Computer Networks · free preview
Requirements and Connectivity
Connectivity must scale to a global Internet, so nodes connect through links into switched networks and networks nest recursively into internetworks, every node carries an address and routing forwards messages toward destinations, sharing links cost-effectively is done by statistical multiplexing rather than fixed time or frequency division, applications need logical channels with properties like reliability and ordering, and three classes of failure plus the burden of manageability complete the requirements a network design must meet.
1. Connectivity at scale
Connectivity at scale
2. Sharing links cost-effectively
Sharing links cost-effectively
3. Channels and common services
Channels and common services
4. Surviving failure and staying manageable
Surviving failure and staying manageable
5. The cost of a full mesh
Why is connecting every host directly to every other host not a viable plan for a large network?
6. The three failure classes
Which list correctly names the three classes of failure a network must survive, from smallest to largest granularity?
7. Why statistical multiplexing wins
Why does statistical multiplexing outperform time-division and frequency-division multiplexing for computer traffic?
8. The two channel types
A videoconference and a file download place opposite demands on a channel. Which pairing of channel to requirement is correct?
9. Lost versus late
What makes a lost packet fundamentally hard to handle, compared with a corrupted one?
10. Stability against velocity
What is the central tension in managing modern networks that the rise of the cloud has sharpened?
11. The sharing technique
What is the name of the multiplexing technique in which flows transmit on demand rather than in fixed slots or bands, and which earns its efficiency from the burstiness of computer traffic? Two words.
12. The forwarding process
What is the name of the systematic process by which switches and routers use a destination address to forward a message toward the node that should receive it?
13. Three failure granularities
At the smallest granularity, interference flips individual symbols in transit. What are these smallest failures called?
14. The semantic gap
The distance between what an application expects from a channel and what the underlying network technology can actually provide is a named design problem. What is it called?
Practice this interactively
Free account · instant grading · spaced review that schedules itself.
Start this course for free