H35-211_V2.5 HCIP-Access V2.5 Question and Answers

Ethernet VLAN identifiers provide a practical limit of about 4K per broadcast domain. As user counts and services grow toward this boundary, Huawei recommends dual-layer VLAN (QinQ) to extend scale and maintain service isolation. Hence the planning threshold is ~4K.

Huawei OLTs support trust/remark mechanisms that can inherit (copy) user packet priorities (for example, trusting incoming 802.1p or DSCP to drive forwarding queues) and also support locally configured priorities (policy-based remarking/mapping on the OLT) to enforce service differentiation. Thus, both priority inheritance and local priority configuration are supported, making the statement Right.

Huawei GPON implements hierarchical traffic control. CAR (Committed Access Rate) can be applied at multiple layers: GEM-port level (GEM-CAR), per-flow (Flow-CAR), and T-CONT level (T-CONT-CAR) for upstream bandwidth policing/shaping and guaranteeing QoS. GEM mapping is a classification/association mechanism (service → GEM/T-CONT) rather than a traffic-control action, so it is not a CAR measure.

Huawei GPON service provisioning supports mapping upstream/downstream service flows to GEM Ports using multiple classifiers, including MAC address, physical port (PORT), VLAN ID, and 802.1p (priority). These classifiers are configured in line/service templates (and pushed via U2000) to ensure precise flow-to-GEM binding.

A: Incorrect — RADIUS typically uses UDP (1812/1813), while HWTACACS/TACACS+ uses TCP (port 49).

B: Incorrect — In RADIUS, only the user-password attribute is encrypted; fields are not selectively encrypted by user choice. TACACS+ can encrypt the entire payload.

D: Incorrect — Because RADIUS uses UDP, transport is not inherently reliable.

C: Correct — RADIUS and HWTACACS differ in message structure and returned attributes after authentication (e.g., RADIUS attribute–value pairs for access control/QoS/VLAN vs. HWTACACS command/privilege oriented responses). Thus the returned content/format differs between the two.

SSH provides secure remote login and secure file transfer (SCP/SFTP) with encryption, integrity, and authentication. Clients validate the server/host and then perform user authentication (password, public key, etc.), often described as two layers of authentication. SSH runs over TCP (port 22) at the application layer, not at the network layer—therefore D is the incorrect statement.

For Huawei uplink protection (STH-NNI port protection) with workmode timedelay, the mandatory steps are:

Create/define the protection group with the protect-target and workmode timedelay;

Add member ports with roles (work/protect);

Enable the protection group.

Interface-level timers like offline-tx-off-time under the physical interface are not required for standard timedelay protection operation and are therefore optional.

In NGN architectures, services (voice, video, data) converge over a packet-based backbone. The core switching layer provides the unified transport using IP/MPLS and related QoS mechanisms, delivering high reliability and large-capacity switching for all upper-layer services. The control and management layers orchestrate services, while the edge/access layer aggregates users—but the packet-based integrated transmission platform itself is the role of the core.

The ONT details pane in eSight displays access-port attributes (admin/state, type), Ethernet rate/speed/duplex, and PON optical module telemetry such as temperature and Tx/Rx optical power for condition monitoring and troubleshooting.

This statement is wrong. In U2000, network elements (NEs) can be added by multiple methods, including batch addition/import and automatic discovery, not just one-by-one during a live access session. U2000 supports template-based onboarding and batch operations to improve efficiency for large-scale deployments.

XG-PON (10G GPON) adopts XGEM framing. Compared with GPON’s 12-bit GEM Port-ID, XGEM expands the Port-ID field to 16 bits, increasing the number of logical data flows that can be identified per ONU/OLT and improving service granularity. Hence the correct length is 16 bits.

In IP multicast, only one copy per multicast group (program) is forwarded upstream, regardless of how many users are watching it. Replication happens downstream at the access/edge, not toward the core. Therefore the upstream traffic equals the number of active multicast groups × per-group bitrate:

20 programs × 2 Mbit/s = 40 Mbit/s.

This behavior is consistent with Huawei’s multicast feature descriptions and general multicast principles (IGMP/PIM), which explain that multicast traffic is sent as a single flow per group and replicated only where branching occurs, not multiplied by user count.

Drag & Drop Answer (NGN layers → devices):

Network control layer → Softswitch

Service management layer → OSS system

Edge access layer → MA5800

Core switching layer → Core router

Why:

Softswitch handles call/session control ⇒ control layer.

OSS provides O&M/service management ⇒ service management layer.

MA5800 is an OLT/access device ⇒ edge access layer.

Core router provides backbone forwarding ⇒ core switching layer.

In FTTB/C voice access with SIP, typical “off-hook but no tone/voice” faults are caused by signaling/control path or device status issues: abnormal MG (Media Gateway/H.248/SIP) interface state on the ONU, unreachable routing between ONU and IMS/softswitch, or a voice server outage. These directly prevent registration, call setup, or media negotiation, resulting in no tone or no audio.

Generic “upper-layer packet loss” on the OLT uplink is not a standard, deterministic root cause specifically for off-hook immediate sound failure; minor loss does not inherently suppress locally generated tones and is not listed as a primary cause in Huawei voice troubleshooting guidance.

Type C dual-homing provides device-level protection with two OLTs; when one OLT/uplink fails, services switch to the peer OLT. In single-homing, ONUs are associated with only one OLT; therefore an OLT failure interrupts services (C). Single-homing has simpler networking and management than dual-homing (D). It does not offer switchover to another OLT (so B is incorrect), and the “important-service” positioning in A is typically applied to dual-homing scenarios that require the highest availability.

In 10G-EPON (IEEE 802.3av), the wavelength plan uses 1577 nm downstream and 1270 nm upstream. The 1270 nm upstream is used in both asymmetric and symmetric 10G-EPON profiles.

In campus/access QoS planning on Huawei devices, queue scheduling commonly supports RR, WRR/DRR, and (optionally) strict-priority. WRR (Weighted Round Robin) assigns larger weights to high-priority queues so they receive more transmission opportunities and bandwidth than lower-priority queues while still serving all queues. Therefore, when ensuring better treatment for high-priority services under round-robin–type scheduling, WRR is the recommended choice.

A is consistent with Huawei L2 multicast practice: global L2 multicast/IGMP snooping parameters can be set and then take effect for multicast VLANs involved in service provisioning.

B reflects the forwarding/CPU-punt behavior on access line cards: IGMP Reports/Leaves from users are recognized by board logic and processed by the control plane to update multicast state.

C matches default service control on BTV/IPTV access where a default program limit (commonly 8) per user applies unless adjusted by policy/profile.

D is incorrect: when IGMP match mode is disabled, multicast programs are not required to be pre-provisioned as channels; groups can be learned dynamically by IGMP. Preconfiguring channels is needed when match mode is enabled (channel matching enforced).

GPON specifies data encryption (commonly AES-128) on downstream traffic so that one ONU cannot eavesdrop another’s data on the shared PON. This cryptographic protection is the standard, practical method to prevent monitoring on GPON access networks; the other options are operational workarounds, not normal GPON security mechanisms.

SNMPv2c uses a community-string model (clear-text “read”/“write” strings) and provides no authentication integrity or encryption. Only SNMPv3 adds auth/privacy (AuthNoPriv/AuthPriv) with hashing and encryption. Therefore, stating that SNMPv2c can encrypt user passwords is false.

OLT-level (device-level) protection is achieved by dual-homing ONUs to two independent OLT interfaces so that services can switch to another OLT when one fails. In Huawei protection taxonomy, the dual-homing scheme (as contrasted with single-homing) provides device-side redundancy for the OLT. By comparison, Ethernet port protection groups and LACP protect uplink links and bandwidth at Layer 2, not the OLT device itself; single-homing does not protect against OLT failure.

IP multicast is designed for one-to-many delivery, allowing a single source stream to be replicated in the network only where branching is required. Compared with sending multiple unicast copies, multicast significantly reduces bandwidth consumption on shared links and provides efficient point-to-multipoint distribution. This is the core rationale for using multicast in video/TV services on campus and access networks.

Terminal (ONT/ONU/CPE) selection adheres to engineering principles: align model and interfaces with the deployment mode and scenario, evaluate TCO/cost, and satisfy special requirements (such as PoE, Wi-Fi specs, voice/FXS, industrial temperature). Ignoring project requirements is contrary to standard design workflows; therefore D is not correct.

For IP-based discovery, eSight relies on Layer-3 reachability and SNMP/ICMP. Being on the same subnet is not required; that condition applies to broadcast/ARP discovery. Required conditions include proper permissions, IP connectivity, and aligned SNMP parameters.

This is wrong. A PON port LED not green may indicate several states—not only a hardware fault. It can mean the port is administratively shutdown, no ONUs are registered, or the port is in an alarm or idle state depending on the LED color and blinking pattern. Fault requires correlation with alarms and performance counters in U2000, not LED color alone.

IGMP is the host-side membership protocol: the first-hop router (querier) uses IGMP Queries/Reports/Leaves to manage host participation in multicast groups and to ensure forwarding state reflects receiver interest. PIM-DM/PIM-SM build distribution trees between routers, and MSDP shares source information between PIM-SM domains—they are not used by hosts.

XG-PON encapsulates user data into XGEM frames. The XGEM Port-ID field is 12 bits, allowing identification of logical connections (GEM ports) per ONU. This mirrors the 12-bit Port-ID used in GPON GEM.

With PITP enabled and Vendor-tag switching allowed, Huawei OLTs support transparent transmission and augmentation: the OLT keeps the user-carried Vendor tag and appends the local Vendor tag (e.g., slot/port/location), ensuring the upstream AAA/BRAS receives both the access-side identity and the OLT’s own location identity. Simply adding only the local tag (A) loses user-side info; forwarding unchanged (B) omits OLT location; discarding (D) contradicts the feature intent.

In Huawei IPTV troubleshooting, mosaic/flash/jerky or intermittent picture typically results from packet loss or bit errors. Common causes include: optical power near the receiver sensitivity threshold causing FEC/BER spikes (A); loss of I-frames (key frames) either in the encoder/compression stage (B) or due to line transmission loss/jitter (D); and device rate-limit (CAR/policing) misconfiguration that drops packets under burst traffic (C). Any of these can yield unstable display.

On Huawei OLTs (e.g., MA5600T/MA5680T), deploying IPTV/multicast commonly uses a program list (channel list) to map program numbers to multicast group addresses—this is configured with igmp program. When programs are statically configured under the multicast VLAN, the program definitions themselves are mandatory; without them, set-top boxes cannot select (zap to) channels by program number/address mapping and the OLT cannot correctly forward the requested multicast streams.

A single-layer VLAN is limited by the IEEE 802.1Q VLAN ID space (1–4094), i.e., about 4K VLANs. In large PON deployments, when you must separate multiple services (e.g., HSI, VoIP, IPTV) and the user scale approaches or exceeds the 4K ceiling, a dual-layer VLAN (QinQ) design is required to extend the VLAN namespace and implement hierarchical isolation (access VLAN + service/aggregation VLAN).

In GPON Type C dual-attribution protection, the ONU has dual PON connections to two OLT PON ports (or two OLTs). When the protection group is configured to track the OLT’s uplink Ethernet port, a Down state on that uplink is an OLT-side trigger. The OLT initiates the protection action (forces the working/standby role swap via PON protection signaling).

A — Correct behavior: After the OLT triggers protection, the ONU’s working PON interface becomes standby, and the protect interface becomes working.

B — Correct behavior: The OLT monitors and updates the ONU uplink status as part of the protection group state machine.

C — Correct behavior: After inversion, services egress the ONU via the protect PON port toward the OLT, i.e., traffic follows the configured protect path.

D — Incorrect (answer): In this scenario the trigger is the OLT’s uplink Ethernet status, not an ONU-detected failure. Therefore the ONU does not detect the fault nor initiate the switchover; the OLT does.

(Reference: HCIP-Access V2.5 – GPON protection mechanisms; Huawei OLT protection group behavior, Type C dual-homing with uplink-tracking.)