JNCIS-SP cheat sheet

Area 0 is the backbone; every non-backbone area must attach to it directly or through a virtual link whose transit area is common to both endpoints and at least one endpoint sits in area 0.

Stub area blocks Type 5 LSAs and the ABR injects a default as a Type 3 summary; totally stubby (Junos: stub with no-summaries) additionally blocks Type 3 and Type 4 LSAs except the default.

NSSA blocks Type 5 but allows Type 7 LSAs originated inside it by an ASBR; the ABR translates Type 7 to Type 5 at the area boundary and can inject a Type 7 default.

Junos default reference bandwidth is 100 Mbps: cost = reference-bandwidth / interface bandwidth, rounded to an integer with a floor of 1, and a manual metric overrides the calculation.

LSA types: 1 router, 2 network, 3 summary, 4 ASBR summary, 5 external, 7 NSSA external.

Path preference in Junos: intra-area 10, inter-area 10, OSPF external 150 - external routes lose to any internal OSPF route; E1 adds internal cost to the ASBR, E2 uses only the redistributed cost and is the default.

Hello/dead timers: 10/40 s on broadcast and point-to-point, 30/120 s on NBMA and point-to-multipoint; fast hellos and BFD shorten detection without changing the dead-interval relationship.

Authentication is configured per area or per interface as simple (plain text) or MD5 with a key-id, and both neighbors must match type and key.

NET = area ID (1–13 bytes) + 6-byte system ID + 1-byte NSEL of 00, e.g. 49.0001.1921.6800.1001.00; the system ID is what identifies the router in the topology.

Level 1 is intra-area; Level 2 is the inter-area/backbone level. A default L1/L2 router sets the attached (ATT) bit, which lets L1 routers install a default route toward the nearest L2-capable neighbor.

There is no backbone area 0: the L2 topology itself is the backbone, and area boundaries exist only between L1 and L2 routers.

Narrow metrics are 6 bits per interface (max 63) with a 10-bit path total (max 1023); wide metrics are 24-bit per interface (max 16,777,215) and are required for MPLS traffic engineering.

On broadcast links the DIS is elected by highest interface priority, then highest MAC/SNPA, and unlike an OSPF DR it is preemptive - a new higher-priority router takes over.

Adjacency requires matching level, adequate MTU (padding is used on IIH), and for Level 1 a matching area address; IIH PDUs carry these fields.

LSPs are refreshed every 900 s with a 1200 s lifetime by default; CSNP/PSNP PDUs synchronize the link-state database and request missing LSPs.

Junos defaults: hello 10 s / hold 30 s (point-to-point uses a 3× multiplier), and the default metric style is narrow unless wide-metrics-only is configured.

Junos default preference: BGP 170 for both iBGP and eBGP; compare with OSPF internal 10, OSPF external 150, IS-IS L1 internal 15 / L2 internal 18, static 5, direct 0.

Junos best-path order: highest local preference → locally originated → shortest AS path → lowest origin (IGP < EGP < incomplete) → lowest MED → eBGP over iBGP → lowest IGP cost to next hop → lowest router ID → shortest cluster list → lowest peer address.

MED is compared only among paths from the same neighboring AS unless always-compare-med is configured; AS-path prepending lengthens the AS path to influence inbound traffic.

iBGP requires a full mesh or a route reflector: an RR reflects client routes to clients and non-clients but never back to the originator, adding ORIGINATOR_ID and CLUSTER_LIST to prevent loops.

A confederation divides one AS into sub-ASes that use eBGP-like rules between them while presenting a single AS number externally; the internal sub-AS path appears in parentheses and is stripped at the confederation border.

Well-known communities: no-export stops advertisement outside the AS, no-advertise stops advertisement to any peer, no-export-subconfed keeps the route inside a confederation sub-AS; communities are matched with a community policy match.

eBGP sets the next hop to itself on advertisement; iBGP preserves the original next hop, so the border router normally needs next-hop-self, or multihop with an increased TTL for non-directly connected peers.

BGP FSM order: Idle → Connect → Active → OpenSent → OpenConfirm → Established; Active means the router is listening for an inbound TCP 179 connection and retrying.

The shim header is 32 bits: 20-bit label, 3-bit EXP (CoS), 1-bit bottom-of-stack S flag, 8-bit TTL.

LDP uses downstream-unsolicited label distribution for IGP-learned prefixes, so an LSP follows the IGP shortest path; label bindings populate the LFIB (in-label → out-label plus next hop).

RSVP-TE signals explicit-path LSPs with bandwidth reservation via PATH and RESV messages, uses CSPF for path computation, and supports link/node fast reroute with bypass tunnels.

In a label stack the outer (transport) label is swapped by P routers; the inner (VPN/service) label is distributed by MP-BGP and is used by the egress PE to select the VRF.

RD makes overlapping customer prefixes unique in VPNv4 MP-BGP; the route target extended community (target:ASN:value or target:IP:value) controls import and export between VRFs.

Junos VRF minimal config: routing-instances X instance-type vrf, interface, route-distinguisher, vrf-target (or vrf-import/vrf-export policies).

PHP pops the transport label at the penultimate hop so the egress PE does one lookup; explicit-null (label 0) is used when the EXP bits must be preserved for egress QoS.

Label allocation modes: per-prefix, per-nexthop, and per-instance (one label per VRF, which reduces label usage but requires the egress to infer the prefix).

802.1Q inserts a 4-byte tag after the source MAC: 12-bit VLAN ID (usable range 1–4094) plus priority and DEI bits; the native VLAN is forwarded untagged on the trunk.

QinQ/802.1ad stacks an outer S-VLAN (TPID 0x88a8 or 0x8100) over the customer C-VLAN, letting the provider carry many customers with overlapping VLAN IDs.

STP root bridge = lowest bridge ID (2-byte priority in steps of 4096, default 32768, then MAC); the root port is the lowest-cost path to the root, ties broken by lowest upstream bridge ID and port ID.

Legacy 802.1D port states: blocking, listening, learning, forwarding (plus disabled); RSTP collapses these to discarding, learning, forwarding and adds edge ports and proposal/agreement.

RSTP (802.1w) provides fast reconvergence; MSTP (802.1s) maps groups of VLANs to instances so each instance runs its own topology and load balancing is possible.

LAG/LACP (802.3ad): members must match in speed, duplex and MTU; LACP runs active/passive, and hashing across members typically uses L2/L3/L4 fields.

Junos examples: set interfaces ae0 aggregated-ether-options lacp active; set protocols rstp interface all; set interfaces ge-0/0/1 unit 0 family ethernet-switching vlan members VLAN10.

Edge protections: BPDU guard disables a port that receives a BPDU, loop protection blocks a port that sees its own BPDU, and root guard prevents an edge device from becoming root.

Junos policy evaluation is top-down: each term matches (from) then acts (then accept/reject/next term/next policy); a term with no terminating action falls through to the next term.

Junos defaults: the default BGP import policy is accept and the default BGP export policy is reject, so an export policy must explicitly accept routes.

Prefix matching uses route-filter with exact, longer, orlonger, upto and prefix-length-range modifiers; route lists and prefix lists can also be referenced from a policy.

maximum-prefix on a BGP group or neighbor caps accepted prefixes and can warn, log, or tear down the session to protect the routing table.

Graceful restart helper mode keeps forwarding for a restarting neighbor during the restart interval and defers best-path recomputation until the End-of-RIB marker (BGP) or database resync completes.

NSR keeps protocol state synchronized between primary and backup Routing Engines so a failover does not reset BGP/OSPF/IS-IS adjacencies and needs no neighbor support; GR instead requires neighbor helper capability.

BFD declares a session down after multiplier × negotiated interval of missed packets; echo mode offloads fast detection to the forwarding plane and BFD can trigger IGP/BGP/static teardown.

Authentication options: OSPF/IS-IS area or interface MD5, BGP TCP MD5, and key chains with rotation for graceful key rollover.