On this page, you will learn how the layout of the Internet is redundant (more than one path from here to there) in order to ensure reliability.


VOCABULARY

  • A path is a sequence of directly connected computing devices that connect a sender to a receiver.
  • Routing is the process of finding a path from sender to receiver.
  • Scalability is the ability of the Internet to keep working as it grows.
  • Redundancy is the inclusion of backup elements in case one part fails.
  • Fault tolerance is the ability of a system to work around problems.


Given the enormous number of devices on the Internet and the reality that pieces of a complex system fail at unexpected times (and often in groups of neighbors such as a whole city), the Internet had to be designed to be reliable. This is achieved by building many redundant connections into the physical systems of the Internet. That way, if part of the Internet fails, data can be re-routed via a different path. And such changes to the path can happen in transit because routing on the Internet is dynamic; it is not specified in advance. Creating such redundancy can require additional resources (such as additional computers and cables) but it also increases the Internet's fault tolerance (ability to work around problems) and helps the Internet scale (expand) to more devices and people.

1. Describe what's going on in this animation.


In the animation, a lightning bolt strikes one of the nodes. There are many reasons a node can fail, such as power failure, chip burnout, etc. The other thing that can happen is that the nodes can all be fine, but the green lines linking the nodes, which represent connections, could fail (e.g., a cable could be cut or disconnected, either accidentally or on purpose).

 

2. Self-Check Question

In this model of a network, what is the minimum number of nodes (connection points) that can stop working before the sender and the receiver can't communicate? (Other than the sender or the receiver themselves, of course.)

 

 

Answer: 2, If the node with six connections goes down and also either of the two to its left, the sender and receiver can't communicate.
 
In the same model network, what is the maximum number of nodes that can fail and still let Sender and Receiver communicate?

Source: https://bjc.edc.org/bjc-r/cur/programming/4-internet/1-reliable-communication/2-network-redundancy.html
This work is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 4.0 License.

Last modified: Monday, April 1, 2024, 11:48 AM