Oracle Exadata Database Machine X8M sets a new bar and changes the dynamics of the database infrastructure market. Exadata X8M combines Intel® Optane™ DC persistent memory and 100 gigabit remote direct memory access (RDMA) over Converged Ethernet (RoCE) to remove storage bottlenecks and dramatically increase performance for the most demanding workloads such as Online Transaction Processing (OLTP), analytics, IoT, fraud detection, and high frequency trading.
“With Exadata X8M, we deliver in-memory performance with all the benefits of shared storage for both OLTP and analytics,” said Juan Loaiza, executive vice president, mission-critical database technologies, Oracle. “Reducing response times by an order of magnitude using direct database access to shared persistent memory accelerates every OLTP application, and is a game changer for applications that need real-time access to large amounts of data such as fraud detection and personalized shopping.”
Exadata X8M helps customers perform existing tasks faster and accelerates time-to-insight, while also enabling deeper and more frequent analyses. Customers continue to realize performance benefits as workloads grow since the Exadata X8M’s performance scales as the platform grows.
“Oracle Exadata X8M significantly raises the bar for OLTP applications and simplifies database consolidation in demanding on-premises and cloud environments,” said Carl Olofson, Research Vice President, Data Management Software, IDC. “Any and all customers who want to improve their Oracle Database performance need to immediately evaluate this new offering. It’s a serious game changer.”
Engineering the Entire System Together is the Key
Typical implementations of persistent memory focus on using it as a faster alternative to flash. Persistent memory is accessed using conventional IO protocols that require OS calls, storage network messages, context switches, and interrupts. These methods were acceptable for existing storage devices such as disks and flash, but are much too slow in the new age of persistent memory.
In contrast, Oracle Exadata X8M uses RDMA directly from the database to access persistent memory in smart storage servers, bypassing the entire OS, IO, and network software stacks. This enables dramatically lower latency and higher throughput. Using RDMA to bypass software stacks also frees CPU resources on storage servers to execute more Smart Scan queries in support of analytics workloads.
Because persistent memory is located in shared storage, all databases consolidated on an Exadata platform get the performance benefits. Exadata Smart System Software automatically migrates the hottest database data to persistent memory while keeping cooler data on flash and disk. In addition, Exadata System Software automatically manages all high availability and redundancy of data stored in persistent memory. No application changes or administration tasks are required to leverage persistent memory with Exadata.
Eliminates Storage Bottlenecks
High-performance OLTP applications require a demanding mixture of high Input/Output Operations Per Second (IOPS) with low latency. Direct database access to shared persistent memory increases peak performance to 16 million SQL read IOPS, 2.5X greater than the industry leading Exadata X8. Additionally, Exadata X8M dramatically reduces the latency of critical database IOs by enabling remote IO latencies below 19 microseconds – more than 10X faster than the Exadata X8. These ultra-low latencies are achieved even for workloads requiring millions of IOs per second.
Performance and Consolidation Outpaces Traditional Storage
With end-to-end integration between database software and smart persistent memory storage, Oracle Exadata X8M goes beyond the performance achievable with other offerings using persistent memory. A single rack Exadata X8M delivers up to 2X the OLTP read IOPS, 3X the throughput, and 5X lower latency than shared storage systems with persistent memory such as a single rack of Dell EMC PowerMax 8000. By simultaneously supporting faster OLTP queries and greater throughput for analytics workloads, Exadata X8M is the ideal platform on which to converge mixed-workload environments to decrease IT costs and complexity.
Performance and Efficiency Dramatically Better than AWS and Azure
Other hyper-scale cloud providers lack a database optimized infrastructure like Exadata. When compared to the fastest Amazon RDS storage for Oracle, Exadata X8M delivers up to 50X lower latency, 200X more IOPS, and 15X more capacity. Compared to Azure SQL Database Service storage, Exadata X8M delivers 100X lower latency, 150X more IOPS, and 300X more capacity. The orders of magnitude performance advantage translates to dramatically faster applications and large cost reductions from the ability to run many more databases using much less infrastructure.
Featuring Machine Learning from Autonomous Database
Exadata X8M delivers the same machine learning capabilities as the Exadata X8, including Automatic Indexing, which continuously learns and tunes the Database as usage patterns change. Based on technology from Oracle Autonomous Database, the entire process is automatic and improves database performance while eliminating manual tuning.
Oracle Zero Data Loss Recovery Appliance X8
Oracle today also announced availability of Oracle Zero Data Loss Recovery Appliance X8M (ZDLRA), which uses new 100Gb RoCE for high throughput internal data transfers between compute and storage servers.
Exadata and ZDLRA customers can now choose between RoCE or InfiniBand-based Engineered Systems for optimal flexibility in their architectural deployments.