Lessons (6, study in order)

Lesson 1 - What Layer 1 Is: OTN/ODF Basics

Beginner · about 8 min

Goal: First see how this module differs from the others, and meet the two most basic physical roles.

This module isn't the same kind of thing as the others

The other modules (OSPF, BGP, VLAN...) all teach how some device makes a forwarding decision based on a field: look up a MAC table, look up a routing table, check a VLAN tag.

★ Layer 1 has none of that. There's no table to look up, no field to read -- which fiber a circuit takes, which devices it passes through, was physically wired during installation, not "computed" by any device. What this module is about isn't an algorithm; it's "exactly which devices does this circuit physically pass through."

★ ODF: the least glamorous, most essential link in the room

An Optical Distribution Frame (ODF) is a patch panel full of fiber connectors. Every fiber jumper in the room terminates here first, then gets cross-connected with a patch cord.

It's a purely passive device -- no amplification, no conversion, no awareness of what signal is inside. It exists purely to keep the cabling tidy and easy to re-patch later -- this is literally where the "how do they connect" in this module's title happens.

★★ The transponder: where client-side meets line-side

A router emits a client signal (e.g. 100GE) over a short-reach "grey" optic. But the trunk fiber between two IDCs carries long-reach "colored" light at specific wavelengths.

The transponder is the boundary between the two: it decodes the client signal, re-frames it with overhead and error correction, then re-transmits it on the line port at one specific wavelength using a tunable laser -- that wavelength is this circuit's identity on the trunk fiber.

Try it: walk through the simplest possible circuit

Go to the Circuit Provisioning page (the two-site scenario is already loaded), pick the two end routers, provision a circuit, and read through the hop-by-hop explanation -- router -> ODF -> transponder -> ROADM -> ... -> ROADM -> transponder -> ODF -> router, not one device skipped.

Key takeaways

• There's no such thing as "computing a path" at layer 1 -- the path was physically wired during installation
• ★ An ODF is purely passive -- it only makes physical patch connections
• ★★ The transponder is the only device that does O-E-O conversion -- it's the boundary between client-side and line-side

Lesson 2 - Wavelengths and DWDM

Beginner · about 8 min

Goal: Understand why a single fiber can carry several mutually independent circuits at once.

A wavelength is just "a color of light"

Infrared light invisible to the eye still has different "colors" at different wavelengths. A transponder's tunable laser can emit many different wavelengths -- each one carries one independent signal, none interfering with the others, much like different radio stations broadcasting on different frequencies through the same air.

★ DWDM: packing many "radio stations" into one fiber

Dense Wavelength Division Multiplexing (DWDM) transmits multiple wavelengths through the same fiber at once. The receiving end just needs a matching filter structure to "pick out" one wavelength and recover that signal -- every other wavelength is completely unaffected.

★ This is also why a ROADM can "drop this wavelength, express that one" -- they are, physically, signals coexisting in the same fiber, but logically completely independent.

This module's simplification: 4 channels, not 40/80/96

Real DWDM systems in the C-band commonly support 40, 80, even 96 channels (100GHz or 50GHz spacing). This module simplifies that down to 4 channels -- so that a resource-planning decision like "we ran out of wavelengths" can be reached in a few steps, instead of needing 40 circuits provisioned first.

Key takeaways

• A wavelength is essentially a color of light -- different colors can travel through the same fiber at once without interfering
• ★ DWDM lets one fiber carry several independent signals at once
• This module simplifies the real 40/80/96 channels down to 4 -- a teaching simplification, not a real system's capacity

Lesson 3 - How a ROADM Decides Where a Line Goes

Advanced · about 12 min

Goal: Understand add / express / drop, and how a path gets picked on a ring -- the core of this module.

Every wavelength at every ROADM has exactly one of three fates

Add: a wavelength from a local transponder is added onto the trunk
Drop: a wavelength is taken off the trunk and delivered locally
Express: neither added nor dropped -- it passes straight through, all-optically, and keeps going

★★★ The key point: express involves no O-E-O conversion at all. No matter how many express ROADMs a circuit passes through, the only extra delay is the light physically traveling that distance in glass -- no electrical-layer processing overhead.

★★ On a ring, diagonally-opposite sites have no direct fiber

On a four-site ring, adjacent sites (e.g. North<->East) are directly connected -- one hop, no express needed. But the two sites diagonally across the ring (e.g. North<->South) have no direct fiber at all -- a circuit between them must pick a direction around the ring and express through the ROADM in between.

And the ring's two directions often have different total lengths -- when provisioning, the system automatically picks the shorter one.

★★★ Try it: provision a circuit that has to go around the ring

Go to the Circuit Provisioning page (the four-site ring scenario is already loaded), provision a North <-> South circuit -- these two are diagonal, with no direct fiber. Check the hop-by-hop explanation: is the ROADM in the middle labeled "Express", which direction did the system pick, and what's the total distance?

Key takeaways

• ★★★ Every wavelength at every ROADM does exactly one of Add / Drop / Express
• ★★ Express does no O-E-O conversion at all -- however many express nodes in a row, the processing overhead is 0
• On a ring, non-adjacent sites must go around one direction and express through intermediate nodes to reach each other

Lesson 4 - Distance and Amplification: Why OLAs Exist

Beginner · about 8 min

Goal: Fiber attenuates signal, and past a certain point the receiver can't read it anymore -- that's exactly what an OLA solves.

Optical power fades with distance

Fiber isn't perfect -- signal power drops as it travels (typically about 0.2-0.25dB per kilometer, plus connector loss). Past a certain point, the receiving optic can no longer read the signal.

★ A relay station roughly every 80-100km

An OLA amplifier site contains a span of erbium-doped fiber (an EDFA) that boosts the signal as a whole -- every wavelength together, without distinguishing or altering any of them.

Check the Distance & Amplification page for the long-haul scenario's budget table: 450km of trunk fiber, an amplifier site every 90km, four of them in total -- without amplification, the optical power would fade below the receiver's sensitivity before arriving.

Key difference: an OLA can't add or drop any wavelength

OLAs and ROADMs both look like "a node in the middle," but their jobs are completely different -- a ROADM decides add/drop/express independently per wavelength; an OLA only amplifies as a whole, and doesn't even know how many wavelengths are inside or whose circuits they belong to.

Key takeaways

• Optical power fades with distance -- typically an amplifier is needed roughly every 80-100km
• An OLA is purely a "relay station" -- it amplifies as a whole and makes no wavelength-level decisions
• Don't confuse OLA with ROADM: whether it can Add/Drop is the fundamental difference

Lesson 5 - Protection and Failure: What Happens When Fiber Gets Cut

Advanced · about 10 min

Goal: An unprotected circuit just goes down when its fiber is cut; ring protection switches over automatically.

Unprotected: cut the fiber, and the service is down

If a circuit has only one physical path, any fiber cut anywhere on that path takes the service down immediately, and it stays down until the fiber is physically repaired or someone intervenes manually.

★★ Ring protection (1+1): both directions lit at once

On a ring topology, a protected circuit is lit simultaneously in both directions -- the same wavelength, two completely disjoint paths.

★★★ The key point: this is not "recompute a path after a failure" -- both paths have already been running since the moment the circuit was provisioned. If one is cut, the other is already hot standby, and the switchover is automatic and millisecond-scale -- exactly why telecom networks favor ring topologies.

★★★ Try it: provision a protected circuit, then cut it

Go to the Circuit Provisioning page, provision a circuit with "ring protection" checked, then click a fiber span on the primary path in the canvas and take it down. Check the circuit's status -- it should stay up, just marked as "running on the alternate path". Compare with an unprotected circuit doing the same thing -- that one just goes down.

Key takeaways

• An unprotected circuit's availability depends entirely on one physical path's luck
• ★★ 1+1 protection has both paths lit already -- switchover needs no "recomputation", just a switch
• The topology must actually contain a disjoint alternate path for protection to mean anything -- checking a box doesn't create one

Lesson 6 - Wavelength Conflicts: A Shared Fiber's Capacity Is Finite

Troubleshooting · about 10 min

Goal: How many circuits a shared trunk fiber can carry at once equals how many channels it supports.

Why a "conflict" happens

How many mutually-independent circuits a trunk fiber can carry at once depends on how many channels it supports -- this module simplifies that to 4 (real systems have 40/80/96). Each circuit occupies one channel; once the channels are used up, that fiber is genuinely "full" -- it's not a configuration problem.

★★ Try it: the 4th succeeds, the 5th doesn't

Go to the Circuit Provisioning page (the wavelength-conflict scenario is already loaded) -- this shared trunk fiber already carries 3 circuits, occupying channels 1-3. Provision a 4th: it succeeds, using the last free channel. Provision a 5th: it fails, reporting that every channel is occupied.

★ That failure message isn't a bug -- it means this fiber's capacity is genuinely used up.

What to do once it's full

In practice the usual options are: add another fiber pair (a physical build, slow and expensive), upgrade to higher-order multiplexing (pricier equipment, more channels on the same fiber), or move part of that route's traffic to a different physical path entirely -- which one to pick is a capacity-planning question this module doesn't go into.

Key takeaways

• A shared fiber's capacity equals its channel count -- once used up, it's genuinely used up, not a misconfiguration
• ★★ The failure message is a teaching signal by design, not an exception