Historical exam notice (October 2026): The JN0-348 examination named in this article’s published title is retired. Juniper replaced JN0-348 with JN0-349 in September 2021, introduced JN0-351 in June 2023, and currently lists JN0-352 for the JNCIS-ENT (Enterprise Routing and Switching, Specialist) credential. The concepts below remain valuable, but this is not a registration guide for JN0-348. Use the current JN0-352 objectives and testing requirements when planning a certification attempt. Junos routing policy and operational verification remain useful after an exam-code transition; distinguish the active specialist requirements from retired exams when navigating the wider Juniper certifications family.
What the current JNCIS-ENT assessment actually covers
JNCIS-ENT is a specialist-level assessment of Junos enterprise routing and switching. It assumes that you can recognize how Ethernet switching, VLANs, routed interfaces, routing policy, and high-availability mechanisms interact. Memorizing a few commands is not sufficient: the official objectives test configuration, monitoring, and fault isolation. Juniper identifies JNCIA-Junos as the prerequisite, a 90-minute assessment, and 65 multiple-choice questions for JN0-352. Verify these details again at scheduling, because exam versions change.
A practical preparation plan starts by splitting the published objectives into labs instead of treating them as a list of isolated terms. The same physical or virtual topology can be used for multiple exercises. For instance, create two access VLANs on a switch, attach a tagged trunk to another switch, and add routed interfaces so clients in one VLAN can reach a different VLAN subject to the correct filtering policy. Write down the expected traffic path before making any configuration changes.
Lab 1: Validate switching, VLAN membership, and Layer 2 protection
Build two switches connected by a tagged interface. Assign host-facing access ports to distinct VLANs, verify the native VLAN expectation at both ends, and compare the MAC tables before and after generating traffic. A frame will not reach a remote VLAN simply because two switches are linked: the VLAN must be carried on the trunk, and inter-VLAN communication needs appropriate Layer 3 forwarding.
Use operational commands such as show vlans, show ethernet-switching table, and show interfaces terse, adjusting for the platform and Junos release. Do not assume that the same command families, interface naming, or ethernet-switching configuration syntax apply across every EX and QFX model. Confirm whether a port is tagged, whether the intended VLAN is present, and where a learned MAC address appears.
Next deliberately introduce a mistake: remove one VLAN from a tagged trunk, disable the member port, or misconfigure the expected VLAN identifier. Explain which observation points distinguish a VLAN membership problem from a down physical link. A missing VLAN on the tagged uplink breaks VLAN segmentation even when the physical link stays up; comparing the access-port membership and trunk allowed-VLAN list isolates the failure.
Layer 2 protection is part of the current objectives. Practice recognizing a topology loop and distinguishing spanning-tree behavior from port-security and DHCP-snooping or ARP-inspection controls. Do not enable protection features on a production port merely to test a hypothesis. A safe lab records the expected outcome, the exact configuration delta, and the rollback procedure.
Lab 2: Explain OSPF neighbors and route selection
Connect three Junos routers in a triangle and advertise distinct prefixes. Give two paths different OSPF costs so the selected route is predictable. Verify the adjacencies and database rather than judging convergence from a ping alone. Commands worth practicing include show ospf neighbor, show ospf database, and show route protocol ospf on platforms that support them.
Break one adjacency deliberately using an area mismatch or incorrect interface setting. Note the resulting neighbor state and what the routing table does when that neighbor stops advertising paths. A correct explanation should distinguish the interface protocol state, OSPF adjacency, link-state advertisement information, the calculated route, and the final forwarding result. If you cannot explain which stage failed, adding more configuration commands is unlikely to help.
Routing policy is not the same as security filtering. Review import/export policy decisions, prefix matching, and the role of routing instances in isolation. In particular, a route’s presence in one routing table does not automatically mean that a host in another routing instance can reach it. Cross-instance communication requires intentional design and a separate security review.
Lab 3: Routing behavior that is independent of any one protocol
Create static, aggregate, and generated route cases, then test which prefixes are active and where traffic actually forwards. Record what happens to a static next hop when its underlying connectivity is removed; not all route types fail or withdraw in the same way. Include a simple filtered forwarding policy and examine the result using both routing-table and interface-level observations.
Juniper routing-instances, routing table groups, and filtered forwarding have precise platform-specific semantics. Treat a copied configuration as a hypothesis, not a deployable recipe. Use commit check, an approved change window, and rollback procedures appropriate to the Junos version. The dedicated guide to Junos operational and configuration modes provides the practical command workflow for safe lab changes.
Lab 4: High availability, not simply another redundant link
Distinguish a link aggregation group (LAG) from an upstream gateway failover mechanism, and distinguish VRRP from routing-process redundancy. For LAG, verify member links, negotiation mode, consistent VLAN settings, and how a single-member failure affects observed throughput. For VRRP, check the virtual gateway, current master state, and whether failure detection produces the expected client behavior.
Current JNCIS-ENT objectives also mention virtual chassis, graceful restart, nonstop active routing, graceful Routing Engine switchover, nonstop bridging, and ISSU. These are not interchangeable. The right lab question is which control-plane or forwarding-plane interruption a mechanism is designed to survive, what hardware and software prerequisites it has, and what observations prove that failover succeeded. Never treat an HA label as a guarantee of zero interruption.
Turn every lab into exam preparation. For each topic, keep a short lab record with five fields: intended topology, relevant configuration, expected observations, injected fault, and recovery evidence. Include commands and output snippets, but omit secrets and real customer device addresses. One coherent failure investigation teaches more than ten disconnected command flashcards.
As you review practice material, classify wrong answers by cause. Missing a VLAN question may reflect confusion about tagging; missing an OSPF question may reflect adjacency formation versus route selection; missing a high-availability question may reflect confusion between a control-plane switchover and data-plane survival. Fix the underlying concept and repeat the lab rather than memorizing the answer wording.
Choosing resources and the correct exam
The original JN0-348 syllabus can still serve as historical background, but its old exam code must not be used for new scheduling. The current Juniper certification program lists JN0-352 for JNCIS-ENT. No approved JN0-352 main-exam URL was found in the ExamTopics.biz destination inventory during this review, so this article does not fabricate a commercial exam link. Candidates should confirm official issuer objectives, registration rules, and the latest exam number before paying for any preparation product.
A productive study plan combines official objectives, a Junos lab, operational verification, and error-driven revision. The credential is useful only insofar as the engineer can explain why a packet is forwarded, filtered, dropped, or redirected in a real enterprise topology. That skill survives exam-number changes.
Diagnostic practice: two faults that look alike
In one exercise, a client on access VLAN 20 cannot ping its configured gateway. Start with the physical port state, then confirm that VLAN 20 exists, that the port is a member of it, and that the router or switch virtual interface serving the VLAN is reachable. Review the learned MAC entries and the host’s ARP result. If ARP never resolves, investigate the Layer 2 path before changing OSPF; a route across the rest of the network cannot repair a local Ethernet VLAN mismatch.
In a second exercise, the client reaches its gateway but not a routed server subnet. Verify the source routing instance, the route for the server prefix, next-hop reachability, policy filters, and reverse traffic path. If OSPF neighbors are down, inspect the specific adjacency conditions; if neighbors are up, compare advertised routes and routing policies. Keep these as distinct fault categories, even if both are reported by the user as “the network is down.”
Design a short, realistic study cycle
On the first study day, build the topology and document working baseline commands. On a later day, apply one controlled VLAN or OSPF fault and gather evidence. After reviewing the relevant official JN0-352 objective, repeat the lab without referring to a step list. On the final day, explain the entire fault flow to someone who was not present, including how the fix was validated and what residual risk remains. This exposes conceptual gaps that a practice-question score may conceal.
Capture the Junos command-output examples, but never publish lab credentials, production addresses, or client configuration files. If an operational command differs on an EX, QFX, or MX release, document the variant and consult release-specific guidance rather than assuming a failed command means the feature does not exist.
Use current issuer objectives as the booking authority
Historical JN0-348 study notes can reinforce OSPF adjacency troubleshooting and switching fundamentals, but they cannot establish the correct current examination, software assumptions, eligibility or fees. Verify the live JNCIS-ENT exam code and blueprint directly in the Juniper certification portal before scheduling, then build labs around the objectives that are actually listed there.