Articles
45 posts
Windows Server Lab: Active Directory, SQL Server, and IIS
How a four-server Windows lab with Active Directory, IIS, SQL Server, batch jobs, and monitoring fits together, and the operations exercises it supports.
CCNP ENCOR: BGP Neighbors, Next Hops, and the Best Path Algorithm
BGP connects autonomous systems and selects the best path using next-hop reachability, weight, local preference, AS-path prepending, and MED.
CCNP ENCOR: Multi-Area OSPF, Stub Areas, and OSPFv3
OSPF builds a shared link-state database and calculates best paths across multi-area OSPFv2, stub areas, summarization, filtering, and OSPFv3.
CCNP ENCOR: EIGRP Metrics, Feasible Successors, and Variance
EIGRP forms adjacencies, calculates its composite metric, and uses DUAL, feasible successors, and variance for fast failover and unequal-cost load balancing.
CCNP ENCOR: Spanning Tree Essentials
STP prevents Layer 2 loops through root elections, path costs, and blocked links, while Rapid-PVST+ speeds convergence and MST scales across VLANs.
CCNP ENCOR: VLANs, Trunking, and EtherChannel
Cisco switching fundamentals come together in hands-on labs with VLANs and SVIs, trunking and router subinterfaces, and static and LACP EtherChannels.
From Vibe Coding to Spec-Driven Development with OpenSpec
I vibe coded for a long time and liked it, but I had no real method. After an Art of Network Engineering episode on spec-driven development, I tried OpenSpec. Here is how vibe coding, SDD, and a few other frameworks compare.
JNCIP-ENT Mega Lab Part 4: DC1 EVPN-VXLAN Fabric
Part 4 builds a self-contained data-center fabric on six vQFX switches: an OSPF underlay with ECMP loopback reachability, an EVPN iBGP overlay with both spines as route reflectors, Layer 2 stretch for two VLANs with anycast IRB gateways, a TENANT1 VRF routed with pure Type 5 over L3 VNI 9999, all-active ESI-LAG multihoming with DF election, and a four-way border handoff back into the AS 65000 enterprise core.
JNCIP-ENT Mega Lab Part 3: Metro Ethernet
Part 3 extends customer VLAN 310 from headquarters to Branch 2 through a Metro Ethernet service. The customer switches retain one VLAN, the provider maps it into service VLAN 3000 with selective Q-in-Q, Layer 2 protocol tunnelling carries customer control frames, and BR2-EDGE01 terminates the subnet on a tagged routed interface.
JNCIP-ENT Mega Lab Part 2: Branch WAN and Basic Multicast
Part 2 extends the mega lab with two different branch WAN designs. Branch 1 uses redundant eBGP sessions through a provider AS and prefers WAN-PE01 with local preference, while Branch 2 uses two routed handoffs in normal OSPF area 1 with unequal costs. A single static RP then supports one PIM sparse-mode ASM service between a source at Branch 2 and a receiver at headquarters.
JNCIP-ENT Mega Lab Part 1: Internet Edge and Core
This is the first build part of the mega lab. It has two providers peering with each other, two enterprise border routers and two core routers, all dual-stack. OSPFv2 and OSPFv3 carry the loopbacks in area 0. The provider transit links stay outside the IGP, so the border routers need next-hop-self. The core routers work as route reflectors in one cluster. Local preference selects the outbound provider, AS path prepend influences the inbound provider, and communities show which border a route entered through. The export policy finishes with a reject term, so the enterprise does not become a transit AS.
JNCIP-ENT: EVPN-VXLAN Layer 3 Gateway and Multihoming
The second EVPN lab in the JNCIP-ENT series. We take the bridged fabric from the previous lab and add Layer 3. The leaves get anycast IRB gateways with one shared virtual IP and MAC, then a tenant VRF and an L3 VNI turn the fabric into symmetric routing with Type 5 routes, so a leaf only holds the VNIs of its own subnets. Then we multihome a host to both leaves with an ESI-LAG and watch EVPN elect a designated forwarder with Type 4 and enable aliasing with Type 1.
JNCIP-ENT: EVPN-VXLAN Fabric and Layer 2 Gateway
The first EVPN lab in the JNCIP-ENT series. We build a two-spine, two-leaf IP fabric on vQFX, make the leaf loopbacks reachable with OSPF and ECMP, and run iBGP with EVPN signaling over it with both spines as route reflectors. The leaves become VXLAN tunnel endpoints and Layer 2 gateways for VLAN 100 as VNI 5100, and two hosts in the same subnet reach each other across the routed core. We verify the Type 3 routes that build the flood list and the Type 2 routes that carry the host MAC addresses.
JNCIP-ENT: Enterprise Routing Policy Construction
The fourth BGP lab in the JNCIP-ENT series. We combine the policy techniques from the earlier labs into one policy set on both border routers. We advertise an aggregate in place of the loopback /32s and suppress the specifics, filter and classify inbound routes with route filters, AS-path regular expressions and communities, then build a policy chain that advertises nothing because the marking term takes no terminating action. We correct it under commit confirmed and validate the result with test policy.
JNCIP-ENT: BGP Route Reflection and Communities
The third BGP lab in the JNCIP-ENT series. We remove one session from the IBGP full mesh to demonstrate the IBGP advertisement rule, then configure R2 as a route reflector and examine the cluster list and originator ID on the reflected route. We add a customer LAN behind R1 and a third AS behind R4, use no-advertise and no-export to stop propagation at two different boundaries, and use a custom community to drive local preference, applying path control by class rather than by prefix.
JNCIP-ENT: BGP Path Attributes and Path Selection
The second BGP lab in the JNCIP-ENT series. We add a second border router to AS 65100, so the same external prefix arrives over two paths. We find out which path the routers select by default and why, then use local preference to set the exit and MED and AS-path prepend to influence the entrance. We also look at origin manipulation, and check every change with a traceroute in both directions.
JNCIP-ENT: BGP Peering and Next-Hop Resolution
The first BGP lab in the JNCIP-ENT series. We build a two-AS topology with EBGP between AS 65100 and AS 65200 and a full-mesh IBGP inside AS 65100, established over loopbacks. Because the EBGP transit subnet is not advertised into OSPF, EBGP-learned prefixes are installed as hidden routes on the non-border routers due to an unresolvable next-hop. We examine this behavior, resolve it with next-hop self and an alternative IGP-based method, and demonstrate the rule that an IBGP-learned route is not re-advertised to another IBGP peer.
JNCIP-ENT: Advanced OSPF Options
The ninth lab in my JNCIP-ENT series. We add a sixth vMX and a new Area 3 that has no connection to the backbone. First we rescue it with a virtual link, then we replace the virtual link with a real backbone link and a multi-area adjacency (RFC 5185). After the topology is fixed, the lab moves to control: prefix-export-limit, the import policy trap, redistribution preference loops, reference bandwidth, and the overload bit.
JNCIP-ENT: OSPFv3 and OSPF Authentication
The eighth lab in my JNCIP-ENT series. The five-router multi-area domain becomes dual-stack without replacing OSPFv2. OSPFv3 has its own neighbors, LSDB, IPv6 routes, LSA types, and area summary. The lab then secures OSPFv2 with MD5 and key rotation, and OSPFv3 with an IPsec security association.
JNCIP-ENT: OSPF Stub and NSSA Areas
The seventh lab in my JNCIP-ENT series. I take the multi-area OSPF domain from Lab 6 and change the area types without touching any cable. Area 1 becomes a stub and then a totally stubby area, and Area 2 becomes an NSSA and then a totally NSSA. At each step we look at how the LSA database becomes smaller: the Type 4 and Type 5 leave the stub, and R5's externals change into Type 7 and then get translated back to Type 5 at the NSSA ABR.
JNCIP-ENT: Multi-Area OSPF Summarization
In this lab, I build a five-router multi-area OSPF topology and use it to understand LSAs in the backbone and route summarization on both the ABR and ASBR.
VXLAN IRB and Anycast Gateway
This part of the VXLAN series explores how to route between VXLAN-backed VLANs and simplify host gateways with anycast.
Classifying and Marking Customer Traffic with Junos CoS
This follow-up to the policing lab adds DSCP classification and rewrite, showing how PE-1 maps customer markings to forwarding classes on ingress and stamps new values on egress.
How to Build a $5 SNMP Temperature Probe
Building a small ESP8266 and DS18B20-XH temperature probe that exposes temperature over SNMP, Prometheus metrics, syslog, and a simple web UI.
Extending VLANs with EVPN VXLAN
Simple lab guide for stretching two bridge domains across three Arista cEOS switches with EVPN, where the red VLAN uses a different VLAN tag on each switch.
JNCIP-ENT: RSTP, MSTP, and VSTP
Comparing RSTP, MSTP, and VSTP in a four-switch Junos lab, with root placement and verification that load-shares VLAN groups across two uplinks.
Extending a Single VLAN with Static VXLAN
Extending VLAN 100 across a routed underlay with static VXLAN on two Arista cEOS switches.
JNCIP-ENT: Understanding Q-in-Q
I expected a clean Q-in-Q lab for JNCIP-ENT, only to find out that vQFX accepts the config but does not actually pass the traffic.
JNCIP-ENT: Understanding Multiple VLAN Registration Protocol (MVRP)
An MVRP lab on vQFX covering configuration, verification, and dynamic VLAN pruning behavior between two switches.
Broke My Server Trying to Clean Up Files
I reclaimed nearly 20GB of Docker data, saw CPU hit 100%, panic-shut down the VM, and ended up with a corrupted filesystem.
Building an ISP Access Lab with PPPoE, DPI, and TR-069
A hands-on ISP access lab showing how PPPoE subscriber access, DPI visibility, and TR-069/GenieACS device management fit together.
JNCIA-DevOps: Junos REST API Automation
Junos has a built-in REST API. It runs over plain HTTP, exposes RPC calls as endpoints, and accepts config as XML in the request body. Because it's just HTTP, you don't need a Juniper-specific library: curl works, `requests` works, anything that can send a POST works. This lab covers reading operational state and pushing config to three routers without opening a single SSH session.
Junos Automation With PyEZ
After passing JNCIS-ENT I needed a break from routing protocols. Weeks of OSPF, BGP, IS-IS, and policy chains. I like networking but I also like doing something different for a while. JNCIA-DevOps looked like the right move: automation focus, still Juniper, and it would give me my third JNCIA cert. I have found out that there is a JNCIA 3x badge and I had been eyeing it for a while. JNCIA-DevOps was the only cert standing between me and it.
JNCIS-ENT + NSE4: From Firewall to ISP
FortiGate HA active-passive, firewall policies for BGP transit and NAT'd internet access, OSPF integration, and R3 as the simulated ISP with eBGP.
JNCIS-ENT: Core Routers, Tunnels, and the Branch
R1/R2 L3 core build: 802.1Q subinterfaces, VRRP split active-active, OSPF+BFD, IS-IS, iBGP/eBGP via FortiGate, GRE/IP-IP branch tunnels and VLAN20 FBF steering.
JNCIS-ENT: Building the L2 Foundation
Building the JNCIS-ENT Layer 2 foundation on vQFX with VLANs, LACP LAG, RSTP, and practical L2 security controls.
How to Enable PNETLab Offline Mode When Online Login Is Down
How to force-enable PNETLab offline mode from the database when the online login servers are unreachable.
I Built an OCR Tool That Reads Hardware Labels So I Don't Have To
How OCR-assisted label scanning turned a tedious data center inventory task into something almost pleasant.
My Disk Manager Script Grew Up and Got a Web Interface
A while ago I wrote about disk.py, a Python script I built to manage a 12-slot SAS backplane after HDSentinel kept losing track of which disks were in which slots. It used PHY symlink paths instead of /dev/sdX names, ran health checks with explicit targets, formatted disks in parallel, and had a terminal-based progress monitor that polled sg_requests independently per device so you could actually read the output. I was pretty happy with it. Since then it grew a web UI and a logs tab.
Per-VLAN BGP Path Steering with Filter Based Forwarding
A case presented at work got me into PNETLab building per-VLAN BGP path steering on a QFX with FBF and virtual-router instances.
Junos QoS: Policing Customer Traffic
Finally facing QoS instead of skipping it. Building a simple policing setup to enforce customer bandwidth limits on Junos.
4-Tunnel IPSec Redundancy Between Two SRX Chassis Clusters
Building four IPSec tunnels across dual ISP links between SRX chassis clusters so traffic survives single-uplink failures with predictable failover.
I Built a Disk Management Tool Because HDSentinel Kept Losing My Disks
I am back with another highly specific script built to solve a problem that probably only exists in my environment.
High Availability IPsec Using Juniper SRX Chassis Cluster
Taking a working IPsec site-to-site VPN and making it highly available with SRX chassis cluster. Active/passive failover in under three seconds.
Routed DCI Between Two EVPN VXLAN Fabrics
Getting full routed connectivity between two EVPN VXLAN fabrics over an MPLS WAN core.