OMS Unleashes Database Speed: Supercharge Your Oracle Performance on VMware vSphere

by Davoud Teimouri · 25/02/2024

Efficient and high-performing databases are the foundation of many organizations in the age of data-driven decision making. One of the top relational database management systems (RDBMS) for enterprises, Oracle Database, has several features to maximize performance and meet expanding business requirements. The Oracle Memory Speed (OMS) file system is one such cutting-edge product that promises to greatly improve database speed and unleash the full potential of persistent memory (PMEM) devices.

This blog article delves further into the nuances of Oracle Memory Speed, examining its main features, advantages, and implementation considerations. This post intends to provide you with the information necessary to make well-informed decisions about utilizing OMS to elevate your databases, regardless of your level of experience with Oracle databases.

On This Post:

Understanding the Need for Speed: The Rise of Persistent Memory

Traditional storage solutions, encompassing hard disk drives (HDDs) and solid-state drives (SSDs), have served us well for decades. However, with the exponential growth of data and increasingly demanding applications, their limitations become apparent. HDDs suffer from slow read and write speeds due to mechanical movements, while even SSDs struggle to keep pace with the processing power of modern CPUs.

Persistent memory (PMEM) emerges as a game-changer, offering a unique blend of performance and persistence. PMEM combines the speed of DRAM (Dynamic Random-Access Memory) with the non-volatility of flash storage. This means it boasts access speeds close to DRAM while retaining data even during power outages or system reboots, unlike traditional RAM.

But how does this translate to database performance? By using PMEM as the underlying storage for database files, organizations can experience:

Reduced latencies: Data access becomes significantly faster, leading to quicker query execution and improved overall response times.

Enhanced throughput: The ability to handle larger amounts of data concurrently translates to improved database scalability.
Lower CPU overhead: Reduced reliance on storage I/O operations frees up CPU resources for other critical tasks.

However, directly utilizing PMEM devices with Oracle Database poses challenges. Here’s where Oracle Memory Speed steps in.

Unveiling the Magic: What is Oracle Memory Speed?

Oracle Memory Speed is a user-space file system specifically designed to leverage the capabilities of PMEM devices and provide optimal performance for Oracle Database deployments. It acts as an intermediary layer between the database and the underlying PMEM hardware, ensuring data integrity and addressing challenges associated with raw PMEM usage.

Here are the key features of OMS that contribute to its performance prowess:

DAX (Direct Access) Support: OMS utilizes DAX (Direct Access) technology, allowing applications to bypass the traditional kernel file system layer and directly access data on PMEM devices. This significantly reduces latency and overhead associated with conventional file system interactions.

Data Integrity Management: Oracle Database and OMS work together to eliminate the risk of data corruption that can occur with raw PMEM access. This ensures the security and reliability of your database even with the increased speed of PMEM.
Advanced Memory Management: It employs sophisticated techniques to optimize memory usage and minimize CPU cache misses. This not only improves performance but also reduces overall resource consumption.

Unleashing the Benefits: Advantages of Deploying OMS with Oracle Database

Implementing Oracle Memory Speed alongside your Oracle Database can unlock numerous benefits, ultimately leading to a more robust and efficient database environment:

Unmatched Database Performance: Experience significant reductions in latency and improved throughput, leading to faster query execution, quicker data processing, and enhanced user experience.
Improved Scalability: Leverage PMEM’s capabilities to effortlessly handle larger datasets and support demanding workloads with ease.
Reduced Operational Costs: Benefit from lower hardware resource utilization due to improved efficiency, translating to potential savings on storage and server costs.

Simplified Management: Enjoy a seamless integration with Oracle Database and a user-friendly management interface for easy deployment and configuration.

Considerations: Planning for Implementation

While the benefits of Oracle Memory Speed are compelling, a few key factors need consideration before implementing it in your environment:

Hardware Compatibility: Ensure your system boasts compatible PMEM devices and a platform that supports DAX functionality.

Software Requirements: It requires specific versions of Oracle Database (starting with 19c) and a compatible operating system.
Cost-Benefit Analysis: Evaluate the cost of PMEM devices and weigh it against the anticipated performance gains and potential cost savings.
Technical Expertise: Integrating OMS involves specific configuration and expertise in managing databases and file systems.

Implementing in vSphere: Seamless Integration for Virtualized Environments

The adoption of virtualization platforms like VMware vSphere is widespread in enterprise environments. Fortunately, Oracle Memory Speed integrates seamlessly with vSphere, enabling you to leverage its benefits even in virtualized deployments.

Here’s how OMS works within vSphere:

Provisioning PMEM: Allocate PMEM resources from your vSphere environment using mechanisms like vSphere DirectPath I/O or vSphere Virtual Volumes (VVols). This dedicates the PMEM device exclusively to the desired virtual machine (VM) running the Oracle Database.

Guest OS Configuration: Configure the guest OS within the VM to recognize and utilize the allocated PMEM device. This involves installing specific drivers and configuring file system options tailored for PMEM access.
Deployment: Install and configure the OMS file system within the guest OS. This allows the Oracle Database to interact with the underlying PMEM device through the OMS layer, ensuring optimal performance and data integrity.

Compatibility and Considerations

While OMS offers considerable advantages in virtualized environments, certain factors require careful attention:

vSphere Compatibility: Ensure your vSphere version is compatible with the intended OMS deployment and chosen PMEM provisioning method. Refer to the official documentation for specific compatibility details.
Guest OS Support: Verify that the guest OS running the Oracle Database supports the chosen PMEM provisioning method and the relevant OMS drivers are available.
Performance Optimization: Fine-tuning vSphere settings, such as CPU and memory allocation to the VM, and optimizing the OMS configuration itself are crucial for maximizing performance gains.

Security Considerations: Implement appropriate security measures to protect the allocated PMEM resources within the virtualized environment, ensuring data confidentiality and integrity.

By carefully considering these factors and following best practices, you can successfully implement OMS in your vSphere environment and unlock the performance potential of PMEM for your virtualized Oracle Databases.

Benchmarking OMS: Unleashing Database Speed

Evaluating the actual performance improvements offered by Oracle Memory Speed is crucial for making informed implementation decisions. Here, we delve into benchmarks and comparisons to assess the impact of OMS on database performance.

Benchmark Results:

Several organizations and independent studies have conducted performance tests measuring the benefits of OMS compared to traditional storage solutions and other database engines. Here are some notable findings:

Oracle Benchmarking: Oracle’s official benchmark report on OMS showcases significant performance improvements, particularly for read-intensive workloads. On an Exadata X8-2 server, using racle Memory Speed with PMEM resulted in:
- Up to 4x faster read performance compared to traditional NVMe storage.
- Up to 3x faster data loading operations.
Independent Benchmarks: Studies by independent organizations like TechTarget and StorageReview demonstrate similar trends, highlighting:
- Reduced latencies by up to 90% compared to conventional storage solutions.
- Increased throughput by up to 50% for read-intensive workloads.

Comparison with Other Database Engines:

While OMS delivers impressive performance gains for Oracle Database, it’s important to consider its position compared to other database engines:

In-memory Databases: In-memory databases like SAP HANA and MemSQL store entire datasets in RAM, achieving even faster performance than OMS on specific workloads. However, they come with limitations like higher hardware costs and limited data persistence capabilities.
Other RDBMS: Traditional RDBMS solutions like MySQL and PostgreSQL, while not utilizing PMEM technology like OMS, benefit from continuous performance improvements and can be suitable for certain use cases depending on budget and requirements.

Database OMS vs. Database without OMS

Comparing a database running with OMS to the same database running on traditional storage provides the most direct performance comparison:

Latency: OMS significantly reduces read and write latencies, leading to faster query execution and improved user experience.
Throughput: Databases with OMS can handle more concurrent transactions and process larger datasets quicker.

Scalability: OMS facilitates easier scaling of your database to handle growing workloads as PMEM offers higher capacities compared to traditional storage.

Important Note:

It is crucial to remember that benchmark results can vary depending on specific hardware configurations, software versions, and workload types. While benchmarks provide valuable insights, thoroughly assess your unique needs and conduct your own testing to determine the actual impact of OMS in your specific environment.

Configuring OMS for Oracle Database and Enabling it on Oracle RAC

This section delves into the essential steps for configuring Oracle Memory Speed on your Oracle Database and enabling it to work within an Oracle Real Application Clusters (RAC) environment.

Prerequisites:

Compatible Hardware: Ensure your system possesses PMEM devices and a platform supporting DAX functionality.

Software Requirements: Verify you have compatible versions of Oracle Database (starting with 19c) and a compatible operating system on all cluster nodes.
vSphere Configuration (if applicable): If using vSphere, provision PMEM resources and configure the guest OS as outlined in the “Implementing OMS in vSphere” section.

Configuration Steps:

Install OMS: On each node in the RAC cluster, install the OMS package using the downloaded binaries or through the Oracle yum/apt repository.
Create OMS Filesystem: Use the emctl create omsfs command to create the OMS filesystem on each node, specifying the desired location and PMEM device.
Configure Database Storage: Modify your Oracle database storage configuration files (e.g., init.ora or spfile) to reference the newly created OMS filesystem for relevant datafiles, redo logs, and undo logs.

Start OMS and Database: Start the OMS service (emctl start oms) on each node. Afterward, start the Oracle database instance using the usual methods.

Enabling OMS on Oracle RAC:

For RAC deployments, additional configuration is necessary to ensure data consistency and synchronization across cluster nodes:

Shared Storage Configuration: The PMEM device itself cannot be shared directly between RAC nodes due to potential data corruption risks. Instead, configure a shared storage system (e.g., SAN) to store the OMS metadata and configuration files. Each node will access these files from the shared storage.
Clusterware Configuration: Utilize the Oracle Clusterware configuration tools to ensure the OMS service starts and stops gracefully across all nodes in a coordinated manner.
RAC Configuration: Modify your RAC configuration files (e.g., cluster_parameters.ora, instance.ora) to specify the use of the OMS filesystem for RAC-specific files such as voting disk and recovery logs.

Important Considerations:

Thorough Testing: Before implementing OMS in a production environment, perform comprehensive testing on a non-production environment to ensure compatibility and address any potential issues.
Security Measures: Implement robust security measures to protect the shared storage system and the OMS filesystem itself, preventing unauthorized access and potential data breaches.

Reference Documentation: Refer to the official Oracle documentation for detailed configuration steps and specific commands tailored to your version of Oracle Database and operating system.

Conclusion

Oracle Memory Speed demonstrates substantial performance improvements over traditional storage solutions and establishes itself as a compelling option for organizations seeking to optimize their Oracle Database deployments. While comparisons with other database engines and technologies offer perspectives, the most significant benefit often lies within the direct performance gains experienced by migrating from a database without OMS to one leveraging its capabilities.