Routing by rumor, just why?

[LetMePass]

Hi there,

I am struggling with the concept of routing by rumor. 

Ft. Worth is going to send information about its own networks to Dallas. Dallas would take that information in conjunction with its own and sends it on to San Jose as if it originated the data. So when San Jose gets the information from Dallas, it does not know that it's not coming directly from Dallas. It may assume that there routes that were learned via Ft. Worth are really Dallas routes. It has not way of differentiating between the two.

Distance-vector routing protocols "route by rumor" - this means that they do not 100% KNOW that a specific neighboring router has the absolute best path for a packet, but the neighboring router tells them that they do, so they believe it and send the packet to the neighboring router. This also means that distance-vector routing protocols do not have complete knowledge of the entire network; they might know a route to a far-off network, but they don't know if that route is the best one to take.

Aside CPU usage, what is the purpose to see a small portion of the network and not to know what route is the best one to take?

What purpose does that serve?

About link-state I got that a Router sees all the routes to its destination in a road map style and then find out the one with the lowest cost but here, I cannot see the logic.

I welcome any comment!

Cheers,

Chris

[Martin L · Accepted solution]

Just to be clear, this "route by rumor" is a flaw m  a drawback of Distance-vector protocols, especially when comparing to link-state ones. Those Distance-vector routing protocols (RIP, IGRP, EIGRP) were developed before link-state protocols like OSPF and were used for many years. Yes, comparing to link-state protocols, Distance-vector protocols have disadvantages described as "route by rumor".  Even with improvements like split-horizon, poison-reverse, and loop-protection in form of EIGRP feasibility condition did not push those protocols over the link-state ones.  RIP 1 and RIPv2 are obsolete and EIGRP -although my favorite one - will be soon forgotten. 

Regards, ML
**Please Rate All Helpful Responses **

[M02@rt37 · Accepted solution]

Hello @LetMePass,

"Route by rumor" is indeed a limitation or drawback of distance-vector routing protocols, including RIP, IGRP, and EIGRP. These protocols were developed earlier in the history of networking and were widely used for many years due to their simplicity and ease of implementation. However, they come with inherent shortcomings when compared to more modern link-state routing protocols like OSPF.

Distance-vector protocols propagate routing information based on what neighboring routers "tell" them, often leading to suboptimal routes or loops in the network due to potentially outdated or incomplete information.

Over time, networking requirements evolved, necessitating more accurate and efficient routing protocols. Link-state protocols, such as OSPF, were developed to address the limitations of distance-vector protocols and provide a better understanding of the network topology.

Improvements like split-horizon, poison-reverse, and EIGRP's feasibility condition were introduced to mitigate the issues with distance-vector protocols, making them more reliable and efficient. However, these enhancements couldn't completely overcome the fundamental "route by rumor" limitation.

Link-state protocols like OSPF have gained popularity due to their ability to provide a complete and accurate view of the network, leading to better route selection and faster convergence. They have become the preferred choice for modern networking environments.

While distance-vector protocols have their historical significance and are still in use in certain environments, they are gradually being phased out in favor of more advanced and efficient routing protocols like OSPF and BGP.