Networking

Last Updated: 3/17/26 Note to Test Takers: This document is a Summary Reference, not a replacement for a comprehensive course and hands on experience in a lab. I recommend the CBT Nuggets course because Knox is great at explaining networking concepts with the right amount of enthusiasm. Table of Contents Networking Fundamentals Junos OS Fundamentals User Interfaces Configuration Basics Operational Monitoring and Maintenance Routing Fundamentals Routing Policy and Firewall Filters Glossary Lab Recommendations Networking Fundamentals The OSI Model Where it all starts and I’m not sure if its really covered in the JNCIA exam. This is an industry standard crafted back in the 1970s. The model has seven layers from physical up through application. The really important bit is the breakdown of the lower layers as you will see L1, L2, L3, L4 all over the place. These are logical separations of the networking stack L1 - Physical layer - Think cabling and electrical or photonic pulses on said cabling L2 - Data-link layer - Think frames, ethernet, and mac addresses L3 - Network layer - Packets, IPs, and routing L4 - Transport layer - Segments, TCP/UDP, and SYN+ACKs Function of routers and switches Routers use L3 information to forward packets between networks Switches use L2 info to forward packets on the lan Ethernet networks Major concept here is Mac addresses Physical address made up of 48 bits and displayed using hexadecimal format Broadcast address is ffff.ffff.ffff Uses mac addresses to forward ethernet frame Ethernet header + trailing checksum ...

IS-IS IS-IS (Intermediate System to Intermediate System) is a link-state routing protocol commonly used in service provider networks and hey, you’re studying for the JNCIS-SP, so you’re in luck. Like OSPF, it uses the Dijkstra SPF algorithm to compute shortest paths, but it was designed to handle all sorts of traffic. So even though it was designed with CLNP in mind, it can carry things like IPv4, IPv6, and label information for mpls and srv6. ...

OSPF OSPF Packet Types Type Name Purpose 1 Hello Neighbor discovery, DR/BDR election, keepalive 2 DBD LSDB table of contents exchange 3 LSR Request specific missing LSAs 4 LSU Delivers actual LSAs 5 LSAck Confirms LSU receipt LSA Types Type Name Originated By Scope Blocked By 1 Router Every router Area — 2 Network DR Area — 3 Summary ABR Domain Totally Stubby, Totally NSSA 4 ASBR Summary ABR Domain Totally Stubby, Totally NSSA, NSSA, Stub 5 AS-External ASBR Domain All stub and NSSA areas 7 NSSA External ASBR (NSSA) NSSA only — (translated to Type 5 at ABR) Area Types Area Type 3 Type 4 Type 5 Type 7 Default Route Standard Yes Yes Yes No No Backbone Yes Yes Yes No No Stub Yes No No No Yes (auto) Totally Stubby No No No No Yes (auto) NSSA Yes No No Yes No (configurable) Totally NSSA No No No Yes Yes (auto) Adjacency States Down → Init → 2-Way → ExStart → Exchange → Loading → Full ...

Broadcom Switch Chipset Families Broadcom dominates merchant silicon for data center and carrier switching. Their three main ASIC families, Tomahawk, Trident, and Jericho, each make different tradeoffs between bandwidth, feature depth, and buffer size. Most Arista, Cisco Nexus, and Juniper QFX/PTX platforms run one of these under the hood. 1. Tomahawk Series Design philosophy: Maximum port density and throughput at the cost of feature depth. These chips use cut-through forwarding, carry shallow on-chip buffers (~50–100 MB), and support little to no L3 routing table depth. The trade-off is intentional; at spine and AI fabric scale, you want wire-rate forwarding with predictable low latency, not a large TCAM. ...

Form Factor Speed Lanes (Elec) Modulation Connector Types Approx. Intro SFP+ 10 Gbps 1 x 10G NRZ LC Duplex, RJ45 2006 SFP28 25 Gbps 1 x 25G NRZ LC Duplex 2014 SFP56 50 Gbps 1 x 50G PAM4 LC Duplex 2019 SFP-DD 100 Gbps 2 x 50G PAM4 LC Duplex 2019/20 QSFP+ 40 Gbps 4 x 10G NRZ MPO-12, LC 2012 QSFP28 100 Gbps 4 x 25G NRZ MPO-12, LC 2014 QSFP56 200 Gbps 4 x 50G PAM4 MPO-12, LC 2019 QSFP-DD 400G / 800G 8 x 50/100G PAM4 MPO-16, LC, CS 2017/21 OSFP 400G / 800G 8 x 50/100G PAM4 MPO-12/16, LC 2019 OSFP1600 1.6 Tbps 8 x 200G PAM4 MPO, Dual LC 2024/25

OSPF (Open Shortest Path First) OSPF is a link-state interior gateway protocol (IGP). Each router floods Link-State Advertisements (LSAs) describing its interfaces and neighbors. Every router builds an identical Link-State Database (LSDB) and runs the Dijkstra SPF algorithm to compute the shortest path tree. OSPF runs directly over IP (protocol 89) and uses multicast for efficiency. Default route preferences in Junos: OSPF Internal routes: 10 OSPF AS External routes: 150 Terms LSDB (Link-State Database) - The topological database. Within a single area, all routers must have an identical LSDB. SPF (Shortest Path First) - The Dijkstra algorithm each router runs against the LSDB to compute best paths. Router ID (RID) - A 32-bit identifier unique to each OSPF router. Junos selects the RID in this order: explicitly configured → highest active loopback IP → highest physical interface IP. Best practice is to configure it explicitly. ABR (Area Border Router) - A router with interfaces in multiple OSPF areas. Generates Type 3 (Summary) LSAs between areas. ASBR (AS Boundary Router) - A router that redistributes routes from outside OSPF into the OSPF domain. Generates Type 5 LSAs. Backbone Router - Any router with at least one interface in Area 0. Internal Router - All interfaces are in the same single area. Remember, best practice to explicitly configure the router-id. See below: ...

IP Tunnels Tunnels encapsulate one protocol inside another, creating a virtual point-to-point link across a network that wouldn’t otherwise carry that traffic. The encapsulating network is called the underlay; the encapsulated traffic and the logical topology it creates is called the overlay. Both GRE and IP-IP are stateless — they hold no session state and provide no encryption or reliability guarantees. Common use cases: Carry IPv6 traffic across an IPv4-only core (6in4) Carry IPv4 traffic across an IPv6-only core (4in6) Extend IGP adjacencies across a WAN that doesn’t support multicast Tunnel MPLS across a non-MPLS network Bridge Layer 2 domains across a routed network Tunnel Concepts Underlay vs Overlay: ...

High Availability Junos provides a layered HA architecture. Link aggregation handles physical link redundancy. Graceful Restart, GRES, and NSR handle control plane failures at increasing levels of sophistication. BFD accelerates failure detection for all routing protocols. Understanding which technology does what — and what each one requires — is the core exam objective for this topic. Link Aggregation Groups (LAG / LACP) LAG bundles multiple physical interfaces into a single logical ae (aggregated Ethernet) interface, providing both redundancy and increased bandwidth. IEEE standard 802.3ad — not to be confused with 802.1ad (Q-in-Q). ...

IPv6 IPv6 was designed to solve IPv4 address exhaustion while simplifying the protocol. The header is fixed-length and streamlined, broadcast is eliminated in favor of multicast, and address configuration can be fully automatic. For service providers, the most exam-relevant areas are address types, NDP, autoconfiguration, and how routing protocols (OSPF, IS-IS) extend to support IPv6. IPv4 vs IPv6 Key Differences Feature IPv4 IPv6 Address size 32 bits 128 bits Header size Variable (20–60 bytes) Fixed 40 bytes Header checksum Yes No (relies on L4) Fragmentation Routers and source Source only Broadcast Yes No — replaced by multicast Address resolution ARP NDP (ICMPv6) Autoconfiguration DHCP only SLAAC + DHCPv6 IPsec Optional Built into extension header framework IPv6 Header The base IPv6 header is always exactly 40 bytes. It is simpler than IPv4 — no checksum, no options field, and no fragmentation fields (those are handled by extension headers when needed). ...

Layer 2 Bridging and VLANs Service provider networks often need to deliver Layer 2 connectivity between geographically separated customer sites. Junos implements this using bridge domains, which define the L2 forwarding boundaries, and 802.1ad (Q-in-Q) to tunnel customer VLAN spaces across the provider network without overlap. Terms Bridge Domain — a Layer 2 forwarding domain. Like a VLAN. It defines which interfaces share the same broadcast domain and MAC table. EVC (Ethernet Virtual Connection) — the L2 service sold by a SP to a customer. It defines the endpoints of a Layer 2 circuit. C-Tag (Customer Tag) — the inner 802.1q tag. Any VLAN 1–4094 from the customer’s space. S-Tag (Service Tag) — the outer 802.1ad tag. Assigned by the SP to identify the customer. Encapsulates all of that customer’s C-Tags. PBN (Provider Bridge Network) — the entire SP Layer 2 fabric. PEB (Provider Edge Bridge) — the SP edge device. Pushes/pops S-Tags on customer-facing ports. S-VLAN Bridge — an interior SP device that only examines and switches based on the S-Tag. Customer ports — PEB ports facing the customer. S-Tags are applied or removed here. Network ports — interior SP ports that carry double-tagged frames without modification. IRB (Integrated Routing and Bridging) — a logical interface that gives a bridge domain an IP address, enabling the router to act as the default gateway for hosts in that domain. 802.1q The standard VLAN tagging protocol. Inserts a 4-byte tag into the Ethernet frame. ...