Bridging loops ..

As we have learned from the basic concept of bridges, what is a transparent bridge . From this understanding we will go one step ahead and find out what are bridging loops.

In any topology we would love to have redundant paths to reach  any destination so that during failure of one path , it seamlessly switches its direction and  chooses another path . 

In a bridged LAN environment, where we have multiple paths bridging loops will exists and drain the network bandwidth by continuously looping the packet around . Let us first see how it actually happens.

  

In the above picture, we see there are two bridges placed for connecting station 1 from network segment 1 to station 2 in network segment 2 . 

Let us imagine both the bridges forwarding table entry is empty and we get a packet from station 1 destined towards station 2.

•            Since both the bridges are in promiscuous mode , both will receive packet in segment 1 , destined to Station 2 .  Now since they are transparent bridges , they will learn that station 1 can be reached  via port connected to segment 1 .         
•             Both the bridges will now plan to send the packet out on all other interfaces except on the received port, as they do not have any entry for reaching station 2 .
•             Since only one bridge can transmit packet out any point of time ( because of CSMA/CD or any other layer 1 protocol), assuming bridge 1 will get time slice for sending packet . This packet will be received by station 2 as well as bridge 2. 
• Now bridge 2 thinks that station 1 has moved its position from segment 1 to segment 2 , it will alter its forwarding table accordingly .
• When bridge 2 sends packet out once its get time slice , Along with  station 2 , bridge 1  will also receive the packet. Now imagine station 2 behavior , it has received same packet twice from two different points - whose behavior is unpredictable here !!
•  Since bridge 1 also received the packet  from bridge 2 , it alters its own forwarding table thinking that station 1 has moved it position from segment 1 to segment 2 .
• Every time a packet is generated , it loops around the paths continuously and bridge forwarding entries will flap every time . Imagine a case when you send a broadcast packet from station 1 / station 2 . 
• This is mere example of what could happen if we have multiple paths , in real life scenario situation is much worse , each of stations will have multiple paths in various segments . 

To avoid loops , we need to run a protocol on the bridges which can determine a possible loop and shuts down one of the ports of the bridges ( logically !! )

Lets look in to such protocol in the next article , which is "spanning tree protocol "