At roughly what time does Mist's global (cloud) RRM run its daily optimisation?

Correct: b. Global RRM runs once daily, around 2-3 AM local time (exact minute auto, non-configurable), using the 30-day trend baseline. Real-time reactions are the separate local RRM layer; Optimize is a manual on-demand extra.

A carpeted Bengaluru office is all dual-band laptops/phones, no 2.4-GHz-only gear, and Capacity SLE is dragging. Which RF Template change gives the biggest capacity lift?

Correct: a. With no 2.4-only clients, letting RRM cancel surplus 2.4 GHz radios and convert some to a second 5 GHz radio (Dual 5 GHz on AP43/45/63) adds 5 GHz cells — pure capacity. 160 MHz everywhere would backfire (too few channels); 2.4-only and single-channel pins are nonsense.

You need 6 GHz disabled on AP24s but enabled on AP45s across 25 sites, managed from one place. What do you configure?

Correct: b. RF Templates are org-scoped and carry per-AP-model settings. Disable 6 GHz on AP24, enable on AP45, in one template, applied to all sites. WLAN templates handle SSID/policy, not radios; per-AP overrides don't scale.

An AP keeps getting moved off DFS channels by RRM and lands on crowded non-DFS channels. Logs show frequent radar hits on that AP. What's happening?

Correct: d. DFS punishment is intentional: RRM tracks which APs/channels see the most radar and steers the worst offenders away from them. Some crowding results — that's the lesser evil vs constant DFS channel jumps. The fix is to find/eliminate the radar source, not to fight RRM.

A previous admin set all 5 GHz radios to fixed max power "for stronger signal". Coverage is fine but throughput is poor and clients report stalls in open areas. Most likely root cause?

Correct: a. Cranking power maximises coverage but destroys capacity — overlapping cells share airtime and collide. Returning power to Automatic with a ceiling lets RRM shrink cells to reduce overlap, raising throughput. Classic coverage-vs-capacity inversion.

A vendor proposes "disable RRM entirely and manually pin every channel + power for full control". For a 600-AP multi-site enterprise, is this sound?

Correct: b. Full manual pinning kills the adaptive layer: no automatic DFS escape, no interference re-tune, no neighbour-offline self-heal, plus 600 hand-tuned radios to maintain. The professional pattern is guard-rails (RF Template: channel list, power range, band plan) + let RRM optimise inside them.

Management asks: "RRM hasn't changed anything in 4 days — is it broken?" Capacity SLE has been steady at 95%. What's the most defensible answer?

Correct: c. RRM optimises toward an outcome, not toward activity. Above ~90% Capacity SLE it deliberately holds steady — churn would risk client disruption for negligible gain. A quiet, healthy site is success. (You can still click Optimize after a physical change if you want a fresh pass.)

Juniper Mist RF & RRM — Auto RF, Coverage vs Capacity & RF Templates

Q: Sneha sees an AP change channel at 6:14 PM after a radar hit, but the dashboard shows the "RRM optimization" timestamp as 2:30 AM. How do you reconcile this?

Correct: a. Real-time channel jumps (DFS, interference, neighbour-offline) are local RRM and happen any time of day. The once-daily 2-3 AM stamp is the cloud's strategic baseline. Seeing both is exactly how a healthy site behaves.

Q: Priya at HCL runs 5 GHz at 160 MHz width in a packed open-plan floor with 40 APs. Capacity SLE is stuck at 68%. What's the most likely capacity killer?

Correct: a. Wider channel = higher per-client speed but fewer channels to reuse. With 40 APs packed together, 160 MHz forces massive co-channel overlap. Narrowing to 40 MHz multiplies the reuse pattern and lifts capacity. (And d is nonsense — 2.4 GHz is 20 MHz only.)

Q: You want 6 GHz off on AP24s but on for AP45s, across 30 sites, changed in one place. What's the right tool?

Correct: c. RF Templates are org-level and hold per-AP-model settings. Disable 6 GHz on AP24, keep it on AP45, in one template, applied to all sites. Change once, propagate everywhere — exactly the design intent.

First — the mistake almost every newcomer makes

Day one on Mist, an engineer opens Site → Radio Management, sees that two APs are on the same 5 GHz channel, panics, and manually pins every radio to a fixed channel and max power. "There, now it's deterministic." Two weeks later the help-desk drowns in "Wi-Fi slow" tickets.

Here is what that engineer got wrong: Mist RRM is not a one-time wizard. It is a living, learning system that re-scores your RF environment continuously. The moment you hard-pin everything, you switch off the very intelligence you paid for — and you freeze the network at one bad snapshot in time.

🧒 ELI5: Imagine a smart thermostat. If you tape the dial at one number forever, you've made it a dumb heater again. Mist RRM is the smart thermostat for your Wi-Fi — let it adjust, just teach it the rules.

🏛 Architect lens: Treat RRM as a closed-loop control system whose reward function is Coverage + Capacity + Connectivity. Your job is to constrain its search space (channel list, power floor/ceiling, band plan) — not to replace it. Over-constraint and zero-constraint are both anti-patterns.

So the right mental model is: you set the guard-rails (RF Template), Mist drives inside them (RRM). Let's build that model from the ground up.

Warm-up · before we start

Gut check: in Mist, who normally picks which 5 GHz channel an AP uses?

a) The admin, manually, per AP — always b) RRM, automatically, within the channel list you allow c) The client device decides d) The ISP

Correct: b. RRM auto-selects channel + width + power within the guard-rails you define in the RF Template. Manual per-AP pinning is the exception (and usually a mistake), not the default.

Warm-up · before we start

If your Capacity SLE already sits at 96%, what will the nightly global RRM most likely do?

a) Aggressively re-shuffle every channel for fun b) Reboot the APs c) Make few or no changes — there's little to gain above ~90% d) Increase power on all radios to max

Correct: c. When Capacity SLE is 90% or above, Mist judges there's not much to gain, so RRM holds steady. Stability beats churn. (We prove this in Section 4.)

Warm-up · before we start

A radar burst hits an AP on a 5 GHz DFS channel. Who reacts first?

a) Nobody until the 2 AM global run b) The local on-AP RRM — instantly jumps to a new channel c) The client roams to another SSID d) You must phone Juniper TAC

Correct: b. DFS is a real-time, safety-of-life requirement — the AP's local RRM moves channels immediately. The cloud only learns from it later. That global/local split is the heart of this lesson.

① The two-tier brain — global cloud vs local AP RRM

Mist runs RRM as two cooperating layers. Miss this and everything else stays fuzzy.

Global RRM lives in the cloud. It is a reinforcement-learning agent that crunches up to 30 days of long-term trend data per site. It runs once a day, around 2:00–3:00 AM local time (the exact minute is automatic and not configurable), and only pushes channel/power baselines down if a change is actually warranted.

Local RRM runs on each AP, in real time. It handles the acute events the cloud can't wait for. There are exactly three local triggers worth memorising — they show up on the JNCIA-MistAI blueprint:

Figure 1 — Mist's two-tier RRM: global cloud (strategy, nightly) + local on-AP (tactics, instant).

Sneha at Infosys Sneha runs a 4-floor campus on AP45s. At 6 PM a radar source nearby trips DFS on channel 100. The local RRM on the affected AP jumps to channel 36 within a second — users barely blink. That radar event is bundled into the AP's daily telemetry. At 2 AM, global RRM notices that AP has seen radar three nights running and quietly stops handing it DFS channels altogether. Two layers, one outcome: stable Wi-Fi.

What RRM is actually allowed to change

Whether global or local, RRM only has five knobs. Know them and you know the whole system.

📻

Channel

tap to flip

Automatic Channel Selection picks a cleaner channel within your allowed list — and remembers past co-channel interference, not just the current snapshot.

📶

TX Power

tap to flip

Raises or lowers transmit power — but only lowers power if coverage isn't hurt. The whole coverage-vs-capacity dance lives here.

↔️

Channel width

tap to flip

2.4 GHz = 20 MHz only. 5 GHz = 20/40/80. 6 GHz = 20/40/80/160/320 (country-dependent). Wider = faster but fewer non-overlapping channels.

🎨

BSS color + band

tap to flip

On 802.11ax APs, RRM tunes BSS color to reduce co-channel collisions, and can steer clients between 2.4 and 5 GHz bands.

Pause & Predict

It's 11 AM and you change the channel list in your RF Template, removing channel 100. When does that take effect on the radios?

RF-template changes (like a new channel list or channel width) push down fairly quickly — they don't wait for the 2 AM run. The nightly global pass is for optimisation based on long-term trends; an explicit config change to the template is applied as a config event. So expect the radios to honour the new list shortly after commit, not 15 hours later. (DFS and interference reactions are even faster — those are local.)

Quick check · Q1 of 10

Sneha sees an AP change channel at 6:14 PM after a radar hit, but the dashboard shows the "RRM optimization" timestamp as 2:30 AM. How do you reconcile this?

a) The 6:14 PM move was local RRM (real-time DFS); the 2:30 AM stamp is the global nightly optimisation. Both are normal. b) The dashboard clock is wrong c) The AP rebooted at 6:14 PM d) Global RRM ran twice that day

Correct: a. Real-time channel jumps (DFS, interference, neighbour-offline) are local RRM and happen any time of day. The once-daily 2-3 AM stamp is the cloud's strategic baseline. Seeing both is exactly how a healthy site behaves.

② Coverage vs Capacity — the one trade-off behind every knob

Every RF decision is the same tug-of-war. Turn power up and each AP covers more area (great coverage) — but cells overlap, clients hear two APs on the same channel, and you get co-channel interference (bad capacity). Turn power down and cells shrink, overlap drops, capacity rises — but push too far and you punch coverage holes.

This is why Mist's power default is Automatic, with a configurable min/max power range. You're not setting a fixed dBm — you're handing RRM a sandbox: "stay between this floor and this ceiling, optimise inside it." In a typical Indian office on 5 GHz you might allow, say, min 8 dBm to max 15 dBm and let RRM choose per-radio.

🧒 ELI5: Power is like how loudly each teacher speaks. Too loud and two classrooms hear each other and get confused. Too soft and the back row hears nothing. Mist finds the "just right" volume per room.

🏛 Architect lens: Coverage is a per-cell SNR floor; capacity is an aggregate airtime function. They conflict because both scale with cell size. Design the band plan for the dominant client density: carpeted office → capacity-led (5/6 GHz, lower power, 20/40 MHz); warehouse/outdoor → coverage-led (higher power, 2.4 GHz retained).

Figure 2 — Coverage vs capacity: the trade-off RRM resolves automatically within your power range.

Rahul at TCS Rahul inherits a TCS floor where a previous admin set every AP to max power "for strong signal". Throughput is awful. He opens the RF Template, sets 5 GHz power to Automatic with max 14 dBm, and lets the next nightly run trim it. Co-channel overlap drops, airtime frees up, Capacity SLE climbs from 71% to 92%. He changed almost nothing — he just stopped fighting RRM.

Channel width — the other capacity lever

Wider channels are faster per client but cost you non-overlapping channels. On 5 GHz in a dense Indian office, 40 MHz is usually the sweet spot; 80 MHz only if you have enough clean spectrum; 160 MHz almost never in carpeted high-density (too few channels, too much overlap). On 6 GHz you finally have the room for 80/160 MHz. On 2.4 GHz you have no choice — 20 MHz only, because there are barely three non-overlapping channels (1, 6, 11) to begin with.

Quick check · Q2 of 10

Priya at HCL runs 5 GHz at 160 MHz width in a packed open-plan floor with 40 APs. Capacity SLE is stuck at 68%. What's the most likely capacity killer?

a) 160 MHz leaves too few non-overlapping channels for 40 APs → heavy co-channel interference. Drop to 40 MHz. b) The APs are too new c) 160 MHz always increases capacity, so it must be a cabling fault d) She must enable 2.4 GHz at 160 MHz too

Correct: a. Wider channel = higher per-client speed but fewer channels to reuse. With 40 APs packed together, 160 MHz forces massive co-channel overlap. Narrowing to 40 MHz multiplies the reuse pattern and lifts capacity. (And d is nonsense — 2.4 GHz is 20 MHz only.)

③ RF Templates — your guard-rails, set once, reused everywhere

An RF Template is where you draw the sandbox RRM plays in. Critically, you build it at the organization level (Organization → Wireless → RF Templates) and then apply it at the site level. One template, many sites — change it once, every assigned site inherits it.

The killer feature: a single RF Template can carry per-AP-model settings. For example, disable the 6 GHz radio on every AP24 (which doesn't need it) while leaving 6 GHz enabled on every AP45 — all in one template.

Figure 3 — RF config hierarchy: org template → site → device profile → per-AP. The most specific override wins.

Auto-cancellation & auto-conversion (Dual 5 GHz)

Here's where Mist gets clever in dense offices. Two RRM features built on the Dual Band Radio Settings (options: 2.4 GHz, 5 GHz, or Auto):

Auto-cancellation — RRM disables surplus 2.4 GHz radios where neighbour coverage already blankets 2.4 GHz. Less 2.4 GHz noise, cleaner band.
Auto-conversion (Dual 5 GHz) — on AP43 / AP45 / AP63, RRM converts the dual-band radio to a second 5 GHz radio, so one AP serves two 5 GHz cells. Pure capacity win for high-density carpeted offices.

You enable the behaviour by setting Band Enabled → Auto on the 2.4 GHz settings in the RF Template. RRM then decides per-AP whether to keep 2.4 GHz, kill it, or convert it.

Figure 4 — Auto-conversion turns a surplus 2.4 GHz radio into a second 5 GHz radio, doubling 5 GHz capacity.

Pause & Predict

A warehouse has rugged barcode scanners that only support 2.4 GHz. Should you enable auto-conversion (Dual 5 GHz) there?

No. Auto-conversion removes 2.4 GHz capacity to add 5 GHz. If you have legacy 2.4-GHz-only clients (scanners, IoT sensors, old printers), killing 2.4 GHz radios will strand them. Auto-conversion is for carpeted offices where almost every device is dual-band. In the warehouse, keep 2.4 GHz — and probably lead with coverage, not capacity.

Quick check · Q3 of 10

You want 6 GHz off on AP24s but on for AP45s, across 30 sites, changed in one place. What's the right tool?

a) Manually edit each of the AP24s on all 30 sites b) You can't mix AP models in one template c) One org-level RF Template with per-AP-model settings, applied to all 30 sites d) Reboot the org

Correct: c. RF Templates are org-level and hold per-AP-model settings. Disable 6 GHz on AP24, keep it on AP45, in one template, applied to all sites. Change once, propagate everywhere — exactly the design intent.

④ "RRM won't move my radio" — the playbook

The most common Mist RRM ticket isn't a crash — it's confusion. "I expected RRM to change the channel and it didn't." Nine times out of ten, RRM is working correctly and the engineer's mental model is off. Walk this ladder.

▶ Watch the nightly global RRM decision

Click Play. Each stage lights up as the cloud decides whether to touch your radios.

① COLLECT Each AP scans all channels all day, scoring interference, client count + usage. Telemetry → Mist cloud.

▼

② NIGHTLY RUN ~2:00–3:00 AM local · RL agent rebuilds the 30-day baseline for the site.

▼

③ SLE GATE Is Capacity SLE ≥ 90%? If yes → make NO changes (little to gain). If no → continue.

▼

④ DFS LEARNING Restrict APs that see the most radar from their worst channels (DFS punishment).

▼

⑤ DECIDE Within your allowed channel list + power range, pick the channel/width/power that maximises Coverage+Capacity+Connectivity.

▼

⑥ PUSH + WATCH Apply new baseline to APs · then keep monitoring the Capacity SLE to confirm the change actually helped.

Press Play to step through a nightly global RRM cycle. Each Next advances one stage.

The #1 "RRM won't move" cause

Symptom you see: you stare at Radio Management for an hour and the channels never change, even though two APs look like neighbours. Cause: your Capacity SLE is already ≥ 90%. RRM judged there's nothing to gain and deliberately held steady — that's correct behaviour, not a bug. Stop expecting churn from a healthy site.

The #2 cause — you pinned it yourself

Symptom you see: one AP never re-tunes while its neighbours do. Cause: a leftover per-AP device-specific channel/power override from a past troubleshooting session. A manual pin disables RRM for that radio. Clear the override and let it rejoin the auto pool. (And if you must pin channel, pin power too — Figure 3.)

When you genuinely need RRM to act now — say after moving furniture or adding APs — don't wait for 2 AM. Go to Site → Radio Management, select the site and band, and click the Optimize button at the top-right to run RRM on demand.

Where to look in the dashboard (no CLI needed on Mist)

Site > Radio Management
  Site:  Lucknow-HQ            Band: [ 5 GHz ▾ ]
  AP45-Floor2   10.20.5.42     ch 149/80   14 dBm   neighbors: 2.6
  AP45-Floor3   10.20.5.43     ch 36/80    13 dBm   neighbors: 1.8
  [ Optimize ▸ ]   ← run RRM on demand for this site + band

Expected output (after Optimize)

RRM run queued for site Lucknow-HQ (5 GHz)
  AP45-Floor2  ch 149 -> 161   (co-channel reduced)
  AP45-Floor3  ch 36  -> 36    (no change needed)
  Capacity SLE: 84% -> monitoring for improvement...
  3 AP events logged · global baseline updates at next 2-3 AM run

Pro tips that separate L2 from L3

1. Don't pin channels to "fix" a one-off — let RRM learn the trend instead. 2. Set a sane power ceiling (e.g. 14 dBm on 5 GHz) so RRM can't shout over itself in dense areas. 3. Keep 160 MHz off in carpeted high-density; reserve it for 6 GHz or low-density. 4. Only enable auto-conversion where 2.4-GHz-only clients are truly absent. 5. If one AP keeps losing DFS channels, that's DFS punishment doing its job — investigate the radar source, don't fight the punishment.

Pause & Predict

You click Optimize and RRM still doesn't change a particular AP's channel. The other APs moved. What's your first suspect?

A per-AP device-specific override on that radio. A manual channel/power pin takes the AP out of RRM's control, so Optimize can't touch it. Check the AP's device-specific config, clear the override, then Optimize again. (Second suspect: that AP is on a DFS-punished channel set and its only "better" options are also restricted.)

🤖 Ask the AI Tutor

Tap any question — instant context-aware answer. No login, no waiting.

Pre-curated from Juniper Mist docs + community Q&A. For live prod issues, share your Radio Management snapshot + Capacity SLE timeline at chat.techclick.in.

🧠 Lock it in — explain it back

In one or two sentences, explain to a teammate why pinning every radio to max power often makes Wi-Fi worse, not better. Teaching it is the fastest way to remember it.

✓ We'll nudge you. Saved locally.

📖 Glossary — say these like an L3

RRM: Radio Resource Management — the automated channel/width/power optimiser.
Global RRM: Cloud layer, reinforcement learning, runs once daily ~2-3 AM on 30-day trends.
Local RRM: On-AP layer, real-time, handles DFS, interference, neighbour-offline self-heal.
Capacity SLE: Mist's measured capacity score; at ≥90% RRM makes little/no change.
DFS punishment: RRM restricting radar-prone APs off the channels they see most radar on.
Auto-cancellation: RRM disabling surplus 2.4 GHz radios where neighbours already cover 2.4 GHz.
Auto-conversion: RRM converting a 2.4 GHz radio into a 2nd 5 GHz radio (Dual 5 GHz) on AP43/45/63.
BSS color: 802.11ax tag that lets radios ignore same-coloured co-channel traffic, cutting collisions.

📚 Sources

Juniper Networks Docs — Radio Resource Management (RRM) (Mist Wireless): global vs local RRM, the five knobs, capacity-SLE 90% gate, DFS punishment. juniper.net/documentation
Mist Documentation — RF Templates: org-level config, site application, per-AP-model settings. mist.com/documentation
Mist Documentation — RRM: Auto Cancellation, Auto Conversion, and Dual 5 GHz. mist.com/documentation
Juniper Networks Docs — RRM Configuration Options & RRM Usage Examples (dual-band): Band Enabled → Auto, channel widths per band. juniper.net/documentation
Juniper Networks Docs — Radio Management: Site → Radio Management dashboard + on-demand Optimize button. juniper.net/documentation
Juniper Certification — Mist AI, Associate (JNCIA-MistAI / JN0-252) & JNCIS-MistAI-Wireless blueprints (RRM change types + timing). juniper.net/training/certification
DCLessons — Juniper Mist's Radio Resource Management (two-tier model field write-up). dclessons.com

What's next?

Now that the radios are tuned, we move up the stack to who gets to use them. Next lesson: building WLANs and WxLAN policy in Mist — templates, WPA3, labels and micro-segmentation done the cloud-native way.

Next · Mist WLAN & WxLAN Policy → ← All lessons

Juniper Mist RF & RRM — Auto RF, Coverage vs Capacity & RF Templates

Pick where to start

The two-tier brain

Coverage vs Capacity

RF Templates

"RRM won't move"

First — the mistake almost every newcomer makes

① The two-tier brain — global cloud vs local AP RRM

What RRM is actually allowed to change

② Coverage vs Capacity — the one trade-off behind every knob

Channel width — the other capacity lever

③ RF Templates — your guard-rails, set once, reused everywhere

Auto-cancellation & auto-conversion (Dual 5 GHz)

④ "RRM won't move my radio" — the playbook

▶ Watch the nightly global RRM decision

🤖 Ask the AI Tutor

📝 Wrap-up — six more

🧠 Lock it in — explain it back

📖 Glossary — say these like an L3

📚 Sources

What's next?