In Apstra 5.1, a routing zone is the primary construct used to represent an L3 domain for multitenant isolation. In traditional network operating system terms, that maps to a VRF (Virtual Routing and Forwarding instance) . Each routing zone is placed “in its own VRF,” which provides independent routing tables and isolates IP traffic so that different tenants can reuse overlapping IP subnets without conflict. This is central to modern EVPN-VXLAN data center design, where tenants typically require clean separation of routing and policy boundaries.

Within a routing zone, you can create one or more virtual networks (often mapped to VXLAN segments) that provide L2 extension across racks while still being contained by the tenant’s VRF. If L3 gateway services are enabled for those virtual networks, their gateway interfaces (for example, IRB interfaces on Junos v24.4 leaf switches) are associated with the routing zone’s VRF so that inter-subnet routing occurs within the tenant boundary.

This terminology distinction is important: an IRB is an interface construct used to provide L3 gateway functionality for a VLAN/VXLAN segment; a VLAN is a Layer 2 segmentation mechanism; and an access list is a policy enforcement tool. A routing zone, however, defines the tenant’s L3 routing context, which is precisely what a VRF provides on Junos.

Verified Juniper sources (URLs):

https://www.juniper.net/documentation/us/en/software/apstra5.0/apstra-user-guide/topics/concept/routing-zones.html

https://www.juniper.net/documentation/us/en/software/apstra4.2/apstra-user-guide/topics/concept/routing-zones.html

In Apstra 5.1, servers are modeled as Generic Systems and must be represented in the blueprint topology so that Apstra can bind an endpoint (the server) to a specific switch interface and then apply the intended connectivity template / virtual network attachment . In the exhibit, the interface ge-0/0/5 on my_border_001_leaf1 is shown as missing from the assignment workflow, which indicates that Apstra does not currently have an endpoint object connected to that port in the blueprint’s staged physical topology (or that the port is not presented as an eligible connectivity point for server attachment).

The correct remediation is to add a Generic System connected to my_border_001_leaf1 ge-0/0/5 in Staged > Physical > Topology , thereby creating a modeled server link on that interface. Once the Generic System exists and the interface is a recognized server-facing connectivity point, you can assign the Tagged VXLAN “finance-app” connectivity template (or the VN assignment action driven by that template) to the server-facing interface and then commit the staged changes.

Changing “deploy mode” may affect whether Apstra actively configures a generic system-facing link, but it does not solve a missing interface in the topology model. Likewise, allocating an IP pool is unrelated to making the port available for attachment, and assigning the VN to the switch node is not how server interfaces are attached in this workflow.

Verified Juniper sources (URLs):

https://www.juniper.net/documentation/us/en/software/apstra5.1/apstra-user-guide/topics/topic-map/internal-generic-system-create.html

https://www.juniper.net/documentation/us/en/software/apstra5.0/apstra-user-guide/topics/topic-map/virtual-network-assignment-update.html

https://cloudlabs.apstra.com/labguide/Cloudlabs/6.0.0/test-drive-guide/lab1-junos-11_adding-gs.html

In Apstra, Deploy Mode = Drain is the operational mechanism used to gracefully remove a switch from active forwarding before performing maintenance such as an OS upgrade. Drain mode is specifically intended to drain traffic while preserving fabric stability , so that maintenance can be executed with minimal to no application impact, provided the fabric design has sufficient redundancy (for example, ECMP in the underlay and dual-homing/ESI for server attachments). In an EVPN-VXLAN IP fabric, taking a leaf or spine abruptly out of service can cause transient loss of reachability as underlay adjacencies reconverge and the overlay recalculates paths. By placing the device into Drain, Apstra adjusts intent so that traffic is shifted away from the device as much as possible, reducing dependency on it before the upgrade begins.

This is different from Undeploy, which removes Apstra-rendered configuration and is generally used for decommissioning; if a device is carrying traffic, Apstra guidance is to drain first. Ready is a pre-deploy state used in lifecycle workflows, not a maintenance traffic-shifting mode. Deploy keeps the device fully participating. Therefore, for a maintenance window where the goal is “upgrade with minimal interruption,” the correct mode is Drain, then perform the Junos v24.4 upgrade, and finally return the device to Deploy.

Verified Juniper sources (URLs):

https://www.juniper.net/documentation/us/en/software/apstra4.2/apstra-drain-mode/apstra-drain-mode.pdf

https://www.juniper.net/documentation/us/en/software/apstra4.2/apstra-user-guide/topics/topic-map/deploy-mode-update-datacenter.html

https://www.juniper.net/documentation/us/en/software/apstra6.0/apstra-user-guide/topics/topic-map/device-config-lifecycle.html

Apstra manages devices using IP connectivity over the management network, which is a Layer 3 relationship. Whether you are using on-box agents or off-box agents, the controller (or cluster) communicates with the fabric devices using IP reachability (for example, to exchange management traffic, retrieve discovery state, collect telemetry, and push configuration). This is why Layer 2 adjacency is not required between the Apstra server and the managed switches; the essential requirement is routable IP connectivity and appropriate access (credentials/agent connectivity) to the device management interfaces.

From a platform perspective, Apstra does not need a dedicated physical NIC per managed device. Instead, the server/VM requires connectivity to the management network through a single network adapter, and that interface can route to all managed devices. In a typical data center deployment, the Apstra controller VM sits on a management VLAN/subnet and reaches the entire fabric through routed management. This scales operationally: adding devices does not require adding additional server NICs; it only requires IP reachability and capacity planning for telemetry and agent workloads. Thus, the correct statements are that Apstra uses Layer 3 to communicate with managed devices and that it requires a single network adapter for that management connectivity model.

The exhibit shows node100 (Generic System) selected, with links from that generic system to two fabric leaf switches (for example, a leaf participating in an ESI pair and another leaf node). In Apstra 5.1, a Generic System represents an endpoint that is not managed as a network device by Apstra (such as a server, appliance, or host), but it is still modeled so Apstra can apply interface intent (LAG vs single link), connectivity templates, and virtual network attachments.

Because the device is shown as a generic system connected on leaf-facing ports inside the fabric topology , this aligns with an internal generic system . Internal generic systems are used for servers or endpoints that reside “inside” the rack/fabric context and consume leaf switch ports as access-facing connections. This is the common representation for endpoints in EVPN-VXLAN data center designs, where the leaf switches provide the VLAN/VNI mapping and, if required, IRB gateway services within the tenant VRF (routing zone).

An external generic system is typically used for devices outside the fabric boundary—most commonly external routers, firewalls, or upstream networks attached at border leafs—where the intent is external connectivity rather than server access. The selected node is neither a peer switch nor an access switch (those are network infrastructure roles), and the UI explicitly labels it as a Generic System, confirming the correct classification as an internal generic system.

A three-stage Clos (leaf–spine) data center fabric separates roles to keep forwarding predictable and scalable. The leaf layer is the attachment point for endpoints, so servers connect to leaf devices . This is where the fabric accepts workload traffic, applies edge policies, and provides tenant/service constructs (for EVPN-VXLAN fabrics, the leafs typically act as VTEPs and host IRB gateways as required). The spine layer provides the non-blocking transit core between leafs using L3 underlay routing and ECMP, but it is not the place where servers attach directly.

For north-south connectivity, a typical three-stage Apstra data center architecture includes border leaf switches. These border devices provide the controlled connection point between the fabric and external networks (Internet/WAN/DCI or upstream services). As a result, fabric traffic enters and exits the border devices, not the spines. This design keeps the spines dedicated to high-speed east-west transit and simplifies operations: external routing policies, security controls, and inter-domain handoffs are concentrated at the border, while the spines remain a uniform L3 transit tier. In Junos v24.4 EVPN-VXLAN deployments, this separation also helps maintain clean overlay boundaries and consistent underlay behavior across the fabric.

Verified Juniper sources (URLs):

https://www.juniper.net/documentation/us/en/software/juniper-data-center-assurance/user-guide/topics/concept/dc-network-topology-on-dc-assurance.html

https://www.juniper.net/documentation/us/en/software/jvd/jvd-3-stage-datacenterdesign-with-juniper-apstra/solution_architecture.html

https://www.juniper.net/content/dam/www/assets/white-papers/us/en/design-considerations-for-spine-and-leaf-ip-fabrics.pdf

Apstra role-based access control separates fabric operations from identity and authorization administration . The administrator role includes full permissions, including the ability to manage users and roles (for example, creating users, assigning permissions, and creating/cloning/editing custom roles where allowed). This enables administrators to govern who can access the system and what they are permitted to change across all blueprints and system settings.

The user role, in contrast, is designed for day-to-day fabric work: viewing and editing supported blueprint elements and operational objects within the scope permitted by the role, but not administering other users’ access or modifying the role structure itself. In other words, a user can work on the network intent and operations, but cannot elevate privileges, change other users’ roles, or otherwise manage user/role administration unless explicitly granted additional permissions through custom roles.

That makes option C the correct statement: the user role cannot make changes to other user types (that is, it lacks the permissions needed to administer identities/roles). Options A, B, and D do not reflect Apstra’s RBAC model: roles are not primarily constrained “per blueprint” in that way, and users are not intended to modify other roles—those are administrator-level capabilities.

Verified Juniper sources (URLs):

https://www.juniper.net/documentation/us/en/software/apstra6.0/apstra-user-guide/topics/concept/user-role-management.html

https://www.juniper.net/documentation/us/en/software/apstra4.2/apstra-user-guide/topics/concept/user-role-management.html

https://www.juniper.net/documentation/us/en/software/apstra5.0/apstra-user-guide/topics/concept/user-role-management.html

In Apstra, tags are an intent-level metadata mechanism used to classify objects and drive automation and reuse. Within a data center blueprint, Apstra supports tagging multiple blueprint objects so operators can apply configuration or policy logic conditionally (for example, applying a connectivity template or a configlet based on a tag match). In this scenario, three valid taggable objects are virtual networks , interfaces , and generic systems .

Virtual network tagging is supported directly from the blueprint’s virtual network table, enabling you to label virtual networks (such as “finance,” “pci,” or “dev”) and then reference those tags elsewhere in blueprint operations and policy application. Interface tagging is also explicitly supported in the blueprint, allowing you to assign tags to switch interfaces and use those tags to control how templates, assignments, or other intent-driven operations apply to those ports. Finally, generic systems (which are modeled endpoint systems such as servers or external routers represented as “systems” in the blueprint) can be tagged so that downstream intent logic can distinguish system roles and apply the correct operations consistently across expansions and changes.

By contrast, property sets are structured data objects used for variable substitution and probe/configlet parameterization, not a primary target for operational tagging in the blueprint UI; and device profiles are catalog artifacts describing hardware/NOS compatibility rather than blueprint objects typically tagged for intent application.

Verified Juniper sources (URLs):

https://www.juniper.net/documentation/us/en/software/apstra5.1/apstra-user-guide/topics/topic-map/tag-interface-add-remove-datacenter.html

https://www.juniper.net/documentation/us/en/software/apstra5.0/apstra-user-guide/topics/task/tag-virtual-network-update.html

https://www.juniper.net/documentation/us/en/software/apstra5.0/apstra-user-guide/topics/topic-map/tag-system-add-remove-freeform.html

In Apstra 5.1, a Virtual Network (VN) is Apstra’s abstraction for a Layer 2 forwarding domain that groups endpoints into a logical segment across the fabric. Apstra defines virtual networks as being constructed using either VLANs or VXLANs . A VLAN-based VN represents a Layer 2 domain identified by a VLAN ID and is typically used where you want traditional VLAN semantics (often in smaller environments, migration scenarios, or designs where an overlay is not required). A VXLAN-based VN represents the same Layer 2 intent but uses a VXLAN VNI for scalable overlay segmentation, which is the common approach in EVPN-VXLAN data center fabrics.

In an IP fabric architecture, VXLAN provides encapsulation to carry tenant segments over the routed underlay, while EVPN provides the control-plane signaling for MAC/IP reachability. Junos v24.4 leaf devices act as VTEPs, mapping local VLANs/bridge-domains to VNIs and participating in EVPN for advertisement and convergence. Apstra’s VN construct allows you to create the segment once (as VLAN or VXLAN type), then consistently attach it to racks, ports, and endpoints through intent-driven workflows (such as connectivity templates and virtual network assignments).

“L2 VPN” and “L3 VPN” are service provider terms and are not the VN “types” in Apstra’s data center reference design. In Apstra, tenant L3 separation is modeled by routing zones (VRFs), while the VN itself is specifically either VLAN-based or VXLAN-based.

Verified Juniper sources (URLs):

https://www.juniper.net/documentation/us/en/software/apstra5.1/apstra-user-guide/topics/concept/virtual-networks.html

https://www.juniper.net/documentation/us/en/software/apstra5.1/apstra-user-guide/topics/topic-map/virtual-network-create.html

Apstra deploy modes control how far a device progresses in the configuration lifecycle and whether it becomes active in the fabric. If you must keep leaf3-sonic present in the blueprint (modeled, cabled, and available for future use) but you do not want it to participate in IP-fabric routing, you use modes that keep the device not active .

Ready mode assigns the device to the blueprint and applies only “Ready (Discovery 2)” level configuration—hostnames, interface descriptions, and port speed/breakout settings—while explicitly keeping the device out of fabric routing. In this mode, Apstra does not configure routing/BGP or L3 interface addressing for the IP fabric, so the switch is staged and visible for validation (for example, LLDP wiring checks) but does not forward as part of the Clos underlay.

Undeploy mode removes the complete Apstra service configuration from the device. Operationally, this also ensures the device is not active in the fabric. It is commonly used when a device must be retained in the blueprint inventory/topology but should not be participating (for example, temporarily withdrawn, decommission preparation, or held as a spare).

By contrast, Deploy makes the device active (full rendered fabric configuration, including BGP), and Drain is a maintenance state used to gracefully remove traffic from an already-active device rather than a state for keeping it non-participatory from the outset.

Verified Juniper sources (URLs):

https://www.juniper.net/documentation/us/en/software/apstra6.0/apstra-user-guide/topics/topic-map/device-config-lifecycle.html

https://www.juniper.net/documentation/us/en/software/apstra4.2/apstra-user-guide/topics/topic-map/device-config-lifecycle.html