Welcome to the course's gym. Modules 1 through 6 gave you knowledge and method; this module gives you what was still missing: fluency. We start at the foundation — the concepts from module 1 — with exercises set in two scenarios: the by-now familiar Grupo Meridiano and a brand-new company, Clínica Azahar, a physiotherapy clinic with headquarters in Alicante and two branch clinics (Elche and Torrevieja) that will stay with us for the whole module. The new scenario is deliberate: if you can only reason about the network you already know by heart, you don't know how to reason about networks yet.

• Solve before you look. Each exercise has its solution right below it; cover it (or don't scroll) until you have your answer written down. Reading other people's solutions produces the feeling of knowing, not knowledge.
• Write down the reasoning, not just the result. In real life nobody asks you "how many links does a 5-node mesh have?": they ask "is it worth it for us?", and that demands the why.
• If you get one wrong, review before moving on. Each exercise marks with 📌 which lesson to review. Getting something wrong is not a problem; dragging it along with you is.
• The difficulty is progressive: the first exercises are single-concept; the last ones combine several.

Clínica Azahar hands you the inventory of its headquarters. Classify each item as an end device, an interconnection device or a transmission medium:

1. The reception PC.
2. The switch in the comms cabinet.
3. The twisted-pair cable running from that switch to the reception PC.
4. The server where patients' medical records are stored.
5. The router that provides the Internet connection.
6. The printer in treatment room 2.
7. The Wi-Fi access point in the waiting room.
8. The radio waves between that access point and a physiotherapist's tablet.

📌 Review if you struggled: lesson 01-01.

Frequent mistakes: classifying the server as "interconnection" because "it serves everyone". The criterion is not how many it serves, but whether it is a source or destination of the data (end device) or merely moves it along (interconnection). A server is as much an end device as a PC. Same goes for the printer.

For each of Clínica Azahar's needs, choose between twisted pair, optical fiber or wireless, and justify your choice:

1. Connecting the 12 fixed workstations on the ground floor to the switch (distances of 5 to 40 m).
2. Linking the main building with the annex rehabilitation building, 350 m away.
3. Providing connectivity for the tablets the physios carry from room to room.
4. A run that has to pass right next to the electrical panel and the elevator motors.

📌 Review if you struggled: lesson 01-01.

1. Twisted pair. Short distances (well below its ~100 m limit), low cost, simple installation: it is the standard for fixed office workstations.
2. Optical fiber. 350 m exceeds twisted pair's reach (~100 m); fiber covers that distance without breaking a sweat and leaves capacity headroom for the future as well.
3. Wireless (Wi-Fi). The requirement is mobility: no guided medium offers it. You accept lower stability than cable in exchange.
4. Optical fiber. It transmits light, not electricity, so the electromagnetic interference from motors and electrical panels doesn't affect it. Twisted pair (electrical signals) would suffer exactly there.

Frequent mistakes: answering "Wi-Fi" for point 2 "because you don't have to run cable". At 350 m and between buildings, a Wi-Fi link is possible but fragile (obstacles, weather, interference); when you have to guarantee service between buildings you own, the guided medium wins.

Azahar's headquarters signs up for 600 Mbps fiber. Every night they upload a 1.5 GB (≈1500 MB) backup to an external service.

1. What is the theoretical maximum speed in MB/s?
2. How long would the backup take under ideal conditions?
3. The Torrevieja branch clinic has 300 Mbps and uploads the same backup. How long does it take?

📌 Review if you struggled: lesson 01-01.

1. Bandwidth is advertised in megabits per second and files are measured in megabytes: 1 byte = 8 bits, so you divide by 8. $$600 \text{ Mbps} \div 8 = 75 \text{ MB/s}$$
2. $$1500 \text{ MB} \div 75 \text{ MB/s} = 20 \text{ seconds}$$
3. $$300 \div 8 = 37.5 \text{ MB/s} \qquad 1500 \div 37.5 = 40 \text{ seconds}$$

In practice it will take somewhat longer (protocol headers, other uses of the line, the remote service), but that is the right order of magnitude.

Frequent mistakes: the classic "I pay for 600 megs but I download at 75 — they're stealing 525 from me!". No: 600 Mb/s is 75 MB/s; they are the same quantity in different units. Before complaining to your carrier, divide by 8.

Two complaints land on Azahar's IT person the same day. For each one, decide whether the problem is bandwidth or latency, and justify it with the data:

1. "Video calls with the Elche branch clinic keep stuttering and there's echo". A speed test on that same PC shows 580 Mbps down. A ping to the videoconferencing server shows times swinging between 40 and 300 ms.
2. "Downloading a 2 GB X-ray study from Torrevieja takes a quarter of an hour". Torrevieja has an old 20 Mbps line.

📌 Review if you struggled: lesson 01-01.

1. Latency (and its variability, the jitter). There is bandwidth to spare (580 Mbps for a video call that needs a few Mbps), but packets take a long time and, worse still, a different time every trip (40–300 ms): the voice arrives in fits and starts. A 10-lane highway is no use if every car takes an unpredictable time to arrive.
2. Bandwidth. Check the math: 20 Mbps ÷ 8 = 2.5 MB/s; 2000 MB ÷ 2.5 MB/s = 800 s ≈ 13.3 minutes. The observed "quarter of an hour" is exactly what the line delivers: there is no fault, there is a narrow pipe. The fix is to buy more capacity, not to "look into the network".

Frequent mistakes: believing that buying more megabits fixes complaint 1. Interactive applications (video calls, remote desktop, gaming) live off latency; large transfers (backups, downloads) live off bandwidth. Step one, always: identify which of the two your application lives off.

Classify each scenario as a PAN, LAN, WLAN, MAN or WAN:

1. The Bluetooth headphones connected to Marta's phone.
2. The wired network of Meridiano's Valencia office.
3. The laptops connected to the Wi-Fi access point in the Valencia meeting room.
4. The city council's own fiber network linking its libraries across the city.
5. The connection between Meridiano's Valencia and Bilbao sites.
6. The Internet.

📌 Review if you struggled: lesson 01-02.

1. PAN: personal reach, a few meters around one person.
2. LAN: a local network, one building or office, managed by the organization itself.
3. WLAN: it is the same local network, in its wireless variant.
4. MAN: city-wide scope, more than a LAN but less than a WAN.
5. WAN: it links geographically distant locations (Valencia–Bilbao) crossing third-party infrastructure.
6. WAN: the WAN par excellence — the network of networks that spans the planet.

Frequent mistakes: hesitating between LAN and WLAN in point 3. The WLAN is not "another network": it is wireless access to the LAN. You classify it as WLAN when the question is about the type of access; the laptops in the meeting room are on the same local network as the wired PCs.

State whether each situation follows the client-server or the P2P (peer-to-peer) model, and what role each machine plays:

1. Marta opens intranet.grupomeridiano.example from her PC.
2. Two architects, on a construction site with no infrastructure, connect their laptops directly over Wi-Fi to exchange drawings.
3. At Azahar, all the tablets look up the medical records held on the central server.
4. A file-sharing application in which every machine downloads chunks from other machines while sharing them at the same time.

Then answer: why should Azahar not keep the medical records "each on its own tablet, shared between peers"?

📌 Review if you struggled: lesson 01-02.

1. Client-server. Marta's PC is the client (it asks); the server 192.168.10.10 is the server (it answers).
2. P2P. There is no server: both laptops request and offer interchangeably — symmetric roles.
3. Client-server. Tablets = clients; medical records server = server.
4. P2P. Every node is client and server at once.

On the final question: with the data scattered P2P across tablets there would be no single place to secure, back up and control access to sensitive information (medical records!). The client-server model concentrates the data where it can be protected: centralized backups, control over who accesses what, a tablet breaks and nothing is lost. The price is that the server becomes a critical point — which is why it gets redundancy and maintenance.

1. Classify as a private network or a public network: (a) the LAN at Azahar's headquarters; (b) the Internet; (c) Meridiano Bilbao's 192.168.20.0/24 network.
2. Jon is working from home today and needs to get into the intranet. What type of VPN does he need?
3. The Elche and Torrevieja branch clinics need permanent, all-day access to the medical records server in Alicante. Which type of VPN is the right fit, and why not the other one?

📌 Review if you struggled: lesson 01-02.

1. (a) Private: access restricted to the organization. (b) Public: shared, open infrastructure. (c) Private — and note the clue: 192.168.x.x is a range reserved precisely for private networks.
2. A remote-access VPN: an individual user, from any location, establishes a tunnel from their machine to the company network. It is the "person-to-network" VPN.
3. A site-to-site VPN: it joins entire networks permanently, router to router; the machines in Elche work against Alicante without installing or activating anything, as if it were all one network. Remote access doesn't fit: it would force every employee at every branch clinic to bring up their own tunnel every day, for a need that belongs to the whole site and never stops.

Frequent mistakes: choosing the VPN type based on "which one is more secure". Both encrypt equally well; you choose based on who connects: a roaming person? Remote access. A whole site, permanently? Site-to-site.

Name the topology described in each case:

1. At Azahar's headquarters, every machine connects to a switch in the comms cabinet.
2. An old factory still runs an installation where every machine hangs off a single trunk cable; if the cable is cut, everything goes down.
3. Each node connects to the next and the last to the first, forming a closed circle around which data travels in one direction.
4. A company with 4 data centers has leased a direct link between every pair of centers.
5. In a 3-story building, the workstations on each floor go to that floor's switch, and the three floor switches connect to a central switch in the basement.

📌 Review if you struggled: lesson 01-03.

1. Star: everything goes through a central node. It is the standard topology of today's LANs.
2. Bus: the single shared medium is its signature — and its Achilles' heel.
3. Ring.
4. Full mesh: everyone with everyone.
5. Tree (or extended/hierarchical star): per-floor stars hanging from an upper tier.

A consultancy with 5 offices is studying linking them all to each other (full mesh).

1. How many links does it need?
2. It is about to open a sixth office: how many links will it need then? How many must be added?
3. Alternative: a star topology with the head office as the central node. How many links with 6 offices? What does it gain and what does it lose compared to the mesh?

📌 Review if you struggled: lesson 01-03.

1. With the full-mesh formula: $$\frac{n \cdot (n-1)}{2} = \frac{5 \cdot 4}{2} = 10 \text{ links}$$
2. $$\frac{6 \cdot 5}{2} = 15 \text{ links}$$ — 5 must be added (the new office connects to each of the 5 existing ones). Notice the pattern: each new office costs as many links as there were offices. The mesh grows quadratically: at 10 offices it would be 45 links.
3. As a star: 5 links (n − 1: each peripheral office to the central one). It wins on cost and simplicity (5 versus 15, and each new office costs 1 link, not n). It loses on two counts: the head office becomes a single point of failure (if it goes down, the peripheral offices are cut off from each other) and all traffic between peripheral offices detours through the center (more latency and more load on the head office).

Frequent mistakes: forgetting to divide by 2 and answering n·(n−1) = 20. Each link joins two offices; counting it from both ends doubles it.

Part A. In Meridiano's Valencia office (a star: the 20 workstations, the printer and the server 192.168.10.10 connected to the switch; the switch connected to the router 192.168.10.1, which provides Internet and the VPN with Bilbao), describe what stops working and what keeps working if each of the following fails, one at a time: (1) the switch, (2) the router, (3) the server, (4) the Wi-Fi access point.

Part B. Clínica Azahar asks you for a topology recommendation: inside each site, and between the 3 sites (headquarters in Alicante, branch clinics in Elche and Torrevieja; the branch clinics work almost exclusively against the Alicante server). Justify both choices and point out the design's main risk.

📌 Review if you struggled: lessons 01-02 and 01-03.

Part A.

The important takeaway: "the network is down" is almost never literally true. Knowing what went down from what stopped working is diagnosis in its purest form — the symptom pattern gives the culprit away.

Part B.

• Inside each site: star (central switch, everything connected to it). It is the standard choice for good reasons: simple cabling, adding a workstation touches nobody else, a damaged cable only takes down that one seat. The single point of failure (the switch) is accepted because replacing a switch is fast and cheap compared with the alternatives.
• Between sites: star (hub-and-spoke) with Alicante at the center: site-to-site VPNs Elche–Alicante and Torrevieja–Alicante. Rationale: the actual traffic is branch clinic ↔ central server, so the links follow the traffic pattern; a full mesh (3·2/2 = 3 links) would add the Elche–Torrevieja link, which would barely be used. With only 3 sites the mesh would be affordable, but it adds no value to the current usage pattern — it would be worth revisiting if the branch clinics started working with each other.
• Main risk: Alicante is a double single point of failure — as the center of the VPN star and as home of the medical records server. If Alicante goes down (power, Internet, the server itself), the branch clinics are left without records. Mitigations worth weighing: an off-site backup, a second Internet line at headquarters.

First gym session complete. You have classified devices and media, converted Mbps into MB/s and into seconds, told apart when the culprit is bandwidth and when it is latency, put names to networks of every size, chosen between client-server and P2P and between the two VPN families, and reasoned through topologies with their arithmetic and their points of failure — no longer just about Meridiano, but about a company you saw today for the first time. That is the sign that module 1 has genuinely sunk in. The next lesson climbs a rung: protocol exercises — encapsulation, frames, ARP, TCP, DNS and friends — where the answers are no longer reasoned out with logic alone, but by reading what the bytes and the command outputs are saying.