VDI · /// Capacity Planning
Stop guessing how many sessions your VDI estate can hold.
Measured concurrency, not spreadsheets. The same scenario engine runs on Citrix, AVD, Horizon, and RDS — capturing actual session-host pressure, FSLogix profile-load, HDX channel saturation, and connection-broker latency. Right-size compute, storage, and license counts with evidence.
Citrix · AVD · Horizon · RDSMeasured concurrency, not guesses±5% host-pool right-sizing
Live orchestration cockpit — measured concurrency across stacks.
The Problem
Spreadsheet sizing breaks first on production weekends.
Vendor-supplied calculators and spreadsheet sizing models give round numbers that look defensible — until a production peak proves them wrong. The components that actually break first hide behind per-host counters that no calculator models.
Pilot pools that lie.
A 50-user pilot proves nothing about a 1,500-user finance-close. Without measured concurrency on the actual session host — Citrix, AVD, or otherwise — capacity claims live on slides, not in production.
Azure / Citrix Cloud invoices don’t care.
Wrong session-host SKU, wrong host-pool count, wrong FSLogix share size — and the monthly invoice tells the truth. Spreadsheet sizing models can be wrong by ±40% on a real workload.
Multi-stack reality.
Most enterprises run more than one VDI stack — Citrix in one division, AVD in another, Horizon in a third, RDS for legacy apps. Capacity planning needs to compare them on the same scenario shape, not three different spreadsheets.
Why LoadGen for capacity planning
One scenario engine. Four VDI stacks. Measured.
Author a scenario once. Replay it on Citrix, AVD, Horizon, RDS — and capture real session-host pressure on each. The same .lgs workload produces real before / after numbers across stacks for direct comparison.
Measured concurrency, not modelled
Full and VDI agents run as actual Citrix / AVD / Horizon / RDS sessions — Activate, Reset, Kill operations land per-session on real session hosts. No thread-based abstractions.
Same scenario, four stacks
The same .lgs workload replays unmodified across Citrix HDX, AVD ARM, Horizon connection-broker, and RDS — with protocol-honest measurement on each.
SUT Monitoring on every host
FSLogix profile-load, AAD auth latency, Citrix Broker, StoreFront, Gateway, RDS Connection Broker counters — bound to every test execution. Capacity correlates with infra health.
Cross-stack validation
The same scenario shape, on every VDI stack.
Author once. Run on Citrix HDX, AVD ARM, Horizon connection-broker, RDS multi-session. Compare HDX p95 vs AVD-specific p95 vs Horizon connection-broker latency on one chart — measured, not modelled.
- Same .lgs scenario runs unmodified on Citrix, AVD, Horizon, RDS.
- Up to 5 runs overlaid on one chart for direct comparison.
- Protocol-honest measurement on each stack — HDX channel pressure, AVD-specific p95, connection-broker latency.
- Numbers feed straight into capacity sign-off — session-host SKU, FSLogix share size, host-pool count.
Multi-test overlay — Citrix vs AVD on the same scenario.
Capability
What ships with capacity-planning workflows.
Per-stack wizards
7-step wizards for Citrix (Basic ICA + Enhanced HDX) and AVD (ARM-native discovery). Horizon and RDS scenarios share the same authoring model.
Learn moreLive orchestration cockpit
vUsers ramp, per-step latency, host-pool density, session-host pressure — visible as the run progresses. Multi-region agents for honest geo-distributed concurrency.
Learn moreSUT Monitoring + infrastructure binding
FSLogix profile-load time, AAD authentication latency, Citrix Broker counters, RDS Connection Broker, Azure App Service status — bound to every test execution.
Learn moreMulti-test comparison overlay
Up to 5 runs overlaid on one chart — compare Citrix vs AVD, pilot vs production scale, week-over-week regression. Drill into Moments, Errors, and per-step deltas.
Learn moreAzure / Cloud cost correlation
Per-host-pool density measurements feed straight into the Azure / Citrix Cloud sizing decision — session-host SKU, FSLogix share size, host-pool count, license counts.
Learn moreEvidence-grade reporting
Every run captured in queryable history — exported to PDF / JSON for the capacity-planning sign-off. Numbers Finance can hand to procurement.
Learn morePeak validation
Spike testing for finance-close, year-end, open enrolment.
Measured before / after — the same scenario at growing scale shows where the host pool breaks first. Phase-by-phase validation between waves; cutover signs on data, not opinion.
- Warm-up / steady / spike / cool-down phases — visible as the run progresses.
- Per-step latency + p95 + error hotspots during the run.
- Compare modelled peak against measured peak — close the gap before production users do.
- Wire into release pipeline to block deploys on regression.
Spike simulation — measured peak vs modelled peak.
Outcomes
Measured capacity planning, not spreadsheet sizing.
Before
±40 %
After
±5 %
Before
7 days
After
4 hrs
Before
8
After
1
Before
Spreadsheet
After
One scenario
Where capacity planning runs
Three windows VDI teams actually plan for.
Pilot → production scaling
Stress the Citrix / AVD / Horizon / RDS pool against a measured concurrency model derived from the pilot. Right-size session-host SKU and license counts with evidence.
See use caseCross-stack migration validation
Compare current Citrix capacity against target AVD or Horizon capacity on the same scenario shape. Plan the cutover with data, not slide decks.
See use casePeak-event readiness
Finance-close, year-end, open enrolment, e-commerce peak — model the spike before the calendar lands on it. Block surprises with measured-peak validation.
See use caseRight-size your VDI estate with measured data.
We’ll author a scenario in the wizard on a call, fire it from Full or VDI agents against your Citrix / AVD / Horizon / RDS pool, and show you measured concurrency — not spreadsheet numbers.
Questions
Frequently asked.
Does LoadGen do capacity planning for Citrix, AVD, Horizon, AND RDS?
Yes. The same scenario engine drives all four. Author once, replay on each stack. Per-stack wizards capture protocol-specific signals natively — HDX / ICA on Citrix, ARM discovery + FSLogix on AVD, connection-broker on Horizon, multi-session host on RDS.
How accurate is the host-pool sizing?
Published target is ±5% host-pool right-sizing accuracy, compared against the ±40% typical of spreadsheet-sizing models. Numbers come from per-host measured concurrency, not modelled assumptions. Source: /platforms/avd OUTCOMES.
Can we compare Citrix capacity against AVD capacity in one engagement?
Yes — that’s the migration-testing primary use case. Same .lgs scenario runs unmodified on both stacks; the cockpit overlays HDX p95 against AVD-specific p95 for direct comparison. See /use-cases/citrix-to-avd-migration.
How does this work for peak-event capacity (finance-close, year-end)?
Spike-simulation phases — warm-up / steady / spike / cool-down — model the peak deliberately. Per-step latency and host-pool density visible as the run progresses; measured peak vs modelled peak gap closes before production users hit the calendar.
What does VDI capacity planning cost?
Load Testing module is €1,099 per week at the 50-vUser tier, scaling to 25,000 vUsers. Terms run from 1 week to 5 years. Capacity-planning engagements typically combine Load Testing with End-to-End Monitoring (€899 / Agent / month) for SLA tracking after sign-off.
What evidence do we hand to procurement?
Every run is captured in queryable history. LoadGen Insight exports per-host density, per-region concurrency, FSLogix profile-load, and per-step p95 to PDF / JSON for sign-off. Finance and procurement get measured numbers, not vendor calculators.
