3dmentat - Fotolia

IBM Spectrum storage refresh targets object, AI metadata

IBM Spectrum storage sharpens its focus on AWS backup and metadata across heterogeneous storage. IBM increases the density of object storage and previews the VersaStack model with NVMe.

IBM Spectrum storage software expanded its scope to cover AI and large-scale analytics, including updates for compliance and deeper integration in Amazon Web Services.

IBM Spectrum is the vendor's brand for storage software. The products released this week extend the reach of IBM Spectrum Discover metadata management to include other vendors' storage. A refreshed IBM Cloud Object Storage supports denser capacity per rack and individual nodes.

In addition, IBM upgraded its Spectrum Protect Plus data protection to enable direct backup of local databases to Amazon's Simple Storage Service (S3). IBM also previewed a new VersaStack converged infrastructure that uses its FlashSystem 9100 NVMe rack-scale storage with Cisco servers.

Spectrum Discover metadata management is a recent addition to the IBM Spectrum storage software portfolio. Spectrum Discover layers on top of storage to ingest and index billions of files and objects stored locally and in the cloud. IBM said Discover can help to classify exabytes of unstructured data.

Originally designed only for IBM storage, Spectrum Discover now supports Dell EMC Isilon NAS and NetApp filers, as well as Ceph and any Amazon S3-compatible storage. 

IBM Spectrum storage added Discover to enable more efficient mining of metadata, said Eric Herzog, chief marketing officer and vice president of worldwide storage channels at IBM Storage. He said the significant feature enhancement is analytics across different storage systems.

With the new release, IBM beefed up Discover's capabilities for hooking metadata directly into AI and big data projects. Science teams can use Discover to search large metadata catalogs and connect third-party data analytics tools via built-in APIs.

Henry Baltazar, an analyst for storage at 451 Research, called IBM's updates "evolutionary, but not revolutionary." He said IBM Spectrum Discover adds important features for regulatory compliance and optimizing storage efficiency.

"What makes Spectrum Discover valuable is being able to see as much data as possible. The big addition with this launch is support for third-party search on other storage arrays. They didn't have that before. The more data people can get their hands on, the more powerful the infrastructure is going to be," Baltazar said.

Customers can use Discover to tag data with keywords and automate detection of personally identifiable information and other sensitive data.

"You can search the content, not just the metadata. We automated detection of certain sensitive data for GDPR and [privacy] regulations coming out of California and Brazil. You can still create custom metadata, but we included some defaults for apps that need to stay in compliance," Herzog said.

IBM Spectrum storage for database backup and object capacity

Amazon customers can protect Db2, Oracle, MongoDB and Microsoft SQL Server databases hosted on AWS, using S3 Intelligent Tiering to move data to IBM Spectrum Protect Plus. IBM also added more data retention options to tape and virtual tape libraries, Amazon S3 Glacier and Microsoft Azure Archive Storage.

IBM Cloud Object Storage arrays are based on technology IBM acquired from Cleversafe in 2015. Customers can purchase IBM object storage software as a cloud service, an on-premises deployment or embedded on IBM hardware. 

The latest Cloud Object Storage arrays use second-generation IBM hardware and bigger drives. The hardware scales to 10 PB in a single 42U rack and 1.3 PB per node, which IBM said equates to 26% more overall capacity. Use cases include AI, big data and secondary workloads.

IBM enables customers to mix and match old and new Cloud Object Storage systems, Baltazar said. "IBM is saying you don't need to drop all your stuff on the new hardware. I think people will take advantage of this right away, since they can mix and match without a forklift upgrade," he said.

IBM VersaStack uses Cisco Unified Computing System servers with IBM storage. The new version attaches to IBM FlashSystem 9100 arrays outfitted with NVMe SSDs and is due out in late 2019. IBM said it will continue to sell VersaStack models that use IBM A900, V9000 and Storwize storage.

Dig Deeper on Storage management and analytics

Disaster Recovery
Data Backup
Data Center
and ESG