Introduction

IBM, as a trusted leader in storage solutions, offers one of the most reliable and high-performance platforms in the industry — IBM Storage FlashSystem. One of the standout features of IBM Storage FlashSystem is its capability for intelligent workload management through the IBM FlashSystem grid. Within a FlashSystem, workloads are organized into units called Storage Partitions. Each system can host multiple Storage Partitions that operate independently, ensuring workload isolation and efficiency. The FlashSystem grid takes this a step further by enabling Storage Partition migration — the seamless movement of workloads between systems in the same grid.

As part of the day to day responsibilities of a storage administrator, a critical responsibility is to maximize both the performance and utilization of the systems under their management. Achieving this involves analysing system behaviour, monitoring performance trends, and carefully studying the requirements of individual Storage Partitions to determine the optimal placement of workloads.

IBM Storage insights AI-driven “Workload Placement Advisor for Storage Partitions” plays a crucial role in simplifying storage management. It provides insights based on historical and forecasted data that help storage owners and planners make informed decisions with ease.

The advisor addresses key questions that typically arise once a storage system is deployed and actively used:

1. Can I balance workloads across systems?
   How do I distribute workloads effectively to avoid hotspots and ensure consistent performance?

2. Am I maximizing storage capability and availability?
   Can I reassign workloads from heavily stressed systems to those with untapped capacity, ensuring both efficiency and resiliency?

3. Which system is most suitable for a given workload?
   How can I quickly identify the best placement option in a storage grid to support existing and future demands?

By delivering clear answers to these questions, the Workload Placement Advisor empowers administrators to optimize performance, utilization, and capacity planning while significantly reducing manual effort.

Feature Introduction

Workload Placement advisor for Storage Partitions is built on top of the existing AI powered feature called Workload placement advisor for new workloads, which deals with assigning a new customizable workload to a suitable system. Refer here for more information.

This feature provides users with the ability to evaluate the capabilities of IBM Storage FlashSystems—whether across the entire IBM FlashSystem grid or within a selected set of systems. It helps determine how well these systems can accommodate an existing workload (Storage Partition) if moved from one system to another.

The evaluation process considers multiple factors to ensure accuracy and relevance:

• Storage Partition (existing workload) requirements – both current and projected needs.

• System KPIs – capacity and performance indicators, along with their behaviour and potential impact.

• Threshold values for KPIs – realistic limits that each system can sustain for long-term stability.

Using these inputs, the advisor analyses all available systems and calculates behavioural scores. It captures the behaviour of key performance indicators (KPIs) over time, assess system sustainability under different workload conditions, and derive fitment scores for each KPI that quantify how well a system can accommodate a Storage Partition with a weighted score called compatibility score.

It takes the partition historical performance data to prepare the workload. Capacity and IOPS forecast data of the partition are used to derive forecast for system with that partition for the selected time period. Each system already has the forecast data for the KPIs available. STORM is used to measure the system’s sustainability against the performance requirement. With the storm provided benchmarks, derived forecasts ,and performance values, the max values and fitment score is computed for each KPI.

Feature Walk-Through

This capability is available within IBM Storage Insights and can be accessed through two different navigation paths:

• From the legacy/classic dashboard, it is listed under the “Insights” menu.
  

• From the new Carbon-based experience, it is part of the “Planning” features.
  

Once inside, simply click on “Workload Placement Advisor” to open the feature page. By default, the advisor opens in the “New Workload” mode, which      provides guidance for placing workloads that are being introduced for the first time.

For evaluating existing workloads (Storage Partitions), users can easily switch to the “Storage Partitions”option. This allows them to leverage the advisor specifically for optimizing and rebalancing workloads already running in their environment.

How to Use the Workload Placement Advisor for Storage Partitions

The new Workload Placement Advisor for Storage Partitions gives users a guided experience to evaluate and optimize existing workloads. Here’s how it works step by step:

1. Select a FlashSystem grid
   Since this feature operates within a grid, the first step is to choose the IBM FlashSystem grid that contains the systems you want to evaluate.
   
   
2. Choose a Storage Partition
   The UI will display all available Storage Partitions across systems in the selected grid. From this list, the user selects the partition to be evaluated.
3. Define target systems
   

   Under “Set target storage system”, all systems in the grid are automatically selected except the one currently hosting the chosen Storage Partition.

   These chosen systems will serve as target candidates for the fitment evaluation.

   • Users can proceed with all of them, or

   • Narrow the scope by manually selecting specific systems
4. Run the evaluation
   After clicking “Evaluate”, the advisor initiates a fitment evaluation and scoring process in the background. Users are free to continue navigating elsewhere.
   
   • Once complete, a toast notification appears for the task, labelled “Storage Partition Advisory”.
     
   • By clicking “View Details” on the notification, users can review the full results.
   • If the user stays on the evaluation page, the results automatically refresh when the task completes.

Understanding the Results

The results closely resemble the Workload Placement Advisor for new workloads, but tailored for existing workloads:

• Target storage systems are listed, showing details such as:
  • Storage system name
  • System family
  • Compatibility score → an aggregated score across all KPIs
    
    
• By expanding each system, users gain deeper insights into the evaluation, including
  • Fitment score per KPI
  • Graph plots of historical data
    • Graph plots of Forecast data (with and without the selected Storage Partition)
      
      

      The KPIs evaluated include:

      1.          Capacity

      2.          Total IOPS

      3.          Latency (response time)

      4.          Data rate (bandwidth)

      5.          CPU utilization

This comprehensive view allows users to make confident, data-backed decisions about where to best host their workloads for long-term efficiency and performance.

Why It Matters

This evaluation process provides better visibility, deeper understanding, and clear explanations to storage administrators, enabling them to make confident choices when selecting target systems for Storage Partition migration.

Moreover, the feature isn’t limited to the UI in IBM Storage Insights:

• It is also exposed via REST APIs, enabling storage owners to integrate AI-driven recommendations into their own workflows.
• Within Element Manager, when planning a Storage Partition migration inside an IBM FlashSystem grid, the same recommendations are surfaced through the API, making system selection easier and more reliable.

With continuous improvements and future enhancements, the Workload Placement Advisor for Storage Partitions is expected to become just as indispensable as the advisor for new workloads—further simplifying planning, boosting efficiency, and helping organizations get the most from their FlashSystem investments.

References:

1. Workload advisory documentation in Storage Insights
   https://www.ibm.com/docs/en/storage-insights?topic=resources-evaluating-workload-placement-advice
2. For details on IBM Storage Insights, please refer our product page https://www.ibm.com/products/storage-insights
3. For a hands-on experience with IBM Storage Insights, please use our demonstration link https://demo.insights.ibm.com/