TTechclick All blogs
Palo Alto · Prisma Access · SASE
L2 / LATERAL ENGINEER

Zero Trust & Prisma Access Complete Guide

How Palo Alto Networks implements Zero Trust architecture across physical NGFW and cloud-delivered SASE — from concept to configuration. Covers the 6 pillars, policy decision logic, mobile-user flow, Strata Cloud Manager setup, HQ firewall IPSec config, and verification commands an L2/L3 security engineer actually runs.

👤 Techclick Infosec · 📅 March 2026 · ⏱ ~25 min read · 🏷 PAN-OS · Prisma Access · SCM

⚡ Quick Answer

Complete Zero Trust & Prisma Access guide: 6 pillars of Zero Trust, Palo Alto NGFW policy logic, Prisma Access SASE, SCM setup, HQ firewall config and

Pick where you want to start

1. Introduction

This document covers how Palo Alto Networks implements Zero Trust — both on physical NGFW and via Prisma Access (cloud-delivered SASE). It explains the architecture, configuration, and policy logic for a company with HQ, branch offices, and remote users.

💡Key Principle

Zero Trust is not a product — it is a security framework. Palo Alto implements it through a combination of features: User-ID, App-ID, HIP, SSL Decryption, DLP, WildFire, and Prisma Access working together.

2. Zero Trust Concept — Old Model vs Zero Trust

✗ Old Model — Trust But Verify
Inside network = trusted automatically
Perimeter firewall is the only control
Once inside LAN, move anywhere freely
VPN gives full network access
No device posture check
IP-based rules only — no user context
✓ Zero Trust — Never Trust, Always Verify
Every access request verified regardless of location
Microsegmentation — no lateral movement
Identity-aware policies (user + group)
Device posture checked before access
Least privilege — only what you need
Continuous re-verification — not one-time
Legend user / endpoint side (royal) Prisma Access / NGFW enforcement (cyan) destination — internet / private app (magenta) verified / allowed failed / denied
Old perimeter model vs Zero Trust — the same access request
Old perimeter model versus Zero Trust Two side-by-side flows. Left: old model — a user crosses one perimeter firewall and is then trusted to move freely across the flat LAN. Right: Zero Trust — every request is verified on identity, device posture and per-app policy before reaching only the one app it is allowed to use. OLD MODEL — trust but verify ZERO TRUST — never trust, always verify User VPN / on LAN Perimeter firewall Flat LAN — trusted, no inspection inside Finance HR ERP lateral movement — free User + device anywhere VERIFY EVERY REQUEST User-ID · HIP posture · App-ID · least-privilege policy SSL decrypt · DLP · threat prevention ERP — allowed this one app only HR — denied no lateral move every zone crossing inspected

In the old model a user who clears one perimeter firewall is trusted to move freely across a flat LAN. Under Zero Trust there is no trusted inside — every request re-proves identity, device posture (HIP) and per-app policy, so a compromised session reaches only the one app it is explicitly allowed and can't pivot to HR or Finance.

Quick check · Zero Trust concept

A laptop is already plugged into the corporate LAN at HQ. Under Zero Trust, what happens when it requests the ERP app?

Correct: b. Zero Trust assumes there is no trusted "inside". Every access request is verified on user, device posture and per-app policy and only ever reaches the specific app it's allowed — there's no flat LAN to roam, so lateral movement is gone.

🔑 Lock in the key terms — tap to flip

🛡️
Zero Trust
tap to flip

A security framework, not a product: "never trust, always verify". Every access request is authenticated, authorised and inspected regardless of location.

🌐
Prisma Access
tap to flip

Palo Alto's cloud-delivered SASE — the NGFW stack as a service. Remote users and branches connect to the nearest PoP; the same inspection runs there, no box per site.

💻
HIP
tap to flip

Host Information Profile. The GlobalProtect agent reports device posture (disk encryption, AV, patch level); policy only allows access when the device is compliant.

⚙️
SCM
tap to flip

Strata Cloud Manager — Palo Alto's cloud-native, single-pane manager for Prisma Access (Service Connections, Mobile Users, policy, commit & push).

3. 6 Pillars of Zero Trust in Palo Alto

iHeads-up

This is Techclick's 6-pillar framing. Palo Alto's official Zero Trust Enterprise whitepaper uses 5 pillars: User, Device, Application, Infrastructure, Data. We split 'Threat' out as a 6th for teaching clarity — when interviewing at Palo Alto, default to their 5.

All 6 pillars must pass for a session to be allowed. Failure in any single pillar results in deny or quarantine.

Pillar 01
User Identity
User-ID + SAML + MFA
Maps every IP to a real username via AD Security Event logs. Policies reference usernames/groups, not IP addresses. Step-up MFA enforced for sensitive resources.
Pillar 02
Device Trust
HIP Check via GlobalProtect
GlobalProtect agent reports device posture — AV installed, disk encrypted, OS patched, domain joined. Valid credentials alone are not enough — device must also pass HIP.
Pillar 03
Application Control
App-ID — Layer 7 inspection
Identifies actual application inside traffic regardless of port. Port 443 could be Office365, Dropbox, TeamViewer, or malware. Allow named apps only — deny everything else.
Pillar 04
Network Segmentation
Zones + Microsegmentation
Every zone crossing is inspected. Finance servers cannot reach HR servers without explicit allow rules. Attacker who compromises one host cannot move laterally.
Pillar 05
Data Security
SSL Decrypt + DLP
90% of traffic is encrypted — without SSL decryption you are blind. Decrypts, inspects for sensitive data (PII, credit cards, IP), re-encrypts. Blocks exfiltration even by trusted users.
Pillar 06
Threat Prevention
IPS + WildFire + DNS Security
Even authenticated, compliant users on allowed apps — traffic itself is scanned. WildFire sandboxes unknown files. DNS Security blocks C2 domains. IPS catches exploit patterns.
Quick check · The 6 pillars

A user has valid AD credentials and is in the allowed group, but their laptop's disk encryption is turned off. What happens?

Correct: b. All 6 pillars must pass for a session to be allowed. Device Trust (HIP) checks posture like disk encryption, AV and OS patch level — valid credentials alone are not enough, so a non-compliant device is denied or quarantined.

4. Zero Trust Policy Decision Logic

This is what Prisma Access evaluates for every single session — in order. All checks must pass.

👤Check 1 — User-IDIs "ram.dixit@corp.com" in allowed AD group?Pass ✓
💻Check 2 — HIP / DeviceDisk encrypted + AV running + OS patched?Pass ✓
📱Check 3 — App-IDApplication = web-browsing to ERP? Not TeamViewer?Pass ✓
🔒Check 4 — Zone PolicySource=Prisma-Zone → Dest=ERP-Zone policy exists?Pass ✓
📄Check 5 — DLPAny sensitive data (PII, credit card) being sent out?Pass ✓
Check 6 — ThreatAny exploit signatures, malware, C2 callbacks?Pass ✓
Result

All 6 checks passed → Session ALLOWED. If any single check fails → DENY or QUARANTINE. That is Zero Trust.

5. Why Prisma Access — Not Just Physical PA

Physical PA firewall protects your building. Prisma Access protects your users and data — wherever they are.

ScenarioPhysical PA OnlyPrisma Access (SASE)
Branch protectionNeed PA box at every branchIPSec tunnel — no hardware needed
Remote / WFH userVPN → backhauled to HQGP → nearest PoP (Mumbai for India)
SaaS inspectionTraffic hairpins through HQInspected in cloud before reaching SaaS
Policy managementPer-device or PanoramaStrata Cloud Manager — single pane
ScalingBuy more hardwareElastic — auto-scales
New branch setupShip hardware, rack, configureConfigure tunnel in SCM — done
CapExHigh — PA-3200/5200 seriesZero — OpEx subscription
Zero Trust coverageWhere hardware existsEverywhere — all users, all locations

6. Mobile User → Private App Flow

This is how a remote user in Pune accesses an internal ERP application at HQ Delhi — without a traditional VPN.

STEP 01
User opens laptop
GP auto-connects
STEP 02
SSL tunnel
To Mumbai PoP
STEP 03
IdP Auth + HIP
SAML + posture
STEP 04
Policy check
SCM security rule
STEP 05
IPSec svc conn
Prisma → HQ PA
STEP 06
Private app
LAN delivery
💡Pro Tip

The Service Connection is the bridge. One IPSec tunnel at HQ serves all remote users globally. You don't need VPN concentrators or per-user firewall rules at HQ. Prisma handles all user sessions in the cloud.

▶ Watch a remote user reach the HQ ERP through Prisma Access

A laptop in Pune opens the internal ERP at HQ Delhi. Press Play for the healthy path, then Break it to see the classic HIP / posture failure — and the fix.

① ConnectThe user opens the laptop. GlobalProtect auto-connects and builds an SSL tunnel to the nearest Prisma PoP (Mumbai for India).
② Auth + HIPThe IdP (Azure AD / Okta) authenticates the user via SAML, and the GlobalProtect agent submits its HIP report — disk encryption, AV, OS patch level.
③ PolicyThe SCM security rule is evaluated: source = mobile-user zone + AD group, HIP = compliant-devices, App-ID matches — all 6 pillars must pass.
④ Service ConnectionPrisma forwards the session over the IPSec Service Connection to the HQ PA firewall — one tunnel serves every remote user globally.
⑤ Private appThe HQ firewall delivers the session to the ERP on the LAN (10.10.0.0/16). The user reaches one app — no full network access, no VPN concentrator.
Press Play to step through the healthy path, then press Break it.
Quick check · Mobile user flow

In the flow above, what carries the remote user's session from the Prisma cloud into the HQ LAN to reach the private ERP?

Correct: c. The Service Connection is the bridge for private-app access: one IPSec tunnel from Prisma to the HQ PA firewall serves every remote user, so you need no VPN concentrators or per-user rules at HQ. DIA is for internet/SaaS, not for reaching the HQ ERP.

Mobile User Flow — Direct Internet Access (DIA)

The most common Prisma Access traffic pattern: user → Prisma Service Edge → direct to SaaS / internet (DIA), NOT backhauled through HQ. Security stack (App-ID, Threat Prevention, URL Filtering, WildFire, DNS Security) runs on the Service Edge in line. The HQ-backhaul flow (Service Connection → on-prem firewall) is for private app access only — Office365, Salesforce, web browsing all go DIA.

Strata Cloud Manager (SCM) vs Panorama with Prisma Access plug-in

Panorama with Prisma Access plug-in remains the dominant management surface for tenants migrating from on-prem PAN. SCM is Palo Alto's strategic direction (cloud-native, multi-product), but if you already operate Panorama at scale, the plug-in path is the lower-risk migration.

Cortex Data Lake (CDL) — where the logs live

Cortex Data Lake (CDL) is where all Prisma Access logs land. Sized in storage tier (1 TB, 10 TB, etc.) — separate license from Prisma Access. Every Prisma operator deals with CDL daily for log queries, AutoFocus integration, and Cortex XSOAR/XSIAM ingest.

Licensing — sized in Gbps, not user count

Prisma Access licensing model: sized in Gbps (peak bandwidth), NOT user count. Sales conversations always start with 'how many Gbps?' — internalize this number for any Prisma role interview.

HIP enforcement mechanism

HIP enforcement: Z-App's HIP report is evaluated against HIP match objects, which can be referenced in security policy as a match condition. Posture violation → policy fires the configured action (block, isolate, restrict). HIP report distribution is per-session; mid-session HIP changes don't tear down existing sessions.

Cert path for Prisma Access roles

Cert path: PCNSE → PCCSE (Prisma Cloud) for cloud-native security roles; PCNSE → PCSAE (Prisma Access SASE Engineer) for SASE/Prisma Access roles. Pick PCSAE if your job is Prisma Access operations.

7. SCM Portal Setup — Step by Step

FieldService Connection Value
NameHQ-Delhi-SC
RegionAsia Pacific — India South (Mumbai)
Subnets to advertise10.10.0.0/16 (your HQ LAN)
IKE VersionIKEv2
EncryptionAES-256-GCM
DH GroupGroup 20
Pre-Shared KeyGenerate strong PSK — save for HQ firewall

8. HQ Firewall IPSec Config — Palo Alto

After completing SCM setup, configure the IPSec peer on your HQ PA firewall.

💡Modern best practice

Certificate-based authentication for Service Connections (PAN-OS 11+) eliminates PSK rotation overhead. PKI: device cert issued by your internal CA, trusted on the Prisma Access side. PSK examples below are kept for legacy compatibility but new deployments should be cert-based.

IKE Crypto Profile
# Network → Network Profiles → IKE Crypto → Add
Name          : Prisma-IKE-Crypto
DH Group      : group20
Authentication: sha256
Encryption    : aes-256-gcm
Key Lifetime  : 8 hours
IPSec Crypto Profile
# Network → Network Profiles → IPSec Crypto → Add
Name          : Prisma-IPSec-Crypto
Protocol      : ESP
Encryption    : aes-256-gcm
Authentication: none (GCM is AEAD — auth is built-in)
DH Group      : group20
Key Lifetime  : 1 hour
IKE Gateway
# Network → Network Profiles → IKE Gateways → Add
Name              : Prisma-SC-GW
Version           : IKEv2 only
Interface         : ethernet1/1 (WAN interface)
Local IP          : 203.x.x.x (HQ WAN IP)
Peer IP           : 34.100.x.x (Prisma SC IP from SCM)
Authentication    : Pre-Shared Key
PSK               : YourStrongPSKHere
IKE Crypto Profile: Prisma-IKE-Crypto
IPSec Tunnel + Static Route
# Network → IPSec Tunnels → Add
Name              : Prisma-SC-Tunnel
Tunnel Interface  : tunnel.10 (zone = Prisma-Zone)
IKE Gateway       : Prisma-SC-GW
IPSec Crypto      : Prisma-IPSec-Crypto

# Static Route — return path for mobile users
Destination       : 100.64.0.0/10 (GP IP pool)
Next Hop          : tunnel.10

9. Verification Commands

Palo Alto CLI
# IKE Phase 1 status
show vpn ike-sa

# IPSec Phase 2 status
show vpn ipsec-sa

# Traffic counter through tunnel
show vpn flow tunnel-id <id>

# User-ID mapping check
show user ip-user-mapping all

# HIP report for a specific IP
show user hip-report ip <ip>

# Active GlobalProtect sessions
show global-protect-gateway current-user
TestExpected Result
GP agent connectsStatus = Connected, IP = 100.64.x.x assigned
ping 10.10.1.50 from laptopReply from HQ server
tracert 10.10.1.50Hops: 100.64.x.x → Prisma PoP → HQ LAN
show vpn ike-saState = ESTABLISHED, Peer = 34.100.x.x
SCM → Monitoring → TrafficUser session visible with App-ID resolved
Non-compliant device (AV off)HIP fails → access denied, remediation page

10. Zero Trust Coverage: Physical PA vs Prisma Access

PA FeatureZero Trust RolePhysical PAPrisma Access
User-IDVerify WHO — username not IPHQ onlyAll locations
GP + HIPVerify WHAT device — posture checkVPN usersAll users globally
App-IDVerify WHAT application — not just portHQ onlyAll locations
Zones + PolicyLeast privilege — no lateral movementWhere hardware existsCloud-wide
SSL DecryptionSee inside encrypted trafficHQ onlyAll users, all apps
DLPProtect what data is leavingLimitedFull SaaS + web
WildFire + IPSVerify traffic is not maliciousHQ onlyAll locations
💡Summary

Zero Trust in Palo Alto is not one toggle — it is the combination of all these features working together, enforced on every session, continuously re-evaluated. Prisma Access extends this enforcement to every user, every location, without requiring hardware at each site.

Quick Lab

Quick Lab (15 min): (1) Log into the Strata Cloud Manager demo or your Prisma Access tenant. Navigate to Service Connections — note auth method (PSK vs cert). (2) Open Cortex Data Lake; run a search for the past hour's threat events. (3) Build a HIP match object requiring disk encryption + AV updated within 7 days. Attach to a security rule. Verify by toggling the criteria on a test endpoint.

🤖 Ask the AI Tutor

Tap any question — instant, scoped to this lesson. The exact framing an interviewer wants to hear.

Pre-curated from this lesson + Palo Alto interview Q&A. For a live tenant issue, paste your show vpn ike-sa output into chat.techclick.in.

📝 Check your understanding

10 scenario questions — same depth you'll see in PCNSE / PCSAE interviews. Pick one answer per question. You need 70% (7 of 10) to mark this lesson complete on your profile.

Q1

An interviewer asks: "Is Zero Trust a product you can switch on in Palo Alto?" What's the most accurate answer?

Correct: (b). Zero Trust is a framework — "never trust, always verify" — not a product or a toggle. Palo Alto implements it through a combination of features working together on every session. (a) and (c) reduce it to one feature; (d) is wrong because Zero Trust applies to both physical NGFW and Prisma Access.
Q2

A developer says: "Just give me a VPN to the internal LAN." Why is per-application Zero Trust access the better answer?

Correct: (c). A classic VPN grants network-location-based trust — once connected you can reach anything routable. Zero Trust grants access per application, not per network: the user is verified, device posture checked, and they only ever reach the specific apps policy allows. (b) is false — VPNs do encrypt; (a)/(d) miss the architectural point.
Q3

Which feature verifies that a connecting laptop has disk encryption on, AV running, and the OS patched before it reaches a sensitive app?

Correct: (c). HIP = Host Information Profile. The GlobalProtect agent reports device posture; the firewall matches it against a HIP object so policy only allows compliant devices. A laptop with disabled disk encryption can be blocked or quarantined. User-ID is "who", App-ID is "what app", WildFire sandboxes files.
Q4

On the same port 443 you must allow Microsoft 365 but block TeamViewer. Which Palo Alto capability makes that possible?

Correct: (a). App-ID performs Layer 7 inspection — port 443 could be Office365, Dropbox, TeamViewer or malware. App-ID identifies the real application regardless of port so you allow named apps and deny everything else. A port rule (b) can't tell those apart on the same port.
Q5

Roughly 90% of traffic is encrypted. Why is SSL/TLS decryption essential to a Zero Trust deployment?

Correct: (d). Most traffic is encrypted, and you can't verify what you can't see. SSL decryption lets the firewall open the session, inspect it for threats, exfiltration and disallowed apps, then re-encrypt. Sensitive categories (banking, health) are typically exempted for privacy — so (c) is backwards.
Q6

A 12-person remote sales team needs the same protection as HQ but you can't ship a firewall to each home. What's the right fit?

Correct: (b). Prisma Access is best for remote users and branches where a firewall per site is impractical — inspection happens in the cloud and remote users connect via GlobalProtect to the nearest PoP (Mumbai for India). On-prem NGFW stays at HQ / data centre; you run both under one policy model.
Q7

When you create the HQ Service Connection in SCM, which crypto settings match the lesson's recommended IPSec config to the HQ PA firewall?

Correct: (b). The lesson's Service Connection uses IKE Version IKEv2, Encryption AES-256-GCM and DH Group 20 — matched by the HQ firewall's IKE/IPSec Crypto profiles. With GCM (an AEAD cipher) the IPSec authentication can be "none" because auth is built in. Legacy IKEv1/3DES/DES choices are weak and not recommended.
Q8

After bringing up the tunnel, which CLI command on the HQ PA firewall confirms IKE Phase 1 is established with the Prisma peer?

Correct: (c). show vpn ike-sa reports IKE Phase 1 status — expected state ESTABLISHED, peer = the Prisma SC IP. show vpn ipsec-sa checks Phase 2. (a) reads device posture, (b) lists GP sessions, (d) checks User-ID mappings — none confirm Phase 1.
Q9

In a flat network an attacker who lands on one host can pivot to others. How does Palo Alto's microsegmentation stop that east-west movement?

Correct: (a). Microsegmentation puts least-privilege policy between workloads. On Palo Alto every zone crossing is inspected with App-ID/User-ID-aware rules, so even traffic inside the data centre is allowed only where explicitly needed — a compromised host has nowhere to spread.
Q10

Your team already runs Panorama at scale on-prem and is adopting Prisma Access. The lesson's guidance on management surface is:

Correct: (d). Panorama with the Prisma Access plug-in remains the dominant management surface for tenants migrating from on-prem PAN — the lower-risk path if you already operate Panorama at scale. SCM (Strata Cloud Manager) is Palo Alto's strategic, cloud-native, multi-product direction.
Lesson complete — saved to your profile.
Almost! Review the sections above and try again — you need 70% (7 of 10) to mark this lesson complete.

Frequently Asked Questions

The Zero Trust and Prisma Access questions that come up in interviews, design reviews, and the first week on a SASE project.

Is Zero Trust a product you can buy?

No. Zero Trust is a security framework, not a single product or a toggle. Its core idea is "never trust, always verify" — every access request is authenticated, authorised, and inspected regardless of where it comes from.

Palo Alto implements Zero Trust by combining features: User-ID (who), App-ID (what), zones + least-privilege policy, SSL decryption, HIP device posture, DLP, and WildFire/IPS — all enforced on every session, continuously re-evaluated.

What is the difference between Zero Trust and a traditional VPN?

A classic VPN drops the remote user onto the network — once connected, they can often reach anything that's routable, with no per-application control. That's implicit trust based on network location.

Zero Trust (and ZTNA specifically) grants access per application, not per network. The user is verified, the device posture is checked, and they only ever reach the specific apps policy allows — there is no "inside the LAN" to move around in.

What is Prisma Access?

Prisma Access is Palo Alto's cloud-delivered SASE — effectively their next-gen firewall stack running as a service in the cloud. Remote users and branch offices connect to the nearest Prisma Access location, and the same security inspection (App-ID, threat prevention, decryption, DLP, URL filtering) runs there instead of on a box at each site.

It extends one consistent Zero Trust policy to every user and location without shipping hardware to each one.

Prisma Access vs an on-prem NGFW — when do I use which?
  • On-prem NGFW: best at HQ / data centres where you control the hardware and need line-rate inspection of local and north-south traffic.
  • Prisma Access: best for remote users and branch offices, where standing up and maintaining a firewall per site is impractical.

In practice most enterprises run both and manage them with one policy model (often via Panorama / Strata Cloud Manager), so a user gets the same protection at the office, at a branch, or working from home.

What is HIP in Palo Alto?

HIP = Host Information Profile. The GlobalProtect agent reports the device's posture — is the disk encrypted, is antivirus running and up to date, is the OS patched, is a required process present — and the firewall matches that against a HIP object.

You then write policy that only allows access when the device is compliant. A laptop with disabled disk encryption can be blocked or quarantined before it ever reaches a sensitive app. This is the "verify the device" pillar of Zero Trust.

How does microsegmentation stop lateral movement?

In a flat network, once an attacker lands on one host they can pivot freely to others (east-west movement). Microsegmentation puts least-privilege policy between workloads — segment A can talk to the database it needs and nothing else.

On Palo Alto this is enforced with zones + App-ID/User-ID-aware rules, so even traffic inside the data centre is inspected and allowed only where it's explicitly needed. A compromised host has nowhere to spread.

What role does SSL decryption play in Zero Trust?

Most traffic is encrypted, and you cannot verify what you cannot see. SSL/TLS decryption lets the firewall (or Prisma Access) open the session, inspect it for threats, data exfiltration, and disallowed apps, then re-encrypt it.

Without decryption, "always verify" breaks down — App-ID, threat prevention, and DLP would all be blind to anything inside HTTPS. Sensitive categories (banking, health) are typically exempted for privacy and to avoid breaking certificate-pinned apps.

What are the pillars of Zero Trust in Palo Alto?

The commonly taught set, all working together on every session:

  • User-ID — verify the user/group, not just an IP.
  • App-ID — verify the application, not just the port.
  • Zones + least-privilege policy — no lateral movement.
  • SSL decryption — see inside encrypted traffic.
  • DLP — control what data leaves.
  • WildFire + IPS — verify traffic is not malicious.
Does Prisma Access need hardware at every site?

No — that's the point of it. Because inspection happens in the cloud, branches connect via Service Connections / IPSec tunnels and remote users connect via the GlobalProtect agent. There's no firewall appliance to buy, rack, patch, or refresh per location.

You still keep on-prem NGFWs where it makes sense (HQ, data centre), but new branches and a growing remote workforce don't each need their own box.

What's the difference between User-ID and App-ID?

User-ID answers who — it maps an IP/session to a real identity (and AD group) so policy can say "Finance group may use this app". App-ID answers what — it identifies the actual application in the traffic regardless of port or evasion, so policy can allow "Microsoft 365" while blocking "BitTorrent" on the same port 443.

Together they replace the old "IP + port" rule with an identity-aware, application-aware rule — the foundation of a Zero Trust policy.

Keep going →

What's next?

Go deeper on the model — how ZTNA actually replaces the VPN, broker by broker, policy by policy.

Next · ZTNA Explained: Zero Trust Network Access That Finally Kills the VPN →Practice on exam.techclick.in →