Infra Estimation Tool

Infra Estimation Tool helps users plan a migration before it begins. Based on the size and composition of the dataset to be migrated, along with the network and infrastructure planned for the migration, the tool predicts the expected migration duration and recommends an infrastructure configuration to support it.

Estimates are generated by a machine learning model trained on real, completed Egnyte migrations and are refreshed over time, so predictions reflect current platform and migration agent performance rather than a fixed formula.

Benefits and Use Cases

• Realistic migration timelines: Users receive a data-driven duration estimate instead of a rough guess, making it easier to plan cutover windows and set expectations with stakeholders.
• Right-sized infrastructure: The tool recommends the number of virtual machines, CPU cores, and memory required, helping avoid over- or under-provisioning.
• Bandwidth planning: Users can see the recommended upload speed for a dataset and understand how the available bandwidth affects the timeline.
• Scenario comparison: Upload speed, read speed, and VM count can be adjusted and the estimate re-run to see the impact before committing resources.
• Re-planning when scope changes: If the dataset grows or the provisioned speeds change, the estimate can be refreshed to produce an updated timeline.

Best Practices

• Enter accurate speeds: Upload speed and drive read speed are the strongest drivers of the estimate. Measured values produce more reliable results than nominal ones.
• Treat the result as a planning estimate, not a commitment: A buffer should be built into the plan, particularly for very large or unusual datasets.
• Use the tool to find the right trade-off: More VMs typically shorten the migration but increase cost; a higher upload speed improves the timeline but may be limited by the source read speed.
• Re-run the estimate after any material change: to the dataset, provisioned speeds, or VM configuration.
• Sanity-check atypical workloads: For datasets dominated by very large files or with an unusual file-to-folder ratio, the estimate should be compared against previous similar migrations.

Configure Infra Estimation

1. In the Migration App, navigate to Migration Assistance and select Infra Estimation from the left panel.

   

2. In the Migration Infra Estimation panel, enter the dataset details and click Submit:
   • Total size of data: the total volume of data to be migrated, with the appropriate unit (for example, TB).
   • Total number of files: the total count of files in the dataset.
   • Total number of folders: the total count of folders in the dataset.
     

3. Review the generated estimate on the canvas. The tool displays:

   • Mapped Drive Card: the source system, with its Drive Read Speed.
   • Virtual Machine Card(s): the recommended number of VMs, each with suggested CPU Cores and RAM.
   • Domain card: the recommended Upload Speed and the predicted Duration of the migration.

   

4. Adjust the parameters to explore alternative scenarios. The predicted duration updates to reflect each change:

   • Edit Drive Read Speed: on the Mapped Drive card if the source storage is faster or slower than initially assumed.
   • Edit Upload Speed: on the Domain card to compare different bandwidth options, for example 1 Gbps against 10 Gbps.
   • Edit CPU Cores or RAM: on a Virtual Machine card, or use Add VM and the remove (×) control to change the number of VMs, for example to compare 4 VMs against 8.

   

5. Users can take several actions for the infra estimation tool:

   • Zoom In: Magnifies the canvas so the Mapped Drive, Virtual Machine, and Domain cards appear larger. Useful when working with a specific card’s values.
   • Zoom Out: Reduces the canvas magnification to show more of the layout at once. Helpful when several Virtual Machine cards have been added and the full configuration no longer fits on screen.
   • Fit View: Automatically adjusts the zoom level and position so that all cards on the canvas are visible within the current window. A quick way to re-center the layout after zooming or panning.
   • Lock or Unlock: Determines whether the values within the cards can be edited. When locked, the fields on the Mapped Drive, Virtual Machine, and Domain cards (such as Drive Read Speed, CPU Cores, RAM, and Upload Speed) are read-only, which prevents accidental changes while reviewing an estimate. Unlock to edit the values and re-run different scenarios.

   

6. Click on Configure and update the dataset size, file count, or folder count to model a change in scope. Click on Submit to view the updated infra estimation. 
   
7. Admins can click on + Add VM to add virtual machine. Information (i) icon provides additional information. Repeat the adjustments until the estimate meets the required timeline and budget.
    

Limitations

• Designed for synchronous migrations: The estimate is modelled on synchronous upload migrations, which the tool is built to support. Asynchronous migrations should be planned using other methods.
• Based on dataset totals: The tool uses the total size, file count, folder count, and speeds to derive an average file size. It does not use a detailed breakdown of individual file sizes, so datasets made up of a small number of very large files may see some variation between the estimate and the actual duration.
• Most precise for datasets with smaller average file sizes: The tool is calibrated on migrations where the average file size is in the single-digit MB range, which reflects the majority of migrations. Datasets with a significantly larger average file size can still be estimated, though users should allow for a wider margin when planning.
• Provides a single expected duration: The tool returns one estimated duration rather than a minimum-to-maximum range. For very large migrations that run over an extended period, actual duration may vary more from the estimate, and users should plan accordingly.
• Relies on the accuracy of the values entered: The estimate is only as precise as the dataset details and speeds provided. Entering measured upload and read speeds, rather than approximations, produces the most reliable result.