RecoverPoint is a continuous data protection solution offered by EMC Corporation which supports asynchronous and synchronous data replication of block-based storage. RecoverPoint was originally created by Kashya, which was bought by EMC in 2006.
Capabilities[edit]
- Block-based journaling.
- Synchronous and asynchronous replication.
- Any-Point-In-Time – Every write is tracked and stored as a different snapshot. Alternatively, groups of writes can be aggregated according to configuration in order to reduce storage space and network traffic.
- Heterogeneous (multi-vendor) storage arrays via Fibre Channel.
- WAN-based compression.
- Tracking multiple volumes as a single consistency group.
Replication[edit]
RecoverPoint continuous data protection (CDP) tracks changes to data at a block level and journals these changes.[1] The journal then allows rolling data to a previous “Point-In-Time” in order to view the drive contents as they were before a certain data corruption. CDP can journal each write individually, hence enabling “Any-Point-In-Time” snapshots, or it can be configured to combine consecutive writes in order to reduce journal space and improve bandwidth. CDP works only over SAN – the RecoverPoint appliances needs to bezoned and masked with both the master, the replica and the journal LUNs.
RecoverPoint continuous remote replication (CRR) enables a replica in a remote site. For such a setup, RecoverPoint appliances clusters are required in both the local and remote sites. These 2 clusters communicate over either FC or IP. RecoverPoint applies compression and de-duplication in order to reduce WAN traffic. As of RecoverPoint 3.4, only one remote site. CRR can be combined with CDP in order to provide concurrent local and remote (CLR) replication.
The consistency group (CG) term is used for grouping several LUNs together in order to ensure write-order consistency over several volumes. This is used for example with a database that stores its data and journal on different logical drives. These logical drives must be kept in-sync on the replica if data-consistency needs to be preserved. Other examples are multi-volume file systems such as ZFS or Windows’ Dynamic Disks. RecoverPoint 3.4 supports up to 128 CGs and 2048 LUNs.[2] Each LUN can contain up to 2 TB, and the total supported capacity can be up to 150 TB.
Write splitting[edit]
Similar to other continuous data protection products, and unlike backup products, RecoverPoint needs to obtain a copy of every write in order to track data changes. RecoverPoint supports three methods of write splitting: host-based, fabric-based and in the storage array. EMC advertises RecoverPoint as heterogenous due to its support of multi-vendor server, network and storage environments.[3]
Host-based write splitting is done using a device driver that is installed on the server accessing the storage volumes. The usage of a host-based splitter allows replication of non-EMC storages. However, splitters are not available for all operating systems and versions.
Available fabric-based splitters are for Brocade SAN switches and for Cisco SANTap. This requires the investment in additional switch blades. This configuration allows splitting from all operating systems regardless of their version, and is agnostic to the storage array vendor.
Storage array splitters are only supported on a subset of EMC storages. This method allows write splitting from all operating systems, and does not require special SAN switching hardware. The RecoverPoint/SE is a slimmed-down version that only supports this type of splitter.
Architecture[edit]
Each site requires installation of a cluster that is composed of 2-8 RecoverPoint appliances. The multiple appliances work together as an high availability cluster. Each appliance is connected via Fibre Channel to the SAN, and must be zoned together with both the server (SCSI initiator) and the storage (SCSI target). Each appliance must also be connected to an IP network for management.
All replication takes place over either FC or standard IP for asynchronous replication. With RecoverPoint 4.0 and later synchronous replication can now take place over FC or IP.
One or more host-, fabric- or array- splitters would split traffic to both the storage and the appliances.
When configuring a consistency group, there is a need to select source LUNs on which the data will be monitored, target LUNs in the same size, and journal LUNs. The management GUI will indicate when the target LUNs are identical to the source LUNs, and will enable selecting an older timestamp in order to roll back the target LUNs to an historical state.
Integration with other products[edit]
Besides integration with EMC products such as AppSync, ViPR, Replication Manager, Control Center and Unisphere, and the CLARiiON,VNX, Symmetrix and VPLEX storage arrays, RecoverPoint integrates with the following products:
Integration with VMware vSphere, VMware Site Recovery Manager and Microsoft Hyper-V allows protection to be specified per VM instead of per volumes that are available to the hypervisor.
Integration with Microsoft Shadow Copy, Exchange and SQL Server and Oracle Database Server allows RecoverPoint to temporarily stop writes by the host in order to take consistent application-specific snapshots.
The usage of APIs/CLIs allows customers to integrate RecoverPoint with custom internal software.[1]