KOHb - Getty Images

Tip

How to use AI in storage management

Organizations can use AI for storage management and AIOps to learn where to make improvements in their storage infrastructure. However, take note of the benefits and drawbacks.

The rise of AI has advanced storage management from storage resource management software to AI for IT operations tools that automate much of the process.

Artificial intelligence for IT operations (AIOps) enables monitoring, diagnostics, predictive analysis and prescriptive functions for storage infrastructure and applications. Essentially, AIOps can tell an organization what is happening with storage and why, what could happen and what to do about it.

By taking much of the manual work out of storage management, AIOps drives efficiencies and frees IT staff to work on other tasks. A variety of vendors provide AI for storage and storage management, which has advantages and disadvantages.

How AI storage management works

AIOps for storage uses machine learning to collect and analyze telemetry data. This information is turned into predictive analytics.

AIOps includes automation, performance management and service management and automates many of the decisions involved in scaling and securing storage systems.

Telemetry data is a key piece of AI storage management tools. Telemetry data is information collected through sensors from storage, servers and networking systems.

AI and machine learning analyze information gathered from these systems on hardware devices, OSes, applications and hypervisors. This helps to detect anomalous activity, such as poorly configured devices, surprising capacity growth or unusual throughput demands. These activities can be used for resource planning and to optimize storage performance.

AI storage management systems are often SaaS applications that perform the analysis on a public cloud. This enables those applications to compare that data to information collection from a wide range of systems, increasing their predictive capabilities.

The differentiation in AI-based storage management tools comes mainly from how well the vendor's algorithms can understand and use the information gathered. Gathering activity logs is not enough; the information must be translated into useful action in real time.

AI tools should also have strong sets of APIs to gather information from as many third-party sources as possible. Also, consider how the systems deliver alerts. These usually come from dashboards, texts, emails or a combination.

AI storage management options

Several vendors offer AI storage management software that is compatible with various systems and applications. Vendors also use AI storage management with their storage-as-a-service offerings.

Vendors and the systems and applications they cover include the following:

  • Dell CloudIQ. All Dell EMC storage; PowerEdge servers; VxRail, PowerFlex and VxBlock converged and hyper-converged infrastructure (HCI); PowerProtect DD and PowerProtect Data Manager data protection; and PowerSwitch and Connectrix networking.
  • HPE InfoSight. Alletra, Primera and Nimble Storage; SimpliVity HCI; ProLiant and Apollo servers; and Synergy composable infrastructure.
  • IBM Storage Insights. All IBM FlashSystem storage, DS8000, Spectrum Accelerate and Spectrum Virtualize. IBM Storage Insights Pro, the paid subscription version, also covers IBM Spectrum Scale, Cloud Object Storage, and select storage systems from Dell EMC, Hitachi Vantara, NetApp and Pure Storage.
  • Infinidat InfiniVerse. InfiniBox, InfiniBox SSA.
  • NetApp Active IQ. OnTap, E-Series, StorageGRID storage and Cloud Backup.
  • Pure Storage Pure1. FlashArray, FlashBlade and Portworx storage.

Advantages of AI storage management

AI-driven storage management removes much of the complexity and manual tasks of traditional storage resource management. Dedicated storage personnel can spend less time monitoring and managing systems. AIOps provides a big help for MSPs by enabling them to remotely manage many customers' storage.

By predicting future events based on current usage patterns, management systems can prevent problems that users may not see otherwise. They also can advise users to add storage capacity and compute resources and complete other upgrades before performance becomes a problem. Users can set the applications to automatically take action to prevent device failures or performance degradation, although IT shops may prefer to receive recommendations to make those changes themselves.

AI in storage management is a potential tool for detecting ransomware or other attacks early enough to take action that mitigates or eliminates damage.

Challenges of AI storage management

One issue with AI storage management systems is that they usually work with only one vendor's products. For example, if an organization has a SAN from one vendor, another vendor's AI storage management system may not be compatible with it.

Another issue is the collection and analysis of data creates more data -- a case of storage management creating a need for more storage. Over time, organizations must decide which data to safely discard.

Not all organizations can allow third parties to connect to their data centers. So-called dark sites cannot use SaaS-based analytics that collect and store data in public clouds or the vendor's site. Vendors can use analytics software on local servers so the telemetry data does not travel, but these users lose some benefit of the analytics when their data is not compared to that of their peers.

Dark site or not, all users should ask their vendor how much information it collects outside of storage and how the vendor will guarantee that data remains anonymous and protected.

Since AI in storage management is a new technology, AI algorithms will improve over time as more information is gathered. There is a long way to go before organizations can manage storage autonomously.

Dig Deeper on Storage management and analytics

SearchDisasterRecovery
SearchDataBackup
SearchConvergedInfrastructure
Close