Leveraging IBM Spectrum Protect De-duplication and Node Replication

Tape restore isn’t dead yet…but the clock is ticking.

There can be nothing scarier for a Recovery Administrator than hearing the words “We lost a drive on a file server…” This typically meant spending long hours watching tape after tape mount while trying to satisfy anxious end-users asking “is it done yet?”

Did you happen to notice I said Recovery Administrator and not Backup Administrator? If you’re a CIO, would you rather hire a Backup Administrator or a Recovery Administrator? The latter is where the rubber hits the road, as a backup environment is only as good as its ability to restore. That being said, regardless of your actual title you should think of yourself as a Recovery Administrator.

The latest release of IBM Spectrum Protect v7.1.3 (formerly Tivoli Storage Manager) has taken disk-based backup/restore to the next level by introducing two new storage pool types.

Directory-Container Storage Pools
The directory-container storage pool provides inline data de-duplication. In addition to the inline de-dup, client-side de-dup is still supported if network bandwidth prohibits the transfer of backup data.

Inline de-duplication provides improved server performance and scalability. IBM has published v7.1.3 capacity at 30 TB daily ingest and up to 4 PB of data per instance. This could potentially allow you to consolidate instances and simplify management.

The caveat for utilizing directory-container storage pools is the data cannot be migrated or copied to tape. It is a true end-to-end disk-based solution. Offsite data protection is accomplished via node replication.

Cloud-Container Storage Pools
The cloud-container storage pool also provides inline or client-side de-duplication with the target disk being from an object-based cloud storage provider. The cloud-container storage pool reduces onsite storage management and provides secure transmission of data via server-side encryption.

These new de-duplication offerings could be a game changer for IBM, since the previous file device class de-dup had limitations in terms of scalability and de-dup ratios. The enhanced de-duplication algorithm is reported to provide greater ratios resulting in improved restore performance and cost savings for customers.

So how do we put these to use? When I’m architecting a backup environment, I am of the mindset that every resource has a purpose and an administrator’s role to utilize them all efficiently. Incorporating these new features into your existing environment could provide you with a tiered service offering.

Tier 1: Production systems would use directory-container de-duplicationnode replication to provide onsite and offsite disk-based backup and recovery.

Cloud-container storage pool could be leveraged for periodic archives of data.

Tier 2: Non-production systems could use traditional disk-to-tape based backup and recovery.

Tier 3: Large database applications that utilize replication (ie. Oracle Data Guard) as primary recovery could make efficient use of tape backup, since large files stream very well to tape over fibre channel.

A tiered system such as outlined above would provide an efficient means to meet recovery SLA’s and save yourself countless hours running TSM administrative processes to maximize tape usage.

In summary, IBM has made great strides in providing a scalable disk-based backup solution resulting in improved performance in data restoration. IBM’s expanding portfolio is literally transforming TSM from an enterprise backup product to IBM Spectrum Protect being a comprehensive data protection solution.

Leave a Reply

Your email address will not be published. Required fields are marked *