JN0-481: Data Center, Specialist (JNCIS-DC)Demo

Configlets use tags to define their application scope. A tag can identify the applicable systems or devices, avoiding the need to list each one individually.

A Range processor marks a matching input as anomalous. With its threshold set to 1,000,000 and Raise Anomaly enabled, received traffic above 1 Mbps generates an actual IBA probe anomaly. IBA probe anomalies appear under Blueprint Analytics, so the Analytics indicator reflects the anomaly. Service anomalies are distinct from analytics anomalies and are associated with service, discovery, or configuration conditions rather than this probe result. Range Processor | Juniper Apstra 6.0

Apstra security policies can use Virtual Network and Routing Zone as source or destination point types. The platform renders the resulting security policies as ACLs at the applicable enforcement points on fabric devices.

A collapsed fabric consolidates leaf, spine, and border-leaf functions into the leaf devices. The leaf layer is interconnected as a full mesh, and those devices perform the EVPN-VXLAN overlay functions.

Deploy mode applies the service configuration to a managed device and makes it active in the IP fabric. Ready mode only stages an assigned device and leaves it inactive.

OOS-READY is the acknowledged, out-of-service state for an Apstra-managed device. The device has its system agent installed and is available to be assigned to a blueprint, although it is not yet assigned or deployed.

Apstra increases off-box-agent capacity by adding worker nodes to an Apstra VM cluster. A worker VM can run the off-box-agent workload, reducing CPU and memory consumption on the controller while preserving off-box-agent operation.

In Juniper Apstra, templates define network policies, including the spine-to-leaf link underlay type. The IPv4 or IPv6 choice is therefore made when the template is created.

The Juniper Apstra data center life cycle consists of the design, deployment, and operational phases. Configuration and installation can occur as tasks within the broader workflow, but they are not the three named life-cycle phases.

Tenant permissions are routing-zone-based and can grant the ability to manage routing zones. Juniper documents tenant-specific permissions as including management of routing zones, resource groups, virtual networks, and virtual-network endpoints.

VXLAN uses the 24-bit Virtual Network Identifier (VNI) to distinguish virtual Layer 2 segments and their broadcast domains across the IP underlay.

The vlan-based MAC-VRF EVPN service type maps one EVPN routing instance to one VLAN and one bridge table. Juniper: service-type (MAC-VRF Routing Instances)

Juniper Apstra resource pools support autonomous system numbers (ASNs), virtual network identifiers (VNIs), and integers. Routing zones and interfaces use assigned resources or configuration constructs, but they are not resource-pool types.

Tenant isolation is implemented with separate routing zones/VRFs. Inter-tenant traffic therefore requires inter-VRF routing, either on a VTEP-enabled device that performs VRF route leaking or through an external device that routes between the tenant VRFs. iBGP alone does not leak routes between VRFs, and virtual networks do not provide inter-VRF routing.

The Freeform reference design places responsibility for device configuration and validation with the network designer, requiring device CLI/configuration expertise. Freeform blueprints support only Juniper devices. The Datacenter reference design abstracts and automates much of the configuration process and is not limited to Junos devices.

An Apstra routing policy’s import policy determines which BGP routes are permitted into the routing table from the external peer. Setting it to allow only the default route permits 0.0.0.0/0 while rejecting the received 172.23.x/24 prefixes, so those routes are not installed in the Finance VRF. An export policy controls routes advertised toward the peer, and the default-route expectation setting only raises an anomaly rather than changing configuration.

A VXLAN virtual network requires a unique virtual-network name and an associated routing zone. The routing zone supplies the tenant’s Layer 3 domain/VRF association. A VNI may be left blank for automatic allocation from a resource pool, and IPv4 or IPv6 connectivity is optional because VXLAN virtual networks can operate as Layer 2-only networks.

Apstra property sets are structured as key-value data and may be written in YAML or JSON. A lone YAML scalar does not satisfy the required dictionary structure; valid YAML mapping syntax, key: value, creates the required dictionary.

In Juniper Apstra, pod-based templates are used to create five-stage Clos networks. A rack-based filter excludes those templates, so selecting the POD BASED filter makes the five-stage template available.

Apstra supports global resource pools that can be used by one or more blueprints, as well as local pools that are created within and specific to a single blueprint. Thus, pool scope can be global or blueprint-specific.

In Juniper Apstra, rack types include generic systems for compute or storage devices connected within the rack. A new server added to that rack type is therefore a rack generic system.

A yellow/orange Uncommitted status indicates pending blueprint changes that can be committed, so those changes have not yet been made active. A red status indicator identifies an issue; in the Active view, red indicators are used to locate anomalies that require investigation or remediation.

Property sets store reusable values that configlets use for variable substitution, and a configlet can reference multiple property sets. Associated property sets are imported into the blueprint with the configlet. Property sets also work with Analytics probes, and they are defined in JSON or YAML rather than with syntax unique to each supported vendor.

Leaf devices provide the access-layer connections for servers, while border devices provide the fabric’s connection to external networks. Spine devices supply transit connectivity between leaf devices rather than directly connecting servers or acting as the external network edge.