Storage integrated snapshots and clones have become standard features in many enterprise storage products. Snapshots can be the foundation of a data protection strategy, providing a point-in-time copy of data that can be used for instant recovery. Clones are similar to snapshots in that they can be created instantaneously and represent a virtual copy of data. Whereas data in a snapshot is immutable, data in a clone can be changed and can be used for operations like data analytics or test/dev. Datera relies on a unique combination of technologies to provide snapshots and clones that are high performance, efficient, resilient and provide flexible management.
Software Defined Storage has matured, and now offers the full range of services traditional storage arrays have offered for two decades. Datera customers enjoy deduplication, compression, encryption and fully integrated snapshot and clone capabilities. These services are even more powerful on a platform with autonomic, intent and policy driven control of workloads and volumes. Datera = SDS flexibility + full enterprise storage services + industry leading automation.
The Datera Storage Platform (DSP) architecture is designed for scale-out and efficiently distributes I/O across storage nodes for resiliency and performance. The system uses a multi-replica copy technique to ensure maximum data resiliency. All data and metadata associated with a snapshot is replicated across nodes like the source data to ensure the same level of resiliency and consistently. Redirect on write (RoW) snapshot implementation reduces storage system resource impact by minimizing the number of writes required when using snapshots. The datastore is implemented using augmented B+ trees for its meta-data. New data written after a snapshot is created is written to a new block. Previously written data in a snapshot does not have to be read and rewritten to storage like some implementations.
Snapshots for Data Protection
Snapshots are an essential component of a complete data protection solution for mission critical applications that have the most demanding service level requirements. This is especially true with large data sets that would take a long time to backup with traditional methods that move a lot of data and have a negative impact on system performance. Taking snapshots does not require data to be moved. Snapshot operations have minimal impact on the system so these snapshot recovery points can be created much more frequently than a traditional host file system based backup. More frequent snapshots provide a lower Recovery Point Objective (RPO) which means that the potential for data loss is much less. The other advantage of snapshots is that the Recovery Time Objective (RTO) process is also much lower, and that minimizes application down time.
The technical implementation of clones are very similar to snapshots, using the same B+ trees and space efficient pointer mechanisms. The major difference is that clones are exported as new read/write volumes. Writes to a clone result in the meta-data for the original volume and the clone to diverge. Another difference is that snapshots belong to the same volume and Application Instance as the source, whereas clones are created as a different volume in a new Application Instance and Storage Instance. One important advantage of Datera clones over most other implementations is that the clones are completely independent of the source volume. Another unique advantage is the ability to change the media or placement policy for the clone to provide different performance or cost characteristics from the parent volume.
Copy Data Management (CDM)
The idea of Copy Data Management (CDM) involves taking a holistic approach to re-using copies of data (virtual copies) for multiple use cases such as:
- Disaster recovery
- DR test
- Operational recovery (backup)
- Data analytics
CDM leverages Datera embedded technologies like snapshots and clones The user points to source data and specifies “virtual copies” of data. DSP supports clones that can be used to quickly make space efficient copies of data for test/dev, analytics and other copy data management workflows. For example, a single remote copy of data can be used for site disaster recovery, DR test and even operational recovery if the recovery points (snapshots) are retained long enough.
Cloud Backup (Datera2Object)
Datera2Object replication is a solution for backing up and restoring Datera primary snapshot data to and from a remote public cloud or on-prem object store. The solution uses replication to create copies of Application Instances and volumes in remote object storage solutions such as Datera system with S3 object services for backup. Backups can also be replicated to Amazon Web Services (AWS) S3, the Google Cloud Platform (GCP) and generic S3 Object Storage systems. The replication takes place directly between the Datera source system and an object storage system, making the process very efficient with low resource impact.
Datera2Object supports these primary use cases:
- Creating space efficient backups of Datera Application Instances and volumes in a remote object store
- Restoring Datera Application Instances and volumes from a remote object store to either the source system or and alternate Datera system
- Migrating Datera Application Instances and volumes to remote object store with the ability to migrate back to the same or another Datera system
Datera plugin for vCenter also enables snapshot data to be cloned and mounted to other ESX hosts to enable granular recovery or support test/dev workflows. Again, the plug-in exposes the power of storage clones to VI admins with a friendly user interface to automate workflows.
Datera vCenter Plug-in for VMware Snapshot List
The Datera Cinder driver provides seamless integration with OpenStack and several advanced functionalities to enable a seamless operator experience. The Container Storage Interface (CSI) is a standard for exposing arbitrary block and file storage systems to containerized workloads on Container Orchestration Systems (COs) like Kubernetes. Datera’s powerful storage classes, and policy driven workloads are a natural fit with Kubernetes and integrate with native snapshots for creating recovery points and doing instantaneous restores. The Datera CloudStack storage plug-in provides Datera storage plugin support for CloudStack Primary storage.
The Bottom Line
The evolution of software defined is indeed quite mature today, and continuing to mature. Software Defined Storage is delivering all the performance, capabilities, resiliency and features of decades old enterprise storage arrays architectures, with the modern simplicity, elasticity and automation that IT professionals need to be able able to support businesses that are operating in a very dynamic environment where agility, efficiency and economics can deliver a significant positive impact to the outcome.
Please click here to read a much more detailed and technical Solution Brief on Datera Snapshots and Clones.
For more general Datera information, we recommend reading our white papers:
We can schedule a demo at any time. Please reach us at firstname.lastname@example.org and share any specific capability you would like to learn more about. We look forward to the opportunity!