Security, Specialist (JNCIS-SEC) — Free Practice Questions
10 free sample questions from a bank of 10, with the correct answers and explanations. No signup required — start practising right now.
1In Juniper high availability (HA) SRX Series device implementations, which interface will be used to exchange session state, configuration files, and ensure session continuity across nodes using the proprietary Trivial Network Protocol?
fab
fxp0
fxp1
swfab
Answer: A
The short version
A — fab. On SRX chassis clusters the fab interface is the data-link between the two nodes; it carries session state and configuration synchronisation so sessions survive a node failure, and the SRX implements that link with TNP.
Key concepts
Chassis cluster: two SRX nodes sharing one management plane so a single node failure does not interrupt service; configured with set chassis cluster cluster-id <n> node <n> plus redundancy-groups.
Control-link (fxp0): the node-to-node control connection running the Junos-level protocol for health monitoring, configuration synchronisation and redundancy-group state.
Data-link (fab): the high-speed link that synchronises session/flow state and, on SRX, the Junos configuration, so the surviving node can continue existing sessions.
TNP (Trivial Network Protocol): the proprietary lightweight protocol SRX uses over the data-link to replicate session state and configuration between cluster nodes.
swfab / fabric-link: the equivalent fabric link used on switches and MX virtualisation chassis, not the SRX chassis-cluster session-state link.
Why A is correct
The question describes two jobs — exchanging session state and configuration files — and mentions TNP. On SRX that traffic goes over the fab interface, the chassis-cluster data-link. TNP is the protocol Junos runs on that link to replicate the session (flow) table and configuration between the node in chassis and its peer, which is what delivers session continuity: after a failover the surviving node knows the existing sessions rather than dropping them. You can see the link and its state with show chassis cluster status, where the fab link health is reported alongside fxp0.
Why the others are wrong
B. fxp0 is the control-link. It runs the Junos-level control protocol for health/failover signalling and configuration state; it is not the high-speed TNP session-state data link the question describes.
C. fxp1 is not the chassis-cluster data-link on SRX. It is an internal/management-style interface and plays no role in TNP session-state synchronisation between cluster nodes.
D. swfab is the fabric link used on switch and MX virtualisation-chassis platforms (for example split-detection and virtual-chassis traffic). Naming it here confuses the switch-platform fabric interface with the SRX chassis-cluster data-link.
JN0-336 exam tip
Chassis-cluster links in one line: fxp0 = control-link (health, config, RG state), fab = data-link (session state, TNP). Remember fab from "session state flows across the fabric". Marking both as wrong: fxp1 and swfab are neither of these.
2What are two ways that Juniper Secure Connect provides flexibility in connection and authentication methods while ensuring that remote users are able to securely access company servers and cloud resources? (Choose two.)
It uses a persistent agent.
It uses Kerberos authentication.
It uses external authentication.
It uses an SSL VPN.
Answer: C, D
The short version
C and D. Juniper Secure Connect gives flexibility by terminating the remote session over SSL VPN (so users connect from a browser or client to servers and cloud resources) and by delegating authentication to an external source such as SAML, RADIUS or LDAP.
Key concepts
Juniper Secure Connect: cloud-delivered remote access (Secure Connect / Connect Secure) providing SSL VPN to internal servers, SaaS and cloud resources for remote and mobile users.
SSL VPN: the connection technology; clientless (browser-based HTML5 portal or plugin) or via the Secure Connect client, tunnelling the user's session to protected resources.
External authentication: delegating the authentication decision to an external identity source — SAML 2.0 identity provider, RADIUS or LDAP directory, or certificate validation — instead of local accounts.
Persistent agent / GlobalProtect always-on: a Juniper firewall (GlobalProtect) or Pulse Secure client feature that keeps the tunnel always connected; it is one specific connection mode, not the flexibility mechanism in question.
Kerberos: a Windows/LAN ticket-based protocol that is not one of the documented remote access or authentication methods the item is describing.
Why C and D are correct
C. External authentication removes the need to manage credentials on the appliance itself. The organisation reuses its existing IdP (SAML), directory (LDAP) or RADIUS server, and policies such as MFA, group mapping and conditional access are enforced centrally — that is exactly the flexibility described.
D. SSL VPN is the connection side of the same flexibility: the same policy can be reached from a browser with no install, from the Secure Connect client, from a mobile device, or from a partner/third-party connection, and the tunnel can terminate onto on-premises servers as well as cloud/SaaS resources.
Why the others are wrong
A. A persistent agent is the always-on, always-connected client approach (associated with GlobalProtect or Pulse Secure deployments). It is one narrow connection style, not the general mechanism that makes connection and authentication flexible in Secure Connect.
B. Kerberos is a Windows/domain ticket protocol and is not the pair of mechanisms this question describes; Secure Connect's documented flexibility comes from the SSL VPN connection technology plus external authentication sources (SAML, RADIUS, LDAP, certificates).
JN0-336 exam tip
Split the question into its two halves: "connection" -> SSL VPN, "authentication" -> external source (SAML/RADIUS/LDAP). Kerberos and persistent agent are the two classic lure options because both are real technologies in adjacent products.
3What are three policy types available in Junos Space Security Director? (Choose three.)
device
local
group
universal
global
Answer: A, C, E
The short version
A, C and E — device, group and global. Junos Space Security Director organises policies into three scopes: a policy applied to one device, a policy applied to a group of devices, and a policy applied globally to the entire security fabric.
Key concepts
Security Director policy scopes: Global, Group and Device — the three scopes Security Director uses to stage, distribute and audit policy across managed SRX devices.
Global policy: written once and distributed to every managed device; used for organisation-wide settings such as standard security policies, NAT and DNS policy.
Group policy: applied to a defined set of devices that share a role or site; useful for tiered rollouts and inheritance.
Device policy: applied to a single managed device, for exceptions or device-specific settings.
Security Director workflow: author in GUI or API -> analyse (edit/audit/compare, policy analysis) -> distribute (with staged push) -> verify and roll back.
Why A, C and E are correct
A. Device scope applies a policy to one managed device — used for single-device changes and exceptions.
C. Group scope applies a policy to a set of devices, which is how a common baseline or a site/tier configuration is pushed once instead of N times.
E. Global scope is the widest: the policy is distributed to all devices managed by Security Director, making it the natural place for organisation-wide policy such as default security policies or shared address books.
Together these three are the complete set of policy scopes in Security Director, which is why a "choose three" item selects exactly them.
Why the others are wrong
B. "Local" is not one of the Security Director policy-scope names. A policy configured on the device itself is a device policy from Security Director's point of view, so the term is simply not a scope in the product.
D. "Universal" is likewise not a Security Director scope name. It is not used for global, group or device policy distribution, and confusing it with "global" is the trap in this item.
JN0-336 exam tip
Security Director scopes, widest to narrowest: Global -> Group -> Device. Both wrong options are invented words that sound like the right ones — "universal" for global and "local" for device — so recognise the three real names cold.
4You are asked to configure a cluster between SRX1 and SRX2.
Which two commands must be used to accomplish this task? (Choose two.)
user@SRX2# set chassis cluster cluster-id 0 node 1
user@SRXl> set chassis cluster cluster-id 1 node 0
user@SRX2> set chassis cluster cluster-id 1 node 1
user@SRXl# set chassis cluster cluster-id 0 node 2
Answer: B, C
The short version
B and C. Both nodes must share the same cluster ID while using different node IDs: SRX1 gets cluster-id 1 node 0 and SRX2 gets cluster-id 1 node 1.
Key concepts
cluster-id: the identifier that groups nodes into the same chassis cluster; it must be identical on both nodes, and it determines the behaviour of the control- and data-links.
node-id: the unique number identifying each physical node inside the cluster — 0 or 1 on SRX; node 0 is the primary at first boot.
RTO vs switch (fabric) mode: chassis clusters use RTO mode by default; the fabric link is configured separately from the control-link.
Redundancy groups: after the cluster forms, devices and interfaces are assigned to redundancy groups (node 0/node 1) to define what fails over where.
Verification: show chassis cluster status must show both nodes, and each interface must be assigned to a redundancy group for failover to work.
Why B and C are correct
B. SRX1 is configured as cluster-id 1 node 0 — cluster ID 1, node number 0.
C. SRX2 is configured as cluster-id 1 node 1 — the same cluster ID 1, but a different node number, 1.
That pair satisfies both rules of chassis clustering: nodes agree on a common cluster ID (which is what lets them form the cluster and agree on RTO/fabric behaviour), and each node presents a unique node ID (which is how they know which is which and where to put each interface in the redundancy groups). Because node 0 is the primary at initial boot, SRX1 takes primary and SRX2 becomes the backup until a failure triggers failover.
Why the others are wrong
A.cluster-id 0 node 1 on SRX2 would give SRX2 a cluster ID of 0 while SRX1 uses 1. Nodes with different cluster IDs never form a chassis cluster, so this can never be part of a valid pair.
D.cluster-id 0 node 2 fails on two counts: the cluster ID does not match the other node, and node 2 does not exist on an SRX chassis cluster — only node 0 and node 1 are valid.
JN0-336 exam tip
Chassis cluster config in three words: same cluster-id, different node-id, only 0 or 1. Any option that breaks one of those three rules (0 vs 1 for the cluster, or node 2) is instantly wrong.
5You need to secure communications from a mobile command center which uses a 5G mobile ISP behind CGNAT to an SRX Series Firewall at headquarters.
Which two actions should be performed on the SRX Series Firewall in this scenario? (Choose two.)
Configure the IPsec VPN to use NAT-T.
Configure the IPsec VPN to use IKEv1 aggressive mode.
Configure the IPsec VPN to use IKEv2 aggressive mode.
Configure the IPsec VPN to use DPD.
Answer: A, D
The short version
A and D. Because the 5G site sits behind CGNAT with no public address, the IPsec tunnel must use NAT-T (UDP 4500 encapsulation of ESP) and must use DPD so dead NAT bindings and roaming peer addresses are detected instead of black-holing traffic.
Key concepts
CGNAT (RFC 6598): carrier-grade NAT using the shared address space 100.64.0.0/10, so many subscribers share one public IP and have no inbound-routable address at all.
NAT-T (RFC 3948): NAT traversal that encapsulates IPsec ESP packets inside UDP port 4500 (with a non-IPSEC marker), so firewalls and carrier NAT devices pass them as ordinary UDP.
DPD (Dead Peer Detection): periodic keepalive probing so a dead SA is torn down and renegotiated instead of continuing to black-hole traffic.
Dynamic peers: a site behind CGNAT must be configured as the dynamic/initiating side (or with dynamic DNS), since its address is not known or stable.
IKEv1 aggressive mode: a legacy IKEv1 mode that skips identity protection and is deprecated; IKEv2 has no aggressive mode at all.
Why A and D are correct
A. Behind CGNAT the mobile command centre has no publicly routable address, and the carrier's NAT will drop raw ESP (IP protocol 50) in most configurations. NAT-T solves this by wrapping the encrypted ESP payload in UDP 4500 so it looks like ordinary UDP traffic that NAT devices handle, allowing the mobile gateway to act as a dynamic peer and keep the tunnel alive through the carrier NAT.
D. CGNAT bindings and 5G handovers expire or move, so the SA can silently die. DPD sends IKE keepalives to prove the peer is still reachable; when they stop, the SRX deletes the SA and renegotiates instead of forwarding traffic into a black hole. It is the standard companion control for VPN peers behind NAT or dynamic addressing.
Why the others are wrong
B. IKEv1 aggressive mode is the opposite of what you want here: it does not protect the peer identities (it does not use the protected/authenticated identification exchange) and is deprecated by RFC 7427. It does nothing about NAT traversal or dead-peer detection.
C. There is no "IKEv2 aggressive mode" — IKEv2 defines only the IKE_SA_INIT and INITIAL exchanges, and it always authenticates and protects the identities. This option does not exist as a configurable mode.
JN0-336 exam tip
"Peer behind CGNAT" = two reflexes: NAT-T (make the packets get through) and DPD (notice when they stop). Any option containing the word "aggressive" is wrong by definition — IKEv2 has no aggressive mode.
6An administrator decides to designate a node as the primary node for the chassis cluster.
Which statement is correct in this scenario?
Configure the burnt-in-address (BIA) to the highest value to bring the node as the primary node.
The node with the highest priority will become a primary node.
The node with the lowest priority will become a primary node.
Nodes with a priority of one are ineligible to participate in the election process.
Answer: B
The short version
B — the node with the highest priority becomes the primary node. Junos compares node priorities in the redundancy group during election, and the highest value wins; if that node fails, the surviving node with the next highest priority is promoted.
Key concepts
Node priority: set with set chassis cluster redundancy-group <n> node <id> priority <0-127>; higher value means more preferred as primary.
Priority 0: makes a node never be selected as primary (it stays backup), it is not the "lowest participating" value.
Redundancy group (RG): the unit over which failover is coordinated — a device, an interface, a routing instance or a logical unit; each has node 0 / node 1 members.
Election triggers: node failure detected on the control-link, control-link loss, or graceful/forced failover (request chassis cluster failover).
BIA (burnt-in address): a separate chassis-cluster setting used in RTO mode to make node 0's MAC the burned-in address so asymmetric routing during failover does not break the path.
Why B is correct
Junos performs chassis-cluster primary election by comparing the priorities configured on the members of a redundancy group; the node with the highest priority is the primary. If that node is shut down, loses the control-link, or fails, the surviving node with the higher priority is promoted and takes over the redundancy group. Because priority is administrator-tunable, you can force a preferred primary during maintenance (for example during a software upgrade) by temporarily raising the surviving node's priority, then lowering it back afterwards.
Why the others are wrong
A. The burnt-in address is a real chassis-cluster configuration item, but it has nothing to do with primary selection — it exists so that a node's MAC stays stable for asymmetric routing across a failover. Raising BIA on a node will not make it primary.
C. Lowest priority winning inverts the rule. A lower priority value makes a node less likely to be primary, not more.
D. Priority 1 is simply the lowest valid positive priority — such a node still fully participates in the election and can become primary if nothing better is available. The value that makes a node ineligible for primary is priority 0.
JN0-336 exam tip
Highest priority wins; priority 0 = never primary, priority 1 = participating but least preferred. BIA belongs to asymmetric routing, not to election — two different chassis-cluster settings, easy to swap in questions.
7Which two statements are correct about client-protection Secure Socket Layer (SSL) proxy configurations? (Choose two.)
Server certificate is required.
Root certificate authority (CA) configuration is required.
Root certificate authority (CA) configuration is not required.
Server certificate is not required.
Answer: A, B
The short version
A and B. In client protection the SRX terminates the client's TLS session, so it must present a server certificate the client trusts and it must trust the root CA of the upstream server.
Key concepts
SSL proxy: an application identification feature that decrypts, inspects and re-encrypts SSL/TLS traffic so policy can be applied to the applications inside the tunnel.
Client protection: the SRX terminates TLS from the client and originates a separate TLS session to the real server — the device plays the server role toward the client.
Server protection: the SRX leaves the client-to-server handshake intact and simply inspects certificate and handshake metadata — no local server certificate needed.
Server certificate / signing certificate: the certificate presented to clients when the device terminates their TLS session; it must be issued by a CA the clients trust.
Trusted CA (root certificate): the CA bundle used to validate the certificate presented by the real server on the device's outbound TLS session.
Why A and B are correct
A. Because the SRX is acting as the TLS server toward the client, it needs its own server (signing) certificate. If clients do not trust the issuing CA, they receive a certificate error and the session fails — so a trusted server certificate is mandatory in client protection.
B. The same profile must also open a TLS session to the actual destination server, so the device needs the root CA (or trusted CA bundle) to validate that server's certificate chain. Without it, the upstream session cannot be established.
Why the others are wrong
C. "Root CA configuration is not required" describes the wrong direction of trust for client protection: the device validates the real server's certificate on the upstream session, so a trusted root CA is required.
D. "Server certificate is not required" is true only in server protection, where the device never terminates the client's TLS session. The stem explicitly says client protection, so the local server certificate is mandatory.
JN0-336 exam tip
Client protection: the box is the server to the client -> needs a server certificate. Server protection: the box is the client to the server -> no server certificate, but CA trust still needed. Learn the swap and both directions are easy.
8You are asked to ensure that traffic that matches an IDP policy is not impacted until administrators have a chance to evaluate it.
In this scenario, which IP action should be configured for the policy?
ip-block
ip-notify
ip-connection-rate-limit
ip-close
Answer: B
The short version
B — ip-notify.ip-notify only logs the IDP event; it takes no action on the flow, so traffic matching the policy keeps flowing while administrators review what was detected.
Key concepts
IDP policy: an ordered list of rules in a policy that match on zone, source/destination, service or rule groups, with a severity threshold and a list of attack databases, followed by an action.
then notify / ip-notify: the action that logs the detected event to Syslog and the IDP log (and can raise alarms) without terminating, resetting or blocking the session — the monitor-first behaviour.
then drop-default-flow / ip-close: terminating actions — the session is dropped or closed, so traffic is definitely impacted.
then ip-block: adds the offending source IP to the SRCF blacklist for a configurable period, blocking subsequent sessions from that IP.
then ip-connection-rate-limit: limits new connections from that IP for a configurable period (a throttle), which also changes the traffic behaviour.
Why B is correct
The requirement is explicitly non-disruptive: let the policy match, but take no action until humans have looked. ip-notify does exactly that — it generates the log entry/alarm for the detection and then returns, leaving the session untouched and its traffic flowing. That makes it the action used when you deploy new signatures or a new policy in "monitor only" mode to measure impact before enforcing.
Why the others are wrong
A.ip-block is a blocking action: the source IP of the offending session is added to the SRCF blacklist so subsequent traffic from that IP is denied. That directly impacts traffic, which is the opposite of the requirement.
C.ip-connection-rate-limit throttles new connections from that IP for a set period. Traffic continues but is rate-limited, so connections are delayed or dropped — still an impact on traffic.
D.ip-close closes/resets the offending session immediately. The attack traffic in that very session is cut, so the session is definitively impacted.
JN0-336 exam tip
Learn the IDP action verbs by impact: notify = log only (no impact), drop/close = kill this session, ip-block = blackhole the source IP, rate-limit = throttle new connections. Monitor-only mode always means ip-notify.
9Your manager asks you to update your SRX Series device’s IDP security package. You perform the required steps; however, when you attempt to install the package, you receive an error.
Referring to the exhibit, which two statements are correct about this error? (Choose two.)
IDP stops inspecting traffic.
The IDP license has expired.
IDP continues to inspect traffic only using the installed signatures.
The IDP license is missing (not installed).
Answer: C, D
The short version
C and D. The install fails because the IDP license is missing (not installed), and while the package is not loaded the device keeps inspecting traffic with the signature set that is already installed.
Key concepts
IDP security package: the policy (rules and signature database) downloaded from Juniper and installed on the SRX; it defines what IDP inspects and how it reacts.
Installation commands: request/download the package, then request security idp security-package install to load it and (optionally) commit it for enforcement.
IDP license: an annual subscription/feature license for IDP content; without it the package cannot be loaded onto the device.
Running vs installed dataplane: the currently loaded policy/signature dataplane stays active in the running configuration until a new package is successfully installed and committed.
Cluster considerations: on a chassis cluster the package must be present and licensed on both nodes, and the secondary node copies from the primary during sync.
Why C and D are correct
C. A failed install does not tear down the dataplane that is already running. The SRX keeps the previously loaded policy and signature database in the running configuration and continues inspecting traffic with it — you simply do not get the new content. That is why an install failure is an operational annoyance rather than an outage.
D. The error returned by the install attempt indicates that the IDP license is missing — that is, it is not installed on the device. The installer aborts before loading the new package content, which produces exactly this failure and leaves the old signature set in place.
Why the others are wrong
A. IDP does not stop inspecting traffic when an install fails. The running dataplane keeps performing inspection with the last successfully loaded policy and signatures; at worst the device logs an alarm about the failed update. Losing inspection entirely would be a serious availability and coverage regression, which the running dataplane design deliberately avoids.
B. An expired license is a different condition from a missing license. The error being described concerns the license not being present/not installed, which is why "expired" is not the correct statement about this error.
JN0-336 exam tip
Rule of thumb for IDP packages: install failure = no change to what is running. The device never stops inspecting because it keeps the last good signature set; and check the license installed state before blaming signatures or the package itself.
10You are asked to set up SSL proxy in SRX Series devices. An SSL proxy profile is already defined for you.
Which two steps are required to complete the setup? (Choose two.)
Enable host-inbound-traffic HTTPS in the security zone in which SSL proxy is referenced.
Reference the SSL proxy profile in a security zone.
Reference the SSL proxy profile in a security policy.
Enable any Layer 7 services in the security policy in which SSL proxy is referenced.
Answer: B, C
The short version
B and C. The SSL proxy profile has to be referenced from the security zone where the traffic terminates and referenced from the security policy, because the policy is what actually invokes the proxy on matching flows.
Key concepts
SSL proxy profile: set security application-profiles ssl-proxy <profile> — defines the server certificate, trusted CA, trusted server certificates and SSL behaviours used when proxying.
Application protocol definition: set security application-protocols application-protocol ssl-proxy <profile> — the object that binds the SSL proxy profile into the policy model.
Security policy: the enforcement point; the policy's then clause references the application protocol, which is what makes the SRX decrypt, inspect and re-encrypt matching traffic.
Zone context: the security zone defines the trust level and interfaces the protected sessions terminate on, and it is where the SSL proxy profile association for that zone is established.
L7 application services (ALG, UTM): separate features attached to policies with their own profiles; they are not a prerequisite for SSL proxy to function.
Why B and C are correct
B. The profile must be referenced from the security zone so the device associates the profile with the zone where the protected sessions arrive and terminate — this establishes the trust context and the trust/certificate settings applied there.
C. The security policy is the enforcement point and must reference the SSL proxy profile (through the application-protocol object). Without the policy-side reference there is no flow for the proxy to act on, so no traffic would be decrypted or inspected.
Why the others are wrong
A.host-inbound-traffic https enables the SRX to terminate management traffic (J-Web, HTTPS services) to the firewall itself. It is not how SSL proxy is set up, and it does not reference an SSL proxy profile.
D. Enabling L7 application services in the policy is unrelated to SSL proxy. ALG and the UTM features (AV, anti-spam, content filtering) are separate profiles attached for separate purposes; SSL proxy requires none of them.
JN0-336 exam tip
SSL proxy is attached at two points: the zone (trust context and termination) and the policy (enforcement on matching flows). host-inbound-traffic is about traffic terminating on the firewall, which is the classic confusion.