Oracle application owners and database admins face tough challenges when it comes to protecting a large number of databases with data volumes continuously expanding due to the growth in database sizes as well as regulatory compliance policies requiring more extended data retention periods. For the most part, some can’t even run protection jobs for specific Oracle databases because of their size and the time that it takes for a protection jobs to complete.
As a result, organizations are forced to look into expensive, complicated and adventurous risky availability solutions to provide some form of data recoverability in the event of failure or corruption. This is a recognized challenge in the data center and one that enterprise organizations are in desperate need to overcome. When it comes to relational database management system, I have said it before, Oracle is by far one of the undeniable leaders in the business, and it’s also one of the most relied upon systems in the enterprise. The challenges organizations face along with their application owners, and database admins are amplified by the exponential data growth and the regulatory compliance demands of our current times requiring more extended data retention periods.
Some technologies can efficiently assist enterprise organizations overcoming the challenges with their expanding Oracle database environments, but it’s the implementation details and capabilities that can determine whether a suitable solution can be achieved. For the most part, the implementation and support of space efficiency features such as compression, deduplication are the main options to tackle these challenges. To set the context for what I want to discuss, let me provide an abbreviated definition to some of the technologies I’ll be referencing.
Compression – compression offers excellent value from a space efficient and size reduction perspective, but its effectiveness and compression ratios are limited to the type of data these functions are being applied. The compression of data can be performed inline or as a post process, all depending on the capabilities of its implementation. Much like everything else, there are pros and cons for the use of compression with Oracle databases but I that discussion is out of the scope of my argument here.
Deduplication – there are two main ways in which deduplication can be performed, and they are source side and target based deduplication. Deduplication techniques can be applied in one of two ways – used on a specific fault domain (single appliance) or globally.
Target deduplication – data is transmitted over a network onto a storage appliance (hardware or software), and then data is deduplicated at
that point. The target appliance approach is often more expensive, but it does provide some performance advantage when dealing with large amounts of data at scale.
Source-side deduplication – eliminates redundant data before transferring to a backup target repository at a server level. Source-side deduplication is implemented as a software component that runs on the server side where data is being protected. This server typically communicates with data protection solution which makes it possible for data to be deduplicated at its current location.
With source-side deduplication data that has been partially transferred or in its entirety is never transferred over the network again. This option reduces the utilization of network bandwidth. This is extremely valuable with respects to relational database management system and databases (Oracle).
For example, when using a file system to store database files, whenever a database is updated with new entries that database file changes and with traditional data protection product that file will be transferred in its entirety regardless of the protection job being flagged to be performed as an incremental protection job. With source-side deduplication, a system can inspect files and identify the changes in bytes and only transfer those changes over the network. This makes the transmission of data more efficient from a data transfers perspective over a network.
Data deduplication is performed at two different levels file-based or block-based. When comparing highest levels of efficiency between the two, file-based is not the most efficient form of deduplication because of the nature of its design to purge duplicated files.
File-based deduplication focuses on performing comparisons on files (data) that are scheduled to be protected or archived against additional replicas of a file that has already been stored. Block-based deduplication is designed to look within files (data) to identify unique values within each block of data. The blocks are divided into multiple chunks of similar fixed length, and their data is processed using either an MD5 or SHA-1 hash algorithm.
During this process, a unique identifier is generated for each chunk and is then stored in an index. At this point when a file is updated, only the changes are saved regardless of the size of the change. Block-based deduplication is more efficient, but it comes at the cost of higher overhead in processing power.
Now why wall all of this rambling? Well because it’s important to understand the value and why’s of what Cohesity’s DataPlatform can do to help enterprise organizations overcome the challenges of data protection and recovery strategies for their Oracle database infrastructures. Also, the fact Cohesity covers all of the technological points I made above with a much better and elegant implementation of the different technologies discussed above.
Cohesity’s hyperconverged data platform that includes a converged data protection solution amongst all of its enterprise storage features and capabilities. It also consists of a converged data protection solution that provides end-to-end data protection that is capable of replacing traditional backup and recovery solutions and simplifying operations.
Cohesity’s DataPlatform and DataProtect provide four different integration capabilities for Oracle databases via the tight integration with Oracle RMAN APIs that enables enterprise organizations to perform application consistent backups and restores combined with risk adverse workflows for safe and efficient operating procedures. From a space efficiency perspective, Cohesity’s DataPlatform provides global space efficiency features from global variable-length dedupe supporting both in-line and post-process dedupe and compression.
Cohesity Offers Four Powerful Oracle Integrations that are capable to meet the Data Protection and Recovery Requirements of the most Demanding Enterprise Oracle Database Environments
Direct mount of Cohesity DataPlatform to Oracle application servers
Cohesity storage abstractions can be presented as NFS volumes, enabling direct mounting to Oracle database servers. No intermediate media servers are required. RMAN automatically leverages Cohesity’s QoS policy for optimal data tiering as well as global deduplication and compression to reduce secondary storage consumption and footprint.
Automated integration with Oracle RMAN
- Automated and configurable Oracle RMAN based protection jobs.
- Significantly reduce recovery time by leveraging Oracle RMAN image copy backups and ensure that backup data is fully hydrated.
- Immutable backup ensures the backups are not comprised to ransomware
- Policy-based backups for Oracle databases; enables centralized scheduling of all RMAN scripts instead of manually scheduling backups across multiple Oracle servers.
- Simplifies administration by providing a single pane of glass for managing backup attributes and reports including:
- Tracking all backup tasks, schedules, alerts from one place
- Reporting on all backup tasks
- Ability to clone Oracle Database from backup copies
- Ability to have DBA registers custom RMAN scripts for data protection
Cohesity source-side dedupe functionality for Oracle database servers provides backup admins with the following advantages:
- Faster backups to fit the ever-shrinking backup window
- More frequent backups that result in lower RPOs
- Reduced network traffic between host servers and backup target
- Less load on the backup target to handle parallel ingest streams
- Simple to use and deploy with existing Oracle instances
- Ability to write Oracle Backup sets simultaneously to more than one Cohesity Appliance, eliminating the need for asynchronous replication
- Supports 11g & 12c versions of Oracle DB deployed on Oracle Enterprise Linux and Redhat Linux
- Supported RMAN backups configurations are Full backups, Incremental backups with or without Oracle change block tracking turned on
Snapshot offload of primary storage arrays
Through integration with primary storage arrays, Cohesity can easily and quickly offload and restore array data & snapshots from and to production environments.
- Optimal snapshot movement using primary storage array snapshot API
- No disruption to database servers as all backups are taken directly from primary storage
- Faster backup and improved RTO & RPO over conventional app server backups
- Provide application-consistent backups
- Reduce storage requirements and overhead on production storage by moving older snapshots to secondary storage
From simplicity of management, risk adverse operational workflows to faster data protection and recovery, Cohesity DataPlatform, DataProtect and its integration with Oracle RMAN can address a majority of the challenges enterprise organizations are currently facing with the data protection and recovery strategies of their large and growing Oracle database infrastructures.
To finish up, here is a demonstration showcasing one of Cohesity’s integration points with Oracle RMAN and the source-side deduplication configuration with multiple streams.
You can find another Cohesity and Oracle related article and demo on the link below:
– Enjoy and don’t forget “The Sky’s the limit.”
For future updates about Cohesity, Hyperconverged Secondary Storage, Cloud Computing, Networking, VMware vSAN, vSphere Virtual Volumes (VVol), vSphere Integrated OpenStack (VIO), and Cloud-Native Applications (CNA), and anything in our wonderful world of technology be sure to follow me on Twitter: @PunchingClouds.