Oracle Exadata Database Machine is an integrated, full-stack platform for deployment in enterprise data centers that's coengineered with Oracle Database to run all types and sizes of database workloads fast and with high availability. It lets organizations get more value from their data with faster transaction processing, analytics, and in-database machine learning (ML) capabilities that improve customer experiences and business insights. With Exadata Database Machine, organizations can eliminate data center complexity by consolidating Oracle databases on less infrastructure, reducing DBA workloads, and lowering total cost of ownership (TCO).
Oracle'ın yaşamsal veritabanı teknolojilerinden sorumlu Başkan Yardımcısı Juan Loaiza ve özel konuğu AMD CTO'su Mark Papermaster'ın Oracle Exadata X10M inovasyonlarını ele aldığı söyleşiyi izleyin.
Exadata Database Machine 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 higher analytic SQL throughput while Exadata RDMA Memory reduces OLTP SQL read latency to less than 17 microseconds.
OLTP applications run faster with Exadata’s scale-out architecture. Exadata Database Machine X10M lets more transactions be processed in parallel with 3X more database server cores and 2X more storage server cores than the previous generation platform. More processing resources, nearly 1 TB/sec of memory bandwidth per server, and latency of less than 17 microseconds meet the transaction processing needs of many organizations.
Organizations run analytics and in-database ML workloads faster by offloading low-level algorithms to intelligent storage servers. Starting configurations with two database servers and three storage servers provide 135 GB/sec of analytics scan throughput and 1.9 PB data warehouses with Oracle Hybrid Columnar Compression (HCC).
With 3X more processor cores and up to 1.5X the database server memory of previous systems, customers can consolidate even more databases in a single Exadata Database Machine X10M than with previous generation systems, providing greater management efficiency and lower costs. Quarter-rack configurations start with two database servers and three storage servers and can be expanded by adding individual servers up to a total of 14 full racks that can be connected to meet the consolidated database needs of virtually any organization.
Application-transparent use of PMEM in Exadata X9M storage servers accelerates crucial customer transactional Oracle Database workloads with 19 microsecond read latencies, ultra-fast log writes, and up to 27.6 million read SQL IOPS per rack.
High-throughput RDMA over Converged Ethernet (RoCE) internal networking reduces system latencies so customers can rapidly process data with up to 1 TB/s throughput from all-flash storage.
Exadata X9M offloads SQL queries, in-memory analytics, and machine learning algorithms from database servers to up to 19 intelligent storage servers per rack to accelerate diverse customer workloads.
IO Resource Manager automatically optimizes resource allocations and prioritizes IOs to deliver consistently fast results for all types of customer database workloads in consolidated environments.
Up to 920 TB of NVMe flash cache per rack gives customers the performance of flash storage with the capacity and cost of disk storage by eliminating up to 95% of all disk IOs.
Exadata Database Machine X10M’s scalable design lets customers size initial deployments to meet current needs and add incremental database and storage servers to support new or increasing requirements.
Customers can utilize as few as 16 licensed CPU cores on Exadata Database Machine X10M or scale up consumption to thousands of cores to meet the needs of demanding workloads. Consolidating databases on fewer systems and managing them together reduces database infrastructure and management complexity by letting more resources be managed as a single unit.
Exadata Database Machine X10M’s high-capacity storage server has 22% more disk storage capacity than the previous generation system, supporting up to 1.2 PB of usable, triple-mirrored storage and more than 1 TB/second of analytics throughput in a single rack. Multirack scalability helps meet the database storage and analytics needs of large organizations.
All-flash Exadata X10M Extreme Flash storage servers provide up to 630 TB of usable, triple-mirrored storage capacity, a 2.4X increase over previous generations. In addition, the new Extreme Flash storage servers increase Exadata Smart Flash Cache capacity to 27.2 TB, further accelerating Exadata Smart Software features.
Exadata Database Machine X10M provides high compute and storage density, letting organizations consolidate their Oracle Database workloads onto smaller configurations that cost less and use less data center space.
Exadata is engineered to provide the highest levels of availability with completely redundant hardware and software. When deployed with Oracle Maximum Availability Architecture (MAA), a set of proven best practices, Exadata provides unsurpassed uptime for running Oracle Database on-premises.
Oracle Automatic Storage Management mirrors permanent data storage across two or three storage servers to maximize the performance and availability of customer databases.
Oracle RAC transparently scales Oracle Database instances across multiple servers to enable higher performance and protect against database server failures.
AI-based fault detection helps identify potential problems with customer database infrastructure and mitigate them before they affect operations.
Monthly releases of pretested updates for Exadata hardware, software, and firmware let customers quickly address security issues for all their Exadata systems.
Exadata consolidates customer databases and infrastructure into fewer externally accessible components, reducing the available surface area that hackers could potentially attack.
Oracle Advanced Security fully encrypts customer databases while they are at rest, in motion, and protected in backups so that unauthorized access doesn’t reveal information.
Role-based administration restricts IT team members’ access to the Exadata database stack, limiting them to the minimum information required to complete their tasks.
Continuous protection of customer Oracle Database instances with Oracle’s Zero Data Loss Recovery Appliance enables rapid recovery to less than one second before an outage or cyberattack.
Machine learning–driven management of Oracle Database indexes quickly adapts to changing workloads so crucial customer workloads run faster and require less manual tuning by DBA teams.
Automated allocation of I/O resources to individual Oracle Database instances based on database type, I/O operation, and end user lets diverse customer databases running in a consolidated environment achieve fast, predictable performance.
Exadata automatically detects CPU, memory, and network issues affecting the performance of customer databases. ML algorithms are then used to determine the root causes of problems, enabling IT teams to operate more efficiently.
Automated storage failure detection and rebalancing enable quick restoration of peak performance for customer databases.
Offloading database operations to storage servers enables more CPU cores to work on crucial customer database queries while freeing database servers to handle more OLTP users and complex queries.
Customers reduce the number of Oracle Database licenses they need for consolidated workloads, as each licensed core on Exadata database servers completes more work per unit of time and storage server CPUs do not require database licenses.
Exadata’s acceleration of converged Oracle Database capabilities eliminates customer requirements to deploy specialized databases. Multiple database models run within Oracle Database with high Exadata performance, availability, and consistent security.
High overall Exadata performance, scale, and management automation allow customers to achieve dramatic consolidation of Oracle Database instances to reduce resource over-provisioning, system counts, and TCO.
Exadata's CoD licensing enables customers to reduce initial costs by starting with as few as 16 Oracle Database licenses on eighth-rack systems or 28 licenses on quarter-rack systems. Customers can increase the number of licenses enabled to closely align costs with workload demands.
The same Oracle Exadata capabilities are available on-premises, in Oracle Cloud Infrastructure, and as Cloud@Customer solutions. Customers can lift-and-shift crucial workloads to cost-effective cloud environments with minimum effort.
Oracle Multitenant enables customers to simplify database management and lower costs by consolidating up to 4,096 pluggable databases per container database on Exadata, 16X more than on other systems.
Oracle Database In-Memory on Exadata simultaneously creates optimized row and column database formats within a single database instance. Dual database formats allow customers to quickly perform real-time analytics on production OLTP databases.
Oracle Machine Learning on Exadata enables customers to improve data-driven insights by easily adding machine learning capabilities to existing database applications. Customers can rapidly build and test ML models by eliminating data exports and transformations required for specialized systems.
Unique Oracle Database compression algorithms reduce typical storage requirements and load times for data warehouses by more than 90% (PDF).
Oracle Active Data Guard lets customers create hot-standby databases that provide business continuity across system failures. The addition of Far Sync extends this capability to remote data centers so organizations can continue operations during complete outages or disasters at their primary data centers.
Customers achieve high Oracle Database performance with an entry-level Exadata system that delivers up to 2.8 M Read IOPS. The system includes 64 CPU cores, 38TB of flash cache, and up to 96 TB of HDD usable storage capacity inclusive of 3-way storage mirroring.
Customers can deploy midrange to high-end Exadata X9M solutions with Oracle’s flexible configuration option to meet company-specific requirements. Start with 128 CPU cores, 4.5 TB of PMEM, and either 44 TB of all-flash or 192 TB of usable HDD storage capacity inclusive of 3-way storage mirroring. Add additional database servers, storage servers, or additional racks as needed.
Customers address requirements for large shared-memory databases with two or three 192-core database servers with up to 6 TB of memory each and up to 14 storage servers per rack.
High-capacity storage servers without PMEM or flash acceleration enable IT teams to build low-cost, scale-out storage for infrequently used data. Ovum reports (PDF) the XT storage has costs similar to that of Hadoop clusters and eliminates the need to secure separate analytics environments or copy data between them.
Large enterprises can scale X9M configurations linearly in performance and capacity with up to 12 system and storage expansion racks to meet enterprise-wide database consolidation needs.
Customers can upgrade existing Exadata X8M systems with additional X9M database servers and intelligent storage servers to economically meet increasing performance and capacity requirements. Infiniband based X8 systems can be upgraded with additional X8 database storage servers.
Thousands of global customers in all industries run business-critical Oracle Database instances on Oracle Exadata, accelerating crucial applications, generating new insights, and reducing TCO.
Migrate Oracle databases from other platforms and consolidate them on Oracle Exadata to respond faster to demand, increase performance, and cut costs by 50%.
Consolidate core banking transactional and analytical workloads on Exadata to achieve a greater return on investment.
Eliminate unplanned downtime and increase the performance of company applications during peak-season demand.
Accelerate online sales, inventory management, and financial management applications with up to 10X greater database performance.
Ashish Ray, Vice President of Product Management for Exadata Platforms and Cloud Services
The next-generation Exadata X10M delivers unrivaled performance and availability for all database workloads. Available starting at the same price as the previous generation, these platforms also provide significantly more capacity and dramatically greater value than previous generations.Read the complete post
See Larry Ellison’s announcement of new Oracle Cloud@Customer capabilities.
Learn how Exadata reduces customer costs, on-premises and in Oracle Cloud Infrastructure.
Find out how Oracle Exadata meets business-specific needs.