Wow, this has been highly on my radar, one of the most anticipated integration point to XtremIO is almost available!

You wouldn’t drive a racecar without a helmet… and you shouldn’t sell XtremIO without data protection. But not just any data protection solution will do – XtremIO requires flash optimized data protection that can protect workloads at the speed of flash. ProtectPoint is the only solution in the industry that delivers this and now it supports XtremIO.

EMC ProtectPoint, an industry first data protection solution that integrates primary storage and industry leading protection storage, now supports XtremIO – the world’s #1 all flash array, which:

Enables up to 20x faster backup to meet application protection SLAs in addition to

Eliminates the backup impact on application servers and

Reduces cost and complexity by eliminating excess infrastructure including a traditional backup application

When buying an all flash array, performance is nonnegotiable and applications on your XtremIO contain some of your business’ most valuable data. So why leave things to chance when it comes to protecting that data? Ensure data protection doesn’t become a bottleneck by leveraging EMC’s flash optimized data protection solution – ProtectPoint for XtremIO.

The secret sauce of ProtectPoint is the combination of best of breed technology – that only EMC could bring together. ProtectPoint combines three components: application integration to ensure an application consistent backup, a change block tracking/data mover engine to ensure minimal infrastructure required and minimal impact on the network, storage and servers and protection storage to ensure cost effective retention and reliable recovery. The exact components may vary with different primary storage arrays, but ProtectPoint still brings the same value no matter what.

By leveraging best of breed technology from across the EMC family, we can:

  • Allow you to leverage common data protection infrastructure for multiple use cases (i.e. backup and replication)
  • Provides seamless integration between different parts of the EMC data center stack and easily scale with your environment
  • By developing ProtectPoint across multiple EMC teams, we are able to drive faster innovation (resulting in a broader ecosystem more quickly vs. having the engineer for each storage array and application separately) and broad investment across EMC exemplifies the strategic importance of ProtectPoint

Now, here’s a quick peek at the technology behind ProtectPoint with XtremIO.

  1. First is the agent installed on the application server – which could be the application agent (for Oracle, SAP or DB2) or the generic file system agent (for other applications). These agents provide an application consistent backup without impacting the host and are designed to empower application owners to control their own backup and recovery and eliminates the requirement for a traditional backup application.
  2. Next, the engine is powered by XtremIO and RecoverPoint technology (via a RecoverPoint appliance). This ensures that there is no impact to the production workloads on the XtremIO and that only unique blocks are sent to Data Domain– minimizing time and bandwidth required to backup an application.
  3. Finally, the Data Domain system has the capability to ingest blocks directly from the XtremIO, deduplicate and store them. But most importantly, even though the system only ingests unique blocks, it stores each backup as a FULL native image copy every time.

At a high level, this is how ProtectPoint works to backup directly from XtremIO to Data Domain. After an initial configuration by the storage administrator, which makes a point in time copy of the data to be protected and seeds the initial data on the Data Domain system, the environment is ready for it’s first full backup via ProtectPoint.

First… to trigger a ProtectPoint backup, an application owner, like an Oracle DBA, triggers a backup at an application consistent checkpoint. This pauses the application momentarily simply to mark the point in time for that backup.

Then… only unique blocks are sent from XtremIO directly to Data Domain. This is enabled by leveraging industry leading XtremIO virtual copy and RecoverPoint technology as the change block tracking engine, which efficiently tracks changes since the last backup.

Finally… the Data Domain system will ingest the unique data and use it to create an independent full backup in native format, which enables greatly simplified recovery.

With ProtectPoint, you do a full backup every time, but only send unique blocks, so the full comes at the cost of an incremental.

In addition, by leveraging RecoverPoint techonlogy, ProtectPoint for XtremIO also offers a new type of backup – triggered by the infrastructure (RPA) to backup data without an application owner triggering it. Let’s take a look at how that works:

First… the infrastructure (specifically the RecoverPoint appliance) triggers a backup based on a previously defined policy.

Then… only unique blocks are sent from XtremIO directly to Data Domain.

Finally… the Data Domain system will ingest and storage the unique data and use it to create a full backup.

[Note: even though she did not trigger it, the app owner can see these backups]

Let’s take a look at how a recovery works with ProtectPoint for XtremIO – first we’ll review a traditional full recovery, which would be recovering an entire LUN.

First, the app owner will trigger the recovery …

Then, the app agent reads the backup image from the Data Domain system and pulls the appropriate data back over the network.

At which point, the RecoverPoint will replace the appropriate production LUN with the recovered copy.

Next let’s take a look at how a full recovery works if it’s done via the storage admin leveraging RecoverPoint, which enables change block tracking on recovery.

First, the storage admin will trigger the recovery …

Then, the RecoverPoint appliance reads the backup image from the Data Domain system and then pulls only the required data back over the network.

At which point, the RecoverPoint will replace the appropriate production LUN with the recovered copy.

This enables a full recovery at the speed of a granular recovery – it is very similar to the Avamar capability with VMware support change block tracking on recovery.

In addition, you can also do a granular recovery with ProtectPoint via instant access. For an Oracle database this might be recovering a specific database, table or record – as opposed to the entire LUN.

First, the app owner or DBA triggers the recovery

Then, the application server connects to the backup image on the Data Domain – but the image doesn’t move off the system. This gives the DBA instant access to their protected data, although it is still on the Data Domain.

At this point, the app owner can recover the specific object she desires to the production database.

With ProtectPoint XtremIO integration, we get a few additional benefits by using RecoverPoint technology as the engine:

  • By leveraging a shared RecoverPoint appliance, ProtectPoint is a simple extension of our best in class replication offering and your customers can leverage their existing RecoverPoint appliances for ProtectPoint backups.
  • All replication and backup can be centrally managed via the RecoverPoint GUI for centralized storage consistent data protection. Alternatively, to empower app owners, day to day backup and recovery can all be centrally controlled by DBA via their native utilizes (i.e. RMAN).
  • By integrating with XtremIO virtual copies and leveraging a RecoverPoint appliance to do the processing, ProtectPoint ensures that the backup does not impact production workload performance on the XtremIO

ProtectPoint offers integration at two layers – either the native integration at the application layer or integration at the “file system” layer for applications that aren’t supported with a native appliance agent. ProtectPoint currently natively integrates with three major applications and databases: Oracle (via RMAN), SAP (via BR*Tools) and IBM DB2 (via Advanced Copy Services).

here’s a good white board discussion

and another overview with a demo at the end of it

1 Comment »

Leave a ReplyCancel reply