FlashGrid Cluster vs Oracle Database@AWS
Introduction
FlashGrid Cluster and the Oracle Database @ AWS service are two distinct solutions that enable AWS cloud customers to run Oracle Databases. Each has advantages in different use cases. The comparison table below helps AWS customers select the most appropriate one for their requirements.
Disclaimer
Due to the lack of publicly available Oracle Database @ AWS documentation as of 2024-09-24, some of the provided information is based on the documentation for Exadata Database Service on Dedicated Infrastructure, a similar service hosted in Oracle Cloud.
The Oracle Database @ AWS information is provided for reference only and may contain errors.
Comparison
| FlashGrid Cluster | Oracle Database@AWS | |
|---|---|---|
| Key advantages | Optimal for mission-critical transaction processing with database uptime SLA of 99.99%, 99.999%, or higher. Maximum uptime with Multi-AZ clustering. Maximum data durability with all-flash storage and two levels of mirroring. Agility (based on AWS EC2). Proven and available in all AWS regions since 2017. | Optimal for extra-large (100+ TB) DWH workloads. Exadata proprietary DWH features (Smart Scan, HCC) Cheaper storage capacity with spinning HDD. |
| Solution type | Virtual clustered appliance hosted on native AWS EC2 instances and EBS storage volumes | Virtual clustered appliance hosted on dedicated Exadata infrastructure |
| Who manages VMs, OS, and databases? | Customer | Customer |
| Oracle RAC option | Yes | Yes |
| Oracle Failover HA option | Yes (available for single instance databases) | Yes (requires RAC One Node) |
| Oracle Enterprise Edition | Yes | Yes |
| Oracle Standard Edition (including SEHA) | Yes | No |
| Database node software updates | Controlled by customer. (including DB, GI, OS, FlashGrid software) | Controlled by customer. (including DB, GI, OS software) |
| Infrastructure software updates | AWS infrastructure software updates are managed by AWS and are fully transparent. | Exadata infrastructure software updates are managed by Oracle. Exadata infrastructure software updates may require patching or upgrading GI and/or DB software. See footnote 1 in Oracle Release Schedule of Current Database Releases (Doc ID 742060.1) |
| Failed hardware replacement | AWS automatically moves affected VMs to new physical hosts. AWS automatically and transparently handles replacement of failed disk storage hardware. | Storage server, HDD, and SSD hardware is exposed to the customer. Customer may need to take certain actions when the hardware fails and needs to be replaced. |
| Data center fault tolerance | Multi-AZ clustering protects against data center failures. | No Multi-AZ clustering. Databases exposed to data center failures. Must rely on data replication to a different AZ or region. Some functionality (control plane, backups, DR replication, etc.) depends on OCI parent region availability. |
| Storage reliability and durability | All-flash storage with AWS EBS volumes. Additional 2-way or 3-way mirroring across availability zones. Durability: * >99.9996% with 2-way mirroring * >99.9999992% with 3-way mirroring | Spinning disks have lower durability than all-flash storage. No durability specifications available. 3-way mirroring within a single server rack. |
| Storage capacity and cost | Flexible capacity from 1 to 1000+ TB usable. Higher cost per TB due to the exclusive use of flash storage for consistent performance and two levels (EBS and Oracle ASM) of mirroring. | Minimum of 190 TB usable (assumes Quarter-rack Exadata). Lower cost per TB due to using spinning HDD. Lower cost per TB when using HCC (for DWH only). |
| Agility | Instant on-demand deployment, usage, and resizing. Based on native AWS EC2 instances and EBS storage volumes. | Must reserve Exadata hardware resources – database servers and storage servers. It is not possible to reduce the reserved storage capacity by removing storage servers from an existing infrastructure instance. |
| Maturity | In production use since 2017 | Preview starting in December 2024 |
| Available in AWS Regions | All (including Gov Cloud, China) | Limited |
| Database versions 11.2, 12.1, 12.2 | Available | Existing information is unclear. Upgrade Support may be required |
| Customer Cloud infra skills required for maintenance | AWS skills only | Oracle Cloud and AWS skills required |
| Oracle licensing | BYOL | BYOL or PAYG |
| Billing | By AWS (infra + Marketplace fees) | By AWS |
Conclusions
AWS customers who require true cloud agility and/or maximum uptime SLA for mission-critical Oracle Database transaction processing are best served by FlashGrid Cluster because:
- It supports multi-AZ configuration and utilizes highly durable AWS EBS storage.
- It is built from native AWS EC2 instances and EBS storage volumes available in all AWS regions.
- It is a mature solution that has been running production Oracle workloads on AWS since 2017.
Customers with large data warehouse databases exceeding 100 TB and the required Oracle Cloud expertise in-house may benefit from Oracle Database @ AWS.
Learn more
For more information, please contact FlashGrid at info@flashgrid.io
