Each Option

Business driver comparison across the five deployment flavors

☁️ Serverless Oracle-managed shared infrastructure
Serverless Standalone pay-per-use
Elastic Pool Pooled compute, shared infra
Ded. Elastic Pool Pooled compute+storage, co-located
🏗️ Dedicated Customer-exclusive Exadata infrastructure
ADB Dedicated Standalone Exadata
EP (Dedicated) Elastic Pool on Exadata

Scenario Analysis

When each flavor makes the most business sense

Decision Matrix

Set importance (0–100) per criterion to find your best fit

How to use this matrix
Each criterion has an Importance slider (0–100) representing how critical that factor is for your scenario — these are independent, not percentages. Each flavor receives a Score (1–5) for how well it meets that criterion. You can click the score dots to adjust them. The weighted result is calculated by normalizing each importance against the total, then multiplying by the score: Weighted Score = Σ (importancei / Σimportances) × scorei
Feature & Capability Matrix Features with differences across flavors or versions — check "Req." to filter Updated: Mar/2026
Flavor Radar — Scores per Criterion
Importance Distribution
Autonomous AI Database — Cost Comparison
BYOL · Transaction Processing · List Price
Estimated Monthly Cost
$0
Database Configuration
Utilization 744 hrs/month
Min 2 per DB
100% — all provisioned ECPUs active
Billing Model Details & Formulas

ATP Serverless

Each DB billed independently for provisioned ECPUs × hours running. Min 2 ECPUs/DB. Auto-scaling bursts up to 3× base.

Compute = Σ ECPUs/DB × $0.0807/hr × 744 h Storage = GB allocated × $0.1156/mo Backup = GB used × $0.0244/mo
Billing Docs

Elastic Pool (Serverless)

Compute billed to pool leader. Members can have 1 ECPU min. Storage billed per-DB.

Compute = Pool shape × tier × $0.0807/hr × 744 h Tiers peak ≤ 1× → 1×; ≤ 2× → 2×; > 2× → 4× Shapes 128, 256, 512, 1024, 2048, 4096
Billing Docs

Dedicated Elastic Pool

Both compute + storage billed to leader. Same tiered billing. Min: 256 ECPUs / 256 TB.

Compute = Pool shape × tier × $0.0807/hr × 744 h Shapes 256, 512, 1024 Storage = Pool total TB × $0.1156/GB/mo
Billing Docs

ADB Dedicated X11M

Compute per provisioned ECPUs + Exadata infra (DB + Storage servers). Backup via Object Storage.

Compute = ECPUs × $0.0807/hr × 744 h DB srvs = max(2, ⌈ECPU÷760⌉) × $2.90/h ST srvs = max(3, ⌈TB÷80⌉) × $2.90/h
Billing Docs

Elastic Pool (Dedicated)

Elastic Pool on Dedicated Exadata. Pool compute billing + Exadata servers. Storage billed per instance.

Compute = Pool shape × tier × $0.0807/hr Tiers same as Serverless EP (1×/2×/4×) Infra = DB + ST servers × $2.90/h
Billing Docs