AI-Enabled Intelligent Infrastructure
Examples: Dell EMC PowerStore, HPE InfoSight, Huawei OceanStor, Pivot3, Pure Storage, Tintri
This is a Press Release edited by StorageNewsletter.com on December 29, 2020 at 2:18 pmThis report was written by Ken Clipperton, lead analyst, storage, DCIG LLC, on December 17, 2020.
The AI-enabled Intelligent Infrastructure Opportunity
Enterprise storage providers are embracing AI and ML to create AI-enabled intelligent infrastructure. The focus and maturity of each vendor’s offering varies. Some are focused on individual array health, relying more on fault data than predictive analytics. Others focus on application-level performance and availability. Both approaches offer significant benefits to customers.
The AI-enabled Intelligent Infrastructure Opportunity
The combination of predictive analytics and proactive support transforms the storage ownership experience. Vendors that have implemented predictive analytics with proactive support have avoided hundreds of significant service disruptions for their customers. Thus, eliminating the direct costs and reputational damage associated with that downtime.
The end goal of predictive analytics for the more visionary storage providers goes beyond eliminating downtime. Their goal is to enable infrastructures to autonomously optimize themselves for application availability, performance, and TCO based on the customer’s priorities.
These intelligent infrastructures automate routine infrastructure management tasks and the time-consuming overhead of tracking down, resolving, and then explaining the cause of application performance problems. As a result, intelligent infrastructures free IT staff to focus on projects that create new business value.
Embedded AI vs. Cloud-Based Storage Analytics
Some storage providers are embedding AI chips into storage arrays or integrating analytics into the storage OS. Either approach enables analytics to run directly on the array. The advantage of this approach includes the ability to adjust in near real time to deliver optimal application performance and safely consolidate many mixed workloads onto a single storage system.
Other storage array vendors collect fault and telemetry data from the entire installed base and run AI/ML on the data in the cloud. This telemetry data includes IO/s, bandwidth, and latency associated with workloads, front end ports, storage pools and more. An advantage of the cloud-based approach is the opportunity to apply predictive analytics and ML algorithms across data collected from the entire installed base, and then leverage that analysis to identify potential problems and optimization opportunities for each array.
Need for Visibility
When infrastructure is running smoothly, it is important to be able to demonstrate to various stakeholders that SLAs are being met. When problems arise, visibility into the infrastructure is even more important. Infrastructure analytics can dramatically accelerate root cause analysis and expedite problem resolution.
Analytics-based visibility can also aid capacity planning that enables the infrastructure to meet performance requirements over time. Capacity planning involves more than extrapolating trend data. As organizations seek to gain more value from their data and IT systems, they add new workloads to the data center.
Some storage vendors now provide analytics that can model the capacity and performance impact of adding a specific workload to the environment. They can determine optimal workload placement, whether upgrades will be needed, and identify the specific upgrades required to meet SLAs.
AI and Importance of Manual Override
When evaluating the AI/ML capabilities of data center infrastructure products, enterprises should look for products that leverage AI/ML by to automate infrastructure management, yet which humans can override based on site-specific priorities, preferably on a granular basis.
After all, when a critical line of business application is not getting the priority it deserves, the last thing you want to hear from your infrastructure is, “I’m sorry, Dave. I’m afraid I can’t do that.”
Examples of AI-Enabled Intelligent Infrastructure
- Dell EMC PowerStore includes an onboard intelligence engine that automates initial volume placements based on available capacity, drive wear, and other factors. PowerStore aids capacity planning by estimating “out of capacity” date and recommending volume migration plans for approval. Dell EMC also includes CloudIQ cloud-based analytics with PowerStore support plans.
- HPE InfoSight for Nimble Storage delivers mature, advanced analytics for infrastructure built on a bedrock of data, telemetry, algorithms and data science. Its cloud-based AI platform that uses ML to predict and prevent 86% of issues before they occur. Cross-stack analytics make infrastructure performance visible at the VM-level. It also provides a prescriptive upgrade path. HPE is now integrating InfoSight across all its data center products.
- Huawei OceanStor Dorado 8000 V6 and Dorado 18000 V6 storage systems incorporate dedicated Ascend AI chips that dynamically optimize the performance of many applications. The company also provides cloud-based analytics focused on array health.
- Pivot3 Dynamic QoS provides policy-based QoS management based on the business value of workloads. The system automatically applies a set of default policies, and dynamically enforces those policies. But administrators can change the policies and change which workloads are assigned to each policy on-the-fly.
- Pure Storage Pure1 META provides cross-stack analytics for rapid root-cause analysis, and sophisticated workload modeling that facilitates capacity planning.
- Tintri built Intelliflash Analytics on a self-optimizing analytics framework. It provides insight into data, VMs, and real-time performance analytics for the storage system. Tintri VMstore embeds intelligent analytics into its storage OS. It uses those capabilities to manage the performance of every workload dynamically, without human intervention.
Questions to Ask Prospective Storage Vendors
• How does the solution use AI and ML to avoid system downtime? Do you provide uptime guarantees?
• How does the solution apply AI and ML to optimize application performance?
• Does the solution provide visibility via cross-stack analytics?
• What tools does the solution offer that facilitate capacity planning?
• How does the solution automate infrastructure?
• What options are available for an administrator to approve or override the automation?