Topic 4a – Networking Fundamentals

Networks and Protocols

1. The computers inside the school building form the LAN (Local Area Network) — a network confined to a single site. The connection from the school to the ISP and out to the Internet is the WAN (Wide Area Network) link. The device at the boundary is a router, which connects the school's internal network to the ISP's external network and directs traffic between them.
2. A protocol is a set of agreed-upon rules that governs how communication takes place between two parties. Networks need protocols because without shared rules, two devices would have no way to know how to format data, when to send it, how to handle errors, or how to indicate that a message has ended. A non-computing example: the protocol for a telephone call includes who speaks first ("Hello"), how to indicate you are done speaking (pausing), and how to close the call ("Goodbye"). Without these shared conventions, callers would constantly talk over each other.
3. CSMA/CD (Collision Detection) works on wired networks because a device can detect a collision while transmitting by simultaneously listening on the wire. On wireless networks, a device cannot hear its own transmissions while also listening, so collision detection is impossible. The hidden terminal problem occurs when two devices can both reach the same access point but cannot hear each other — they cannot detect that the other is transmitting. CSMA/CA (Collision Avoidance) addresses this by having devices request permission to transmit before sending, so the access point can coordinate who goes when.
4. An open network uses publicly documented standards that any vendor can implement (Ethernet, Wi-Fi, TCP/IP). A proprietary network uses standards controlled by a single company, requiring its hardware and software. The open Internet succeeded because anyone could connect any device, build any application, and operate any service without permission from a central authority. Proprietary alternatives (like early AOL or CompuServe) required users to stay within their ecosystem, limiting what was possible.

Network Hardware

5. • Wired lab to wireless student network (different protocol): router (or bridge/gateway) — connecting two networks with different protocols requires a device that can translate between them.
   • Extending Ethernet to a distant wing, no filtering needed: repeater — a repeater simply amplifies and retransmits the signal with no intelligence or filtering.
   • Wireless access for laptops/phones/tablets: access point — an access point bridges wireless devices onto the wired network.
   • Three Ethernet segments with minimized cross-segment traffic: switch (or bridges) — a switch forwards traffic only to the segment where the destination device is located, reducing unnecessary traffic on other segments.
6. Devices that extend a single network (repeaters, bridges, switches) operate within one network — they move data between devices that share the same addressing scheme and protocol. A router connects two or more different networks, each with its own addressing. A router must preserve and work with IP addresses (network-layer addresses) to determine how to route packets between networks; a switch only needs to know hardware (MAC) addresses within a single network.
7. "An internet" (lowercase) is any collection of interconnected networks — your school's network connecting multiple buildings, for example, is technically an internet. "The Internet" (uppercase) refers specifically to the global public network using the TCP/IP protocol suite that connects billions of devices worldwide. A school network that is not accessible from the public Internet is an internet but not part of the Internet. A school network that connects to the public Internet is both an internet (internally) and a node on the Internet (externally).
8. C. Since they can still reach the outside Internet, the router is working. The problem is likely with the local hardware (the switch) that connects the devices within that specific room.

Communication Models and Distributed Systems

9. Your browser is the client; the machine hosting the website is the server. The client sends an HTTP request for a specific page (identified by URL). The server locates the requested resource and sends back the HTML, images, and other files that make up the page.
10. Two key differences: (1) Roles: in client/server, devices have dedicated roles — servers always serve, clients always request. In peer-to-peer, every device can act as both. Example: a streaming service (client/server) vs. BitTorrent (peer-to-peer, where each downloader also uploads to others). (2) Centralization: client/server centralizes resources and control on the server; peer-to-peer distributes resources across all participants. Example: a school file server (client/server) vs. students sharing files directly with each other (peer-to-peer).
11. The student is confusing file sharing behavior with network architecture. A peer-to-peer network is one where devices communicate directly with each other without a central server managing the exchange. Most school networks are client/server architectures — students' devices connect to central servers (file servers, print servers, authentication servers) even when "sharing files." True peer-to-peer would mean devices connect and transfer directly without any central coordination.
12. Cloud computing analogy: think of it like electricity from the power grid. You do not generate your own electricity; you subscribe to a utility and use as much as you need, paying for what you consume. Cloud computing works the same way — you rent computing resources (storage, processing, applications) from a provider's data center rather than running your own hardware. It relates to the mainframe/terminal model in that computation is centralized in a remote facility and users access it over a network — the difference is that cloud resources are virtualized, globally distributed, and billed per use rather than time-shared on a single machine.
13. Cluster computing connects multiple machines in the same location via a high-speed local network to work together as a single system — all machines are under unified management, tightly coordinated. Grid computing connects machines in different locations (often across organizations) over a standard network to collaborate on large tasks, with looser coordination. A website that never goes down is more likely backed by cluster computing — clusters are designed for high availability with redundancy built in, so if one node fails others take over instantly.