Oracle Exadata Exascale: Bye-Bye ASM?

This is from my memories of Oracle CloudWorld 2023. I attended various sessions and this is one of the many topics that stuck so wanted to share what I learned. On September 20th, 2023, Oracle introduced its next-generation Exadata System Software – Exadata Exascale. It was a preview session so, I just have my notes to share. 😊

What is Exadata Exascale?

Exadata Exascale enables database instances to adapt and scale seamlessly, catering to a broad spectrum of needs, from small configurations with just a handful of cores and a few gigabytes of storage to large deployments spanning thousands of cores and petabytes of intelligent storage. Moreover, it has database-optimized, storage-efficient snapshot and cloning capabilities, facilitating agile development and testing for applications of any size, ensuring adaptability and efficiency at every level. Remember our beloved ASM (Automatic Storage Management)? From what I understand, it is the replacement of ASM.

ASM serves as kind of a volume manager and file system for Oracle database files. It simplifies storage management and helps with disk configuration, mirroring, and striping for database files. Exascale focuses on database storage efficiency, performance, and elasticity, designed to enhance the Oracle Exadata platform. While ASM is compatible with a wide range of Oracle database versions and can be used in various storage configurations, Exascale is only available from Oracle Database 23c.

Key Features of Oracle Exadata Exascale

1. Cloud Capabilities

Oracle Exadata Exascale incorporates essential cloud features such as multitenancy, resource pooling, and a pay-as-you-go model. This combination allows businesses to experience the full potential of Exadata while benefiting from cloud-like flexibility and cost-efficiency. It’s an ideal solution for both on-premises and Oracle Cloud deployments, making it accessible to a wide range of customers, from small businesses to large enterprises.

2. State-of-the-Art Storage-Efficient Database Cloning

One of the standout features of Exadata Exascale is its advanced database cloning capabilities. This feature not only increases developer productivity but also reduces costs. Thin cloning allows for efficient snapshotting and cloning of pluggable databases within the same container database or across different container databases, see below examples:

Create pluggable database pdb1clone from pdb1 snapshot copy;

create pluggable database pdb1clone from pdb1@cdb1_link snapshot copy;

No more sparse clones: Eliminates the need of having a test master or even the sparse diskgroup.

Snapshot Carousel: Exascale allows you to take new snapshots of pluggable databases at regular intervals. This feature enhances data management and simplifies the process of recovering from errors or rolling back to previous states.

alter pluggable database pdb1 snapshot mode every 3 hours;

3. Elastic Infrastructure

Exadata Exascale offers a highly elastic database infrastructure, which can adapt to varying workloads and resource demands. It enables resource isolation and network security for different tenants, ensuring that your data remains secure and isolated. This architecture can pool multiple tenant clusters and databases for maximum efficiency.

4. Hybrid Deployment

Oracle Exadata Exascale is designed to work seamlessly in both on-premises and cloud environments. It’s essential to note that it requires Oracle Database 23c for optimal performance. This flexibility ensures that businesses can choose the deployment model that best suits their needs.

Combining the Best of Exadata and the Cloud

Oracle Exadata Exascale combines the strengths of both Exadata and cloud technologies, offering the following:

• Database Intelligent Storage, which optimizes data access and retrieval.
• Fast RDMA (Remote Direct Memory Access) for high-speed data transfer.
• Disk, Flash, and DRAM auto-tiering for efficient data storage and retrieval.
• Multitenancy for sharing resources securely.
• Pooled resources to ensure efficient resource utilization.
• Hyper-elasticity, allowing the infrastructure to scale as needed.

Cloud Automation for Exadata and Compute

Exadata Exascale simplifies cloud automation:

• Physical servers are pooled and hidden from tenants. Tenants only interact with databases and VM clusters.
• Customers retain root access to VMs for greater control.
• Future support for live VM migration is planned to minimize disruptions during cloud infrastructure updates.

Cloud Automation for Storage

Exadata Exascale introduces automated storage allocation:

• Allocate and pay only for the storage capacity needed for your databases.
• Virtualization is fully automated through the cloud control plane, removing limitations on local drives and enabling rapid migration.
• Linux VM storage now uses network-attached Exascale storage, eliminating the need for multiple storage tiers, reducing latency and bottlenecks.

Exascale Storage Architecture

The Exascale storage architecture differs from traditional cloud storage solutions:

1. Each client database caches the mapping table, allowing it to send IOs directly to storage servers.
2. This direct IO architecture, with no intermediary nodes, is faster and required for RDMA.

Can Exadata Exascale and ASM co-exist?

Based on my discussions, it appears that both ASM and Exascale can co-exist while you upgrade from the existing database version to 23c. I think you would no longer need ASM once you start using Exadata Exascale completely.

That is all I could re-collect from the “preview” session and can’t wait to learn more about it and test it in action!