An Ethernet switch uses a memory buffering technique to store frames before forwarding to the destination. The switch uses Buffering when the destination port is busy due to congestion. As a result, frames must be buffered until transmitted. So without an effective memory buffering scheme, frames are more likely to be dropped anytime traffic oversubscription or congestion occurs.
During congestion at the port, the switch stores the frame until it transmitted. The memory buffer is the area where the switch store the data. There are two methods of buffering:-
Port-based Memory Buffering
In this type, all frames have stored common memory buffer and in queues; that linked to specific incoming ports. Switches utilizing port buffered memory in this type of buffering. In port buffering switch give each Ethernet port with a certain amount of high-speed memory to buffer frames until transmitted. A disadvantage of port buffered memory is the dropping of frames when a port runs out of buffers. It is also possible for a single frame to delay the transmission of all the frames in memory because of a busy destination port. This delay occurs even if the other frames could be transmitted to open destination ports.
Some of the earliest Cisco switches use a shared memory design for port buffering. Shared buffering deposits all frames into a common memory buffer that all the ports on the switch share. The amount of buffer memory required by a port dynamically allocated. The frames in the buffer have dynamically connected to the destination port. This allows the packet to receive on one port and then transmitted on another port, without moving it to a different queue.