A storm of dust has knocked out the link between Hub North and mission control. The plant team is waiting for tonight's readings.
"Our shortest path is gone," Comet says. "Do we just wait until it is fixed?"
Wren spreads the relay map across the table. "What do you notice? Hub North still has another link."
Nova hovers above Ridge Station. "A network with spare paths can bend without breaking," she says.
Comet grabs a stack of index cards. "Then let's play it out. Every card is a packet, and we are the routers."
Wren writes the first address. "To: Mission Control."
You will run the relay network on paper, send a message as packets, break a link, and reroute.
Then you will decide how fault tolerant the relay really is.
| Round | Links removed | A path still exists? | Path you used |
|---|---|---|---|
| 1 | none | ||
| 2 | Hub North to Mission Control | ||
| 3 | Hub North to Mission Control, and Ridge Station to Mission Control |
Redundancy means including extra parts that can take over if other parts fail.
One way to build network redundancy is to have more than one path between any two connected devices.
If a device or link on the Internet fails, later data are sent by a different route, if one is possible.
A system that keeps working when parts fail is called fault tolerant. Redundancy often needs extra resources, but it gives that fault tolerance.
| What the lab shows | True or false? |
|---|---|
| Removing one link did not stop the message in round 2. | ? |
| The relay is fault tolerant for any two broken links. | ? |
| Extra links are an example of redundancy. | ? |
| Redundancy usually needs no extra resources at all. | ? |
Great relay work. Tomorrow you will look inside the layers of a device.