Network Engineering · BCIT CISA

Enterprise WAN &
Infrastructure Design

By Austin Webber — Co-authored with Wesley Kitane, Paul Kim, Kaiden Koran & Duc Anh Nguyen

Type Academic Project
Course Advanced Routing Services
Program BCIT CISA
Scope Multi-site SP + CE WAN
Cisco IOS DMVPN MPLS OSPF HSRP MSTP EtherChannel NAT IPv4 / IPv6 Provider Edge Customer Edge

Overview

This project was the final deliverable for BCIT's Advanced Routing Services course, part of the Computing Information Systems Administration (CISA) program. The goal was to design and implement a production-representative network infrastructure spanning both Service Provider (SP) and Customer Edge (CE) layers using Cisco hardware.

The topology simulates a real-world multi-site enterprise WAN: a central Provider Edge router connects to four Customer Edge routers across geographically separate sites, with all inter-site traffic tunnelled securely through DMVPN.

Key Technologies

  • DMVPNDynamic Multipoint VPN — hub-and-spoke tunnels (GRE Multipoint) providing secure site-to-site connectivity through the Provider Edge.
  • MPLSMulti-Protocol Label Switching — label-switched forwarding on CE routers for optimised path selection across the provider backbone.
  • HSRPHot Standby Router Protocol — virtual gateway redundancy across L3 switches; immediate failover if the active router goes down.
  • MSTPMultiple Spanning Tree Protocol — VLAN-aware STP instances for loop prevention and load balancing across the switch block.
  • EtherChannelLogical link aggregation on S1, S2, and S3 — provides fault tolerance and additional bandwidth between switching devices.
  • NATNetwork Address Translation — configured on both PE and CE routers to allow internal networks internet access.
  • OSPFOpen Shortest Path First — routing protocol across all devices for neighbour discovery and route propagation; Area 16.

Topology

The network is structured around a central Provider Edge router (R1) that connects to the internet via DHCP and acts as the NAT boundary and OSPF default-information originator for the entire domain. Four Customer Edge routers (R2–R5) connect to R1 through provider-facing VLANs and form a DMVPN spoke network.

Below R4 and R5 sits the switch block: two Layer 3 switches (S1 and S2) running HSRP and MSTP, interconnected with EtherChannel port-channels, feeding a Layer 2 access switch that connects end devices.

Figure 1 — Network Infrastructure Topology Figure 1 — Network Infrastructure Topology
Topology Summary
R1 (PE) → R2/R3/R4/R5 (CE) via DMVPN tunnels · R4/R5 dual-home site 1 LAN · S1/S2 L3 switch block with HSRP VIP · L2 access switch → end devices. The serial link between R4 and R5 (S0/0/0) provides a direct site 1 redundancy path.

Switch Block

EtherChannel

All three switches use EtherChannel to bundle physical links into single logical port-channels, providing both redundancy and aggregated bandwidth. S1 initiates channels in Desirable / Active mode; S2 responds in Auto / Passive mode for PAgP and LACP compatibility respectively. All port-channel interfaces carry Native VLAN 333 and allow VLANs 10–15, 222, and 333.

DevicePort-ChannelInterfacesModeType
S1Po1G1/0/1-2DesirableLayer 2 trunk
S1Po2G1/0/3-4DesirableLayer 2 trunk
S1Po3G1/0/5-6ActiveLayer 3 routed
S2Po1G1/0/1-2DesirableLayer 2 trunk
S2Po2G1/0/3-4AutoLayer 2 trunk
S2Po3G1/0/5-6PassiveLayer 3 routed
L2-S1Po1Fa0/1-2AutoLayer 2 trunk
L2-S1Po2Fa0/3-4AutoLayer 2 trunk

MSTP & HSRP

Multiple Spanning Tree Protocol is configured with two instances: S1 is the primary root for VLANs 10–11; S2 is the primary root for VLANs 12–14. This distributes STP convergence load and prevents a single root election bottleneck.

HSRP creates a virtual gateway (VIP) for each VLAN. S1 holds active status at priority 110; S2 is standby at priority 90. On S1 failure, S2 immediately assumes the virtual IP and traffic continues without manual intervention.

VLANNameS1 IPS2 IPHSRP VIP
V10Data10.16.10.25410.16.10.25310.16.10.252
V11Voice10.16.11.25410.16.11.25310.16.11.252
V12Server10.16.12.25410.16.12.25310.16.12.252
V13Wireless10.16.13.25410.16.13.25310.16.13.252
V14Guest10.16.14.25410.16.14.25310.16.14.252
V15NetMgmt10.16.15.25410.16.15.25310.16.15.252

Routing & WAN

OSPF

OSPF Area 16 runs across all devices. R1 uses default-information originate to push the internet default route into the OSPF domain. All CE routers advertise their site-local networks (10.16.0.0 and 198.18.0.0 summaries) via OSPF, ensuring full reachability between sites without static routing.

DMVPN

Dynamic Multipoint VPN Phase 2 tunnels are established between all CE routers through the Provider Edge. R2 acts as the hub router; R3, R4, and R5 are spokes. Tunnel interfaces use GRE Multipoint mode with G0/0/0 as the tunnel source on each CE router. This allows spoke-to-spoke traffic to build dynamic tunnels on demand rather than routing all inter-site traffic through the hub.

RouterRoleTunnel IPSite Network
R2Hub50.16.1.210.16.100.0/24
R3Spoke50.16.1.310.16.200.0/24
R4Spoke50.16.1.410.16.1.0/24 (Site 1)
R5Spoke50.16.1.510.16.2.0/24 (Site 1)

MPLS

Multi-Protocol Label Switching is enabled on CE routers (R2–R5) on both their WAN-facing and LAN-facing interfaces. Labels are applied to each packet, allowing the provider backbone to forward traffic along the shortest label-switched path rather than performing full IP lookups at each hop.

NAT Configuration
NAT is configured on both R1 (PE) and each CE router. Internal interfaces (VLANs 11–14 on R1; LAN interfaces on CE routers) are marked ip nat inside; the DHCP-facing WAN interface on R1 is marked ip nat outside. This allows all internal hosts to reach the internet through a single translated address.

IP Addressing

DeviceInterfaceIPv4 AddressIPv6 Address
R1 (PE)G0/0/1DHCP
G0/1/0 (V11)198.18.1.254/242001:198:18:1::254/64
G0/1/1 (V12)198.18.2.254/242001:198:18:2::254/64
G0/1/2 (V13)198.18.3.254/242001:198:18:3::254/64
G0/1/3 (V14)198.18.4.254/242001:198:18:4::254/64
R4 (CE)G0/0198.18.1.1/242001:198:18:1::1/64
G0/1 (Site 1)10.16.1.254/242001:10:16:1::254/64
S0/0/010.16.4.100/242001:10:16:4::100/64
R5 (CE)G0/0198.18.2.1/24
G0/1 (Site 1)10.16.2.254/242001:10:16:2::254/64
S0/0/010.16.4.200/242001:10:16:4::200/64
R2 (CE)G0/0/0198.18.3.1/242001:198:18:3::1/64
G0/0/1 (Site 2)10.16.100.1/242001:172:16:1::1/64
R3 (CE)G0/0/0198.18.4.1/242001:198:18:5::1/64
G0/0/1 (Site 3)10.16.200.1/242001:172:16:2::1/64
S1G1/0/24 Routed10.16.1.1/242001:10:16:1::1/64
S2G1/0/24 Routed10.16.1.2/242001:10:16:1::2/64

Results

All primary objectives were successfully verified through ping tests, traceroutes, and deliberate failure injection.

Layer 2 — Switching

  • MSTPVLAN 10 traffic routed through S1 (primary root); VLAN 30 traffic routed through S2 — confirming instance-based load balancing.
  • EtherChannelShutdown of G1/0/5 on Po3 (routed link between S1 and S2) — S1 maintained connectivity to S2 through the surviving port, confirming fault tolerance.

Layer 3 — Routing & WAN

  • OSPFFull route propagation and neighbour adjacency confirmed across all routers. All sites reached each other, the ISP, and the internet.
  • DMVPNSuccessful pings through GRE tunnels from all spokes to the hub and between spokes.
  • HSRPS1 powered down — end device on VLAN 10 maintained connectivity to the R2 site via S2 assuming the virtual IP.
  • MPLSTraceroute confirmed label-switched forwarding; each CE router selected the shortest label path.
Outcome
All Layer 1, 2, and 3 objectives passed verification. End-to-end connectivity was confirmed from workstation → access switch → distribution → CE router → DMVPN tunnel → internet.

Known Limitations & Improvements

R4 Redundancy Gap
No direct link was implemented between R4 and R1. Traffic destined for R4 is dropped when the R4 → R5 serial path fails, as no floating static route exists to re-route it. Fix: add a direct serial link between R4 and R5 and configure OSPF floating static routes to restore redundancy.
  • OSPFOSPF neighbour adjacency uses broadcast mode. Switching to multicast would reduce hello traffic overhead and improve convergence speed across the domain.
  • HSRPSerial link IP addresses on R4 and R5 were not configured — HSRP tracking against these interfaces is not possible until addressing is applied.