February 4, 2014 | Written by: Jonathan Savage
Share this post:
For anyone experiencing out-of-control storage, there’s a good chance that excess “copy data” from systems like backup, snapshot, disaster recovery, and test/dev is the culprit. IBM is using new technology delivered in a storage appliance, specifically to cure this disease.
IBM has just announced the first release of the SmartCloud Data Virtualization (SCDV), which is the latest service offering in the IBM BCRS SmartCloud Resilience portfolio. This offering leverages an appliance which uses snapshot and deduplication technology to help save costs on replicated disk storage in a DR environment.
In this blog post, I’ll share a brief synopsis of the underlying technology, what it is, how it works and how IBM is leveraging it in its new solution. I’ll also discuss what it does well and its best position to the market
What is it and how does it work?
The underlying appliance leverages the IBM SVC (San Volume Controller) at its core. On top of the SVC this technology has a slick, user-friendly Adobe Air GUI, which offers hooks into VMware VADP (VMware API for Data Protection) and other agnostic storage arrays via iSCSI and FCP. The solution is then able to discover the Virtual Machines running on that host and utilize VM snapshots to create known recovery points for the VM. This procedure or technology, is termed as the Virtual Data Pipeline (VDP).
VDP also takes advantage of Changed Block Tracking capabilities after the initial snapshot backup of a virtual machine. VDP in combination with SVC technology from IBM is the basis for the product. SVC allows VDP to take advantage of SVC volume management, snapshot capabilities, and other capabilities built into SVC. The solution can also take point-in-time snapshots of physical servers as well, either by using an agent-based “connector,” which can be installed on the source server, or by leveraging the SVC engine. This allows the solution to be used as the storage target for a source server. Again, these snapshots are deduplicated using the Changed Block Tracking capabilities after the initial snapshot is performed.
Once VDP has received and stored the VM snapshot data and subsequent changed blocks for each VM, this technology now has the ability to restore VM data through a unique process. Typical backup/recovery tools will stream the data back to the VM to place the data back onto the original datastore of the VSphere (ESX) server.
This is where SCDV separates itself. Since this technology stores the data blocks of the original VM, and the volumes containing the snapshot pool are SVC-based, this technology takes advantage of SVC to allow the snapshot pool to have snapshot versions of the pool externally mounted by the VSphere (ESX) host that the user wants to restore the VM data to. With this in mind, VDP does not “restore” the data to the original ESX location. VDP takes a snapshot from the snapshot pool and only the data blocks that represent the VM being restored and presents that “restoration” volume back to the VSphere server to be mounted as a new datastore. With the snapshot now presented back to the VSphere server, this technology automatically adds the VM to the VSphere server as a new VM and sets this new, restoration VM to be started on the host.
This is critically different than any other backup/recovery technology. At this point, the appliance is acting as a primary disk device to the VSphere server on which VCenter can now start that VM and run it in seconds/minutes in it’s entirety. There is not the long wait that is needed by traditional backup/recovery products of getting each data block transferred from the backup media to the VSphere datastore; the appliance IS the datastore. Thus restoration of the entire VM is measured in seconds or minutes.
What does it do well?
1. Copy Data Management
The native ‘Golden Image’ technology allows for only the changed blocks for protection targets to be transferred (snapshot or replication), which greatly reduces replication bandwidth and duration. Depending on size and change rate, snapshots run in just seconds (to just minutes at most), essentially eliminating traditional backup windows. This reduction can result in up to a 70 percent improvement in network optimization. This technology also greatly reduces the amount of data to be stored in the data pools on the appliance as only the unique data is kept across all secondary data. This can result in up to a 90 percent improvement in global deduplication.
2. Disaster Recovery for Virtual Servers
With dedicated infrastructure in place (compute, network, etc.), the appliance can be used for extremely quick recovery of a virtual environment. Entire virtual machines can be restored from the appliance in just minutes, while mounts of the snapshots from the appliance can take place in just seconds. This allows customers to spin up just a few of their critical VMs in a DR scenario so production data can actually be served off the appliance for a short time.
What is the best fit for it in the market?
The CDS appliance can handle backup just fine and does have some nice things going for it in that area (VDP, elimination of backup windows). Snapshot data from the CDS appliance can also be used to clone new VMs relatively quickly, which can then be used for test/dev.. As one of our marketing slides relays, “SCDV combines the recovery capability of Virtualized Server Recovery with the retention capabilities of Smart Cloud Managed Backup.”
What value does IBM Bring?
IBM has tested and verified the capabilities of this product through IBM’s engineering labs. This vetting process has caused great improvement in the underlying technology for a much more refined offering to the market for the GA release. IBM standardizes with two sites (source, destination), under the assumption that the second site will be in a BCRS recovery center, to leverage the 40 years of recovery experience in BCRS. Like other offerings in the SmartCloud Resilience portfolio, SCDV will receive 24/7 monitoring and management by the BCRS Delivery and Steady State teams. IBM has also created a proprietary and innovative customer portal to make the management and reporting process seamless.