Exadata Cloud@Customer

Exadata Cloud@Customer X10M

Distributed cloud database platform

Exadata Cloud@Customer X10M brings the cloud automation and consumption pricing of Exadata Database Service and the fully managed Autonomous Database in Oracle Cloud Infrastructure (OCI) to an organization’s data centers. Owned and remotely operated by Oracle, it lets organizations easily start using cloud resources for crucial applications, connect to current on-premises systems and data sources, and meet data residency requirements.

High performance architecture

Exadata Cloud@Customer X10M is based on a unique scale-out architecture that integrates fourth-generation AMD EPYC processors in database servers and intelligent storage servers to provide more processing cores and higher per-core performance than previous generation platforms. Intelligent storage servers offload SQL processing from database servers to enable faster SQL I/O while Exadata RDMA Memory reduces SQL read I/O latency to less than 17 microseconds.

High transactional performance

OLTP applications run faster with Exadata’s unique, RDMA-enabled scale-out architecture. Storage servers automatically cache the most frequently used data in the extremely low latency Exadata RDMA Memory Data Accelerator and large, ultrafast NVMe flash cache to boost overall system performance. Exadata Cloud@Customer X10M starts with 380 available processor cores, 5.6M SQL low-latency read I/O operations per second, and nearly 2 TB/sec of memory bandwidth in quarter-rack configurations that can be expanded to meet the transaction processing needs of virtually any organization.

Accelerated analytics and machine learning

Analytics and in-database machine learning (ML) workloads run faster by eliminating I/O operations and offloading analytic SQL queries and ML algorithms to storage servers. Extreme analytics throughput, parallel processing across all storage servers, and unique compression algorithms accelerate data warehousing workloads of any size. Exadata Cloud@Customer X10M starts with 135 GB/sec of scan throughput, 190 TB of uncompressed database capacity, and the ability to hold up to 1.9 PB of data warehouses using Oracle Hybrid Columnar Compression in a quarter rack. Its scale-out architecture lets you use up to 2,880 GB/sec of scan throughput, 4 PB of uncompressed databases, and 40 PB of Hybrid Columnar Compression compressed data warehouse capacity for data-intensive workloads.

Independent database compute and storage scaling

Database servers provide 190 usable processor cores and up to 2,800 GB of memory. Storage servers provide 64 processor cores and 80 TB of usable storage. Database and storage resources can be scaled independently by adding individual servers, up to a total of 16 servers in the initial rack and up to 32 database servers and 64 storage servers across multiple racks. Optimizations to the entire software stack ensure that database performance scales linearly when running on the Exadata X10M servers with a high core count and large memory capacity.

Greater database consolidation

Customers can consolidate even more databases in a single Exadata Cloud@Customer X10M than with the previous generation platform, providing greater management efficiency and lowering costs. Organizations can concurrently run the automated Exadata Database Service and fully managed Autonomous Database on the same Exadata Cloud@Customer infrastructure, letting them efficiently consolidate all types of workloads.

Learn more

Exadata performance, scale, and availability

High transactional performance

Business-critical OLTP applications run faster with Exadata’s unique scale-out architecture. A starter system provides 190 processor cores and 2 TB/sec of memory bandwidth in database servers, 17 microsecond SQL read IO latency, and 16x compute scalability to meet the needs of virtually any organization.

Accelerate analytics and machine learning

Organizations run analytics and in-database machine learning applications faster by processing low-level SQL queries, analytics, and machine learning (ML) algorithms in intelligent storage servers. A starter system provides 192 processing cores for SQL processing in storage servers, 135 GB/sec of scan analytics throughput, and storage for 192TB of uncompressed databases or 1.9PB of compressed data warehouses – with 21x additional storage scalability.

Greater database consolidation

All types of Autonomous Database and Exadata Database workloads can be consolidated on a single set of Exadata Cloud@Customer infrastructure, reducing infrastructure and management costs while maintaining operational isolation.

Oracle Real Application Clusters (Oracle RAC)

Autonomous Database Service and Exadata Database Service use Oracle RAC to enable online performance scaling and high availability for mission-critical database workloads. Configure multiple clusters on a single system and manage them separately to support diverse workload and organizational needs.

Oracle Maximum Availability Architecture (Oracle MAA)

Oracle MAA best practices for protecting mission-critical Oracle databases against outages and data loss are easily accessible on Exadata Cloud@Customer from the OCI console and are built into the Autonomous Database Service.

Learn more

Oracle Autonomous Database

Automated tuning and scaling

Oracle Database workloads are automatically analyzed, tuned, and scaled to optimize performance for OLTP, analytics, and application development workloads without interrupting their operations.

Pay-per-use economics

Autonomous Database helps customers reduce costs by only charging for actual consumption, instead of anticipated consumption. Online compute scaling based on workloads’ immediate needs helps increase performance during spikes in demand and lower costs when there’s less demand.

Automated provisioning

Developers are more productive with a self-service database cloud that lets them quickly provision their own databases whenever needed.

Automated fault management and patching

Protect crucial customer databases from downtime due to failures, maintenance, user errors, and cyberattacks with automatic application of the latest software updates and security patches.

Increased staff efficiency

By automating everything, including performance monitoring and data protection, Autonomous Database enables scarce IT resources to focus on innovation and business requirements instead of manual operations.

Customizable environments

Organizations can optimize support for multiple workloads by using separate VM clusters on the same service for different workloads and different database lifecycle stages. Development, staging, and production VM clusters can have different configurations, access rules, quotas, and licensing (Oracle Bring Your Own License or license included) to optimize performance and cost.

Learn more

Oracle Database Cloud

Converged database

Organizations can use Autonomous Database and Exadata Database Service with virtually any type of data, any workload, and any development style. By standardizing on a single database, they can eliminate data silos and the management and security fragmentation caused by multiple, single-purpose databases.

Modern application development

Developers can create modern applications with any type of data, for any workload, and with their choice of development models. Applications automatically benefit from high performance, consistent security models, and mission-critical capabilities so developers can focus more on innovation and less on integration. Built-in tools, such as the Oracle APEX low-code development environment, further increase developer productivity and shorten development lifecycles.

Consolidated database management

IT teams operate more efficiently because they can manage hundreds of databases at the same time using Oracle Multitenant capabilities and consolidate thousands of OLTP, analytics, and mixed database workloads on a single Exadata Cloud@Customer.

Multiple database versions

Organizations simplify moving to the cloud with the ability to use multiple Oracle Database versions—all on the same Exadata Cloud@Customer.

In-database machine learning

Organizations can easily add ML capabilities to Oracle Database applications by eliminating the data extraction, reformatting, and movement required by standalone ML services. In-database ML algorithms are accelerated with high throughput intelligent storage servers.

Oracle Real Application Clusters (RAC)

Autonomous Database and Exadata Database Service use Oracle RAC to enable online performance scaling and high availability for mission-critical database workloads. Configure multiple VM clusters on a single system and manage them separately to support diverse workload and organizational needs.

Learn more

Security and resiliency

Defense in-depth

OCI helps protect crucial Oracle databases with multiple levels of security to identify and respond to threats in real time. Exadata Cloud@Customer includes the Oracle Data Safe database security console at no extra charge, letting customers monitor and manage their Oracle Database security posture across their database fleet.

Operational isolation

Exadata Cloud@Customer lets Exadata Database Service and Autonomous Database run on multiple isolated VM clusters with separate network configurations to improve security postures in customer data centers.

End-to-end encryption

Private data in Oracle databases is protected with always-on encryption for data at rest, in motion, and backed up with customer-owned encryption keys.

Operator Access Control

Oracle Operator Access Control increases security by letting organizations fully supervise what Exadata Cloud@Customer resources that Oracle Cloud operators can access and what actions they can take during remote management sessions.

Full-stack patching by Oracle

Oracle creates full-stack patch sets for Exadata Cloud@Customer and coordinates their rapid deployment with enterprises’ IT teams to reduce threats and risks without interrupting business.

Oracle Maximum Availability Architecture

Oracle Maximum Availability Architecture best practices for protecting Oracle databases against outages and data loss are easily accessible on Exadata Cloud@Customer from the OCI console and are built into Autonomous Database.

Learn more

High availability

Oracle Maximum Availability Architecture

Maximize the availability and protection of vitally important Oracle databases using proven Oracle Maximum Availability Architecture best practices.

No single point of failure

Enable crucial customer Oracle databases to operate 24/7 across failures with Exadata’s full server, storage, and networking redundancy.

Proactive fault management

Autonomous Database on Exadata Cloud@Customer helps increase availability for customers’ Oracle databases by using AI to detect and remediate potential problems before users notice.

Oracle Real Application Clusters (Oracle RAC)

Oracle Exadata Cloud@Customer enables local performance scaling and high availability for important customer databases with Oracle RAC.

Oracle Active Data Guard

For business continuity and disaster recovery, Active Data Guard maintains both synchronized standby and replica Oracle databases in customers’ data centers or the cloud.

Learn more

Lower costs

Automate maintenance operations with Exadata Fleet Update

For large-scale database deployments, Exadata Fleet Update automates maintenance operations by grouping databases and grid infrastructures in a single maintenance schedule and cycle.

Lower consumption costs

Consolidating databases on Exadata Cloud@Customer and automatically scaling consumption with Autonomous Database lets organizations run their database fleets on 65% to 75% less infrastructure and costs than other hybrid cloud database services, as reported by Wikibon.

Scalable consumption

Organizations only pay for the CPU cores that database services consume using online scaling of Exadata Database Service consumption and fully automatic scaling with Autonomous Database. CPU core scaling is performed online to enable continuous operations.

Zero consumption databases

Autonomous Database on Exadata Cloud@Customer lets organizations create persistent databases with zero consumption costs at the time of creation or when consumption is set to zero.

More efficient database consolidation

Exadata Cloud@Customer X10M provides high compute and storage density, letting organizations consolidate their Oracle Database workloads onto smaller configurations that cost less and use less power and data center space.

Less management

Consolidating databases onto Exadata Cloud@Customer with Oracle managed infrastructure improves management efficiency and infrastructure utilization. Automating database management with Autonomous Database reduces database administration, freeing up resources to accelerate innovation.

Software features

Concurrently run multiple database services

Autonomous Database and Exadata Database Service concurrently run in isolated VM clusters on the same Exadata Cloud@Customer infrastructure, making it easy for customers to use a fully autonomous database service with no up-front costs.

Smart Scan

Applications processing large amounts of data run faster because SQL queries, analytics, and ML algorithms are offloaded to intelligent storage servers, minimizing data transfers and freeing database server resources for other workloads.

Exadata RDMA Memory optimization

An organization’s crucial transactional applications run faster because database servers directly access Exadata RDMA Memory in shared storage servers with no operating system or networking overhead, reducing read latencies to less than 17 microseconds.

Automated resource management

Built-in I/O resource management and ML-based automated indexing and tuning of databases let them run at peak speed without DBA management.

VM cluster customization

Multiple VM clusters for Exadata Database Service and Autonomous Database can be created with customized mixtures of compute and storage resources to support different workloads and operational isolation requirements for crucial workloads.

Cloud control plane

Users securely provision and manage databases through a secure connection to the OCI control plane using an encrypted web browser, command-line interface, or REST API.

• Analyst report: Reimagining the data-driven bank (PDF)
• Infographic: Improving healthcare data management with Oracle Exadata (PDF)

Exadata Cloud@Customer Shapes

Exadata Cloud@Customer Base System

Enable 0 to 48 OCPUs with 38 TB of NVMe flash cache, 73 TB of usable storage, and up to 562,500 SQL IOPS.

Exadata Cloud@Customer X9M Quarter Rack

Enable 0 to 124 OCPUs with 4.5 TB of PMEM, 76 TB of NVMe flash cache, and 190 TB of usable storage with up to 5.6 million SQL IOPS and 144 cores of SQL processing power in intelligent storage servers.

Exadata Cloud@Customer X9M Half Rack

Enable 0 to 248 OCPUs with 9 TB of PMEM, 153TB of PCIe 4.0 NVMe flash cache, and 381 TB of usable storage with up to 11.2 million SQL IOPS and 288 cores of SQL processing power in intelligent storage servers.

Exadata Cloud@Customer X9M Full Rack

Enable 0 to 496 OCPUs with 18 TB of PMEM, 307 TB of PCIe 4.0 NVMe flash cache, and 763 TB of usable storage, as well as up to 22.4 million SQL IOPS and 576 cores of SQL processing power in intelligent storage servers.

Exadata Cloud@Customer Storage Servers

Customers may use Exadata Cloud@Customer elastic storage expansion to increase storage capacity and intelligent storage processing capabilities without adding additional compute servers. Organizations with base, quarter, and half-rack systems can expand them up to a total of 12 storage servers.

High-performance networking

PCIe 4.0 dual-port active-active 100Gb/sec RDMA over Converged Ethernet (RoCE) networking provides 80% more throughput between storage and compute servers than to X8M systems, allowing customers’ consolidated workloads to run faster. Communication with application servers and multi-cloud environments is optimized across up to sixteen 25 Gb/sec Ethernet connections.

Cloud control plane servers

Two local servers enable remote Oracle infrastructure management and local operations if networks are unavailable.