Aruba Switching Expert (Written) study materials
Aruba Switching Expert (Written) study resources

Aruba Switching Expert (Written) cheat sheet

Aruba Switching Expert (Written) · Keep the key ideas within reach.

Quick reference

Switching Fundamentals and Network Design

  • Store-and-forward switches buffer the entire frame and verify FCS before forwarding; cut-through forwards after reading the destination MAC, so it can propagate errored frames.

  • A Layer 2 switch builds its MAC address table by learning source MACs per ingress port; unknown unicast, broadcast, and multicast frames are flooded to all ports in the VLAN.

  • Collision domains end at a switch port; broadcast domains end at a router or Layer 3 boundary.

  • Access port carries one untagged VLAN; trunk port carries multiple VLANs with 802.1Q tags (native VLAN untagged by default).

  • Campus design tiers: access (endpoint density, PoE), distribution (policy, inter-VLAN routing), core (high-speed transit); collapsed core merges distribution and core.

  • Oversubscription ratio = total downstream access bandwidth / uplink bandwidth; typical access-to-distribution design targets 20:1 or better.

  • Spanning a VLAN across many access switches enlarges the failure domain; keep VLANs local to a wiring closet where possible.

  • MTU mismatch causes silent drops for large frames; jumbo frames require consistent MTU on every Layer 2 hop and the routed path.

Quick reference

VLANs, Trunking, and Layer 2 Segmentation

  • 802.1Q inserts a 4-byte tag; TPID 0x8100, 3-bit PCP priority, 1-bit DEI, 12-bit VLAN ID (usable range 1-4094; 0 and 4095 reserved).

  • Native VLAN frames cross a trunk untagged; a native VLAN mismatch between two ends causes VLAN leaking and is flagged by CDP/LLDP-style checks.

  • VLAN 1 is the default VLAN on most switches and should not carry user traffic; management VLAN should be separate and unused elsewhere.

  • Allowed VLAN lists on a trunk filter which tagged VLANs pass; pruning reduces flooded traffic to switches that need the VLAN.

  • Voice VLAN separates IP phone traffic, often using LLDP-MED to advertise the voice VLAN and QoS policy to phones.

  • Private VLANs: promiscuous port reaches all; isolated ports reach only promiscuous; community ports reach each other and promiscuous.

  • GVRP/MVRP dynamically propagate VLAN registrations; static VLAN configuration is preferred for deterministic behavior.

  • A Layer 2 loop with no STP causes broadcast storms, MAC table instability, and duplicate frames within seconds.

Quick reference

Spanning Tree Protocols and Loop Prevention

  • Root bridge election: lowest bridge ID wins; bridge ID = bridge priority (default 32768, multiples of 4096) + system MAC.

  • 802.1D port states: Blocking, Listening, Forwarding, Disabled; 802.1w (RSTP) uses Discarding, Learning, Forwarding.

  • RSTP port roles: Root, Designated, Alternate (backup for root), Backup (backup for designated); edge ports skip listening/learning.

  • Path cost is cumulative to root; lower root path cost wins for root port; ties broken by lowest sender bridge ID, then lowest port ID.

  • Convergence: 802.1D up to 30-50 s; RSTP sub-second to a few seconds; MSTP converges per instance.

  • MSTP maps VLANs to instances; region must match on name, revision, and VLAN-to-instance mapping or the switch forms a separate region.

  • BPDU Guard disables a port that receives BPDUs (protects edge ports); Root Guard blocks superior BPDUs to protect the root location.

  • Loop Guard prevents an alternate/root port from becoming designated when BPDUs stop arriving; it blocks the port instead of forwarding.

Quick reference

Link Aggregation and Redundancy

  • LACP (802.3ad) negotiates aggregation with LACPDUs; static/manual trunks skip negotiation and can blackhole on misconfiguration.

  • All member links in an aggregation must match speed, duplex, MTU, and VLAN membership; mismatches keep the link out of the bundle.

  • LACP modes: active (initiates), passive (responds); active-active or active-passive both form a bundle; passive-passive does not.

  • Load-balancing hash uses Layer 2/3/4 fields (src/dst MAC, IP, port); a single flow uses one link, so per-flow bandwidth is limited to one member.

  • An aggregation is one logical STP port; bundling links reduces STP topology and eliminates blocked redundant links.

  • Multi-chassis aggregation (MLAG/VSX-style) lets two switches share one logical bundle toward a server or downstream switch.

  • LACP fallback/static mode keeps a minimum number of links active if LACP negotiation fails, depending on platform support.

  • Redundant uplinks without aggregation rely on STP; with aggregation, use LACP plus a routing or MLAG design to avoid loops.

Quick reference

Stacking and Switch Virtualization

  • A stack presents one management IP and one control plane; member switches share a distributed forwarding plane.

  • Stack roles: commander/master (control), standby, and members; election is by priority then MAC.

  • Stack cabling is typically a ring for redundancy; a broken ring can split the stack into independent fragments if split detection is not configured.

  • Stack split (brain split) can create duplicate IPs and loops; configure split detection and redundant management paths.

  • Firmware must match across members; a mismatch can prevent a member from joining or cause an unstable stack.

  • Virtual switching (e.g., VSX/MLAG pairs) keeps separate control planes but synchronizes forwarding state, avoiding single-control-plane risk.

  • Stack ports have finite bandwidth; oversubscribing stack links between members causes inter-member traffic to bottleneck.

  • Adding a member with a different model may be restricted by mixed-stacking rules; verify supported combinations before deployment.

Quick reference

Layer 3 Switching, Routing, and IP Services

  • An SVI is a logical Layer 3 interface bound to a VLAN; it needs an IP address and the VLAN must be active with at least one up port (platform-dependent).

  • Inter-VLAN routing options: router-on-a-stick (subinterfaces on a trunk), SVIs on a multilayer switch, or routed ports.

  • Longest-prefix match wins in the routing table; administrative distance breaks ties between protocols; metric breaks ties within a protocol.

  • Static route default AD is 1; connected is 0; OSPF is 110; eBGP is 20; iBGP is 200 (Cisco-style values; verify platform defaults).

  • OSPF: neighbors must match area, hello/dead timers, network type, and authentication; DR/BDR elected on broadcast networks.

  • DHCP relay (helper address) forwards client broadcasts to a server; the relayed packet carries the gateway IP so the server can pick the pool.

  • ARP resolves IPv4 to MAC within a subnet; proxy ARP answers for hosts that believe a remote host is local.

  • VRRP/HSRP provide a virtual gateway IP and MAC; priority and preemption determine the active router.

Quick reference

Network Security and Access Control

  • Port security limits learned MACs per port; violation actions are protect (drop), restrict (drop and log), and shutdown.

  • 802.1X roles: supplicant, authenticator (switch), authentication server (RADIUS); EAPoL carries authentication between supplicant and switch.

  • MAB (MAC Authentication Bypass) authenticates devices that lack an 802.1X supplicant by their MAC against RADIUS.

  • DHCP snooping trusts only designated uplink ports; untrusted ports cannot offer DHCP; the binding table feeds Dynamic ARP Inspection.

  • Dynamic ARP Inspection validates ARP packets against the DHCP snooping binding table to block ARP spoofing.

  • IP Source Guard filters traffic by source IP/MAC binding, preventing address spoofing on access ports.

  • Storm control rate-limits broadcast, multicast, or unknown unicast; exceeding the threshold can drop or shut the port.

  • Management plane hardening: SSH instead of Telnet, AAA with RADIUS/TACACS+, ACLs on management VLAN, and disable unused services.

Quick reference

Multicast, QoS, and Traffic Management

  • IPv4 multicast range 224.0.0.0/4; 224.0.0.0/24 is link-local and never forwarded by routers.

  • IGMP snooping lets a switch forward multicast only to ports that joined the group; without it, multicast floods like broadcast.

  • PIM-SM uses a rendezvous point; PIM-DM floods then prunes; IGMPv2/v3 differ in leave and source-specific join support.

  • QoS pipeline: classification, marking (DSCP 6-bit / CoS 3-bit), queuing, scheduling, and congestion avoidance.

  • DSCP per-hop behaviors: EF (46) for voice, AF classes with drop precedence, CS values for class selectors; default is 0 (best effort).

  • Trust boundaries: trust DSCP/CoS at the access edge only for known devices; re-mark untrusted traffic at the boundary.

  • Weighted random early detection (WRED) drops packets before queue overflow to signal TCP senders; tail drop waits until the queue is full.

  • Shaping buffers and delays excess traffic; policing drops or remarks excess traffic immediately.

Quick reference

Monitoring, Troubleshooting, and Operations

  • Interface counters: runts (<64 bytes), giants (>MTU), CRC errors (physical/cable), late collisions (duplex mismatch), and input/output discards.

  • Duplex mismatch symptom: late collisions, FCS errors, and poor throughput on one side; verify both ends and avoid hardcoding one side.

  • Show commands to sequence: interface status, MAC table, ARP, VLAN/trunk status, STP topology, then routing table.

  • Syslog severity 0 (emergency) to 7 (debug); SNMP traps and RMON provide event and statistics collection.

  • Port mirroring (SPAN) copies traffic to a monitor port; a remote mirror (RSPAN) carries mirrored traffic in a dedicated VLAN.

  • LLDP/CDP reveal neighbor identity, port, and capabilities; mismatched native VLAN or MTU often shows here first.

  • Configuration management: save running config to startup, keep backups, and use version control or a config server for rollback.

  • NTP synchronization is required for accurate log correlation and for time-based ACLs or certificates.

Quick reference

Switching Fundamentals and Network Design

  • Store-and-forward switches buffer the entire frame and verify FCS before forwarding; cut-through forwards after reading the destination MAC, so it can propagate errored frames.

  • A Layer 2 switch builds its MAC address table by learning source MACs per ingress port; unknown unicast, broadcast, and multicast frames are flooded to all ports in the VLAN.

  • Collision domains end at a switch port; broadcast domains end at a router or Layer 3 boundary.

  • Access port carries one untagged VLAN; trunk port carries multiple VLANs with 802.1Q tags (native VLAN untagged by default).

  • Campus design tiers: access (endpoint density, PoE), distribution (policy, inter-VLAN routing), core (high-speed transit); collapsed core merges distribution and core.

  • Oversubscription ratio = total downstream access bandwidth / uplink bandwidth; typical access-to-distribution design targets 20:1 or better.

  • Spanning a VLAN across many access switches enlarges the failure domain; keep VLANs local to a wiring closet where possible.

  • MTU mismatch causes silent drops for large frames; jumbo frames require consistent MTU on every Layer 2 hop and the routed path.

Quick reference

VLANs, Trunking, and Layer 2 Segmentation

  • 802.1Q inserts a 4-byte tag; TPID 0x8100, 3-bit PCP priority, 1-bit DEI, 12-bit VLAN ID (usable range 1-4094; 0 and 4095 reserved).

  • Native VLAN frames cross a trunk untagged; a native VLAN mismatch between two ends causes VLAN leaking and is flagged by CDP/LLDP-style checks.

  • VLAN 1 is the default VLAN on most switches and should not carry user traffic; management VLAN should be separate and unused elsewhere.

  • Allowed VLAN lists on a trunk filter which tagged VLANs pass; pruning reduces flooded traffic to switches that need the VLAN.

  • Voice VLAN separates IP phone traffic, often using LLDP-MED to advertise the voice VLAN and QoS policy to phones.

  • Private VLANs: promiscuous port reaches all; isolated ports reach only promiscuous; community ports reach each other and promiscuous.

  • GVRP/MVRP dynamically propagate VLAN registrations; static VLAN configuration is preferred for deterministic behavior.

  • A Layer 2 loop with no STP causes broadcast storms, MAC table instability, and duplicate frames within seconds.

Quick reference

Spanning Tree Protocols and Loop Prevention

  • Root bridge election: lowest bridge ID wins; bridge ID = bridge priority (default 32768, multiples of 4096) + system MAC.

  • 802.1D port states: Blocking, Listening, Forwarding, Disabled; 802.1w (RSTP) uses Discarding, Learning, Forwarding.

  • RSTP port roles: Root, Designated, Alternate (backup for root), Backup (backup for designated); edge ports skip listening/learning.

  • Path cost is cumulative to root; lower root path cost wins for root port; ties broken by lowest sender bridge ID, then lowest port ID.

  • Convergence: 802.1D up to 30-50 s; RSTP sub-second to a few seconds; MSTP converges per instance.

  • MSTP maps VLANs to instances; region must match on name, revision, and VLAN-to-instance mapping or the switch forms a separate region.

  • BPDU Guard disables a port that receives BPDUs (protects edge ports); Root Guard blocks superior BPDUs to protect the root location.

  • Loop Guard prevents an alternate/root port from becoming designated when BPDUs stop arriving; it blocks the port instead of forwarding.

Quick reference

Link Aggregation and Redundancy

  • LACP (802.3ad) negotiates aggregation with LACPDUs; static/manual trunks skip negotiation and can blackhole on misconfiguration.

  • All member links in an aggregation must match speed, duplex, MTU, and VLAN membership; mismatches keep the link out of the bundle.

  • LACP modes: active (initiates), passive (responds); active-active or active-passive both form a bundle; passive-passive does not.

  • Load-balancing hash uses Layer 2/3/4 fields (src/dst MAC, IP, port); a single flow uses one link, so per-flow bandwidth is limited to one member.

  • An aggregation is one logical STP port; bundling links reduces STP topology and eliminates blocked redundant links.

  • Multi-chassis aggregation (MLAG/VSX-style) lets two switches share one logical bundle toward a server or downstream switch.

  • LACP fallback/static mode keeps a minimum number of links active if LACP negotiation fails, depending on platform support.

  • Redundant uplinks without aggregation rely on STP; with aggregation, use LACP plus a routing or MLAG design to avoid loops.

Quick reference

Stacking and Switch Virtualization

  • A stack presents one management IP and one control plane; member switches share a distributed forwarding plane.

  • Stack roles: commander/master (control), standby, and members; election is by priority then MAC.

  • Stack cabling is typically a ring for redundancy; a broken ring can split the stack into independent fragments if split detection is not configured.

  • Stack split (brain split) can create duplicate IPs and loops; configure split detection and redundant management paths.

  • Firmware must match across members; a mismatch can prevent a member from joining or cause an unstable stack.

  • Virtual switching (e.g., VSX/MLAG pairs) keeps separate control planes but synchronizes forwarding state, avoiding single-control-plane risk.

  • Stack ports have finite bandwidth; oversubscribing stack links between members causes inter-member traffic to bottleneck.

  • Adding a member with a different model may be restricted by mixed-stacking rules; verify supported combinations before deployment.

Quick reference

Layer 3 Switching, Routing, and IP Services

  • An SVI is a logical Layer 3 interface bound to a VLAN; it needs an IP address and the VLAN must be active with at least one up port (platform-dependent).

  • Inter-VLAN routing options: router-on-a-stick (subinterfaces on a trunk), SVIs on a multilayer switch, or routed ports.

  • Longest-prefix match wins in the routing table; administrative distance breaks ties between protocols; metric breaks ties within a protocol.

  • Static route default AD is 1; connected is 0; OSPF is 110; eBGP is 20; iBGP is 200 (Cisco-style values; verify platform defaults).

  • OSPF: neighbors must match area, hello/dead timers, network type, and authentication; DR/BDR elected on broadcast networks.

  • DHCP relay (helper address) forwards client broadcasts to a server; the relayed packet carries the gateway IP so the server can pick the pool.

  • ARP resolves IPv4 to MAC within a subnet; proxy ARP answers for hosts that believe a remote host is local.

  • VRRP/HSRP provide a virtual gateway IP and MAC; priority and preemption determine the active router.

Quick reference

Network Security and Access Control

  • Port security limits learned MACs per port; violation actions are protect (drop), restrict (drop and log), and shutdown.

  • 802.1X roles: supplicant, authenticator (switch), authentication server (RADIUS); EAPoL carries authentication between supplicant and switch.

  • MAB (MAC Authentication Bypass) authenticates devices that lack an 802.1X supplicant by their MAC against RADIUS.

  • DHCP snooping trusts only designated uplink ports; untrusted ports cannot offer DHCP; the binding table feeds Dynamic ARP Inspection.

  • Dynamic ARP Inspection validates ARP packets against the DHCP snooping binding table to block ARP spoofing.

  • IP Source Guard filters traffic by source IP/MAC binding, preventing address spoofing on access ports.

  • Storm control rate-limits broadcast, multicast, or unknown unicast; exceeding the threshold can drop or shut the port.

  • Management plane hardening: SSH instead of Telnet, AAA with RADIUS/TACACS+, ACLs on management VLAN, and disable unused services.

Quick reference

Multicast, QoS, and Traffic Management

  • IPv4 multicast range 224.0.0.0/4; 224.0.0.0/24 is link-local and never forwarded by routers.

  • IGMP snooping lets a switch forward multicast only to ports that joined the group; without it, multicast floods like broadcast.

  • PIM-SM uses a rendezvous point; PIM-DM floods then prunes; IGMPv2/v3 differ in leave and source-specific join support.

  • QoS pipeline: classification, marking (DSCP 6-bit / CoS 3-bit), queuing, scheduling, and congestion avoidance.

  • DSCP per-hop behaviors: EF (46) for voice, AF classes with drop precedence, CS values for class selectors; default is 0 (best effort).

  • Trust boundaries: trust DSCP/CoS at the access edge only for known devices; re-mark untrusted traffic at the boundary.

  • Weighted random early detection (WRED) drops packets before queue overflow to signal TCP senders; tail drop waits until the queue is full.

  • Shaping buffers and delays excess traffic; policing drops or remarks excess traffic immediately.

Quick reference

Monitoring, Troubleshooting, and Operations

  • Interface counters: runts (<64 bytes), giants (>MTU), CRC errors (physical/cable), late collisions (duplex mismatch), and input/output discards.

  • Duplex mismatch symptom: late collisions, FCS errors, and poor throughput on one side; verify both ends and avoid hardcoding one side.

  • Show commands to sequence: interface status, MAC table, ARP, VLAN/trunk status, STP topology, then routing table.

  • Syslog severity 0 (emergency) to 7 (debug); SNMP traps and RMON provide event and statistics collection.

  • Port mirroring (SPAN) copies traffic to a monitor port; a remote mirror (RSPAN) carries mirrored traffic in a dedicated VLAN.

  • LLDP/CDP reveal neighbor identity, port, and capabilities; mismatched native VLAN or MTU often shows here first.

  • Configuration management: save running config to startup, keep backups, and use version control or a config server for rollback.

  • NTP synchronization is required for accurate log correlation and for time-based ACLs or certificates.