Q1. Concept of polling and its effect on network congestion and saturationThe switched Ethernet was developed to cub the problem of increased load in a network system that would often result to saturation in a classic Ethernet system. Polling is done on the switch to establish the required bandwidth for a real time system established within the protocol 802.16. In so doing, the capacity of the network is efficiently utilized. Whereas collisions are avoided in the use of a switch relative to the hub, without necessarily using the CDMA/CD algorithm, it is still significant to note that a single port within a switch may receive two frames at the same instance of time.
Thus buffering is still a requisite tool of consideration. With many frames send to the buffer and with relative speeds of transmission, dropping the frames is bound to happen. Thus polling could still solve the potential event of discarding the frames when the limited buffer space is filled. It could be applied by allowing network subscribers being polled at set intervals to determine the required bandwidth and hence the switch device could sort the frames based on the bandwidth before porting them to the destination address.
A hub also can be ported and polled to one of the destination for concentration of the frames and hence free the port buffers during a busy state. The advantage of polling the destination thus stands in that if a large bandwidth is required for transmission, the extra frames could be send to the concentration hub and hence queued and would avoid dropping. The main disadvantage would be a slight reduction in speed and also complex design for the switch to include the operation in its algorithm.
(A Tanenbaum, 1944, 290) Token bus data transmission on devicesThe token bus technology, though apparently has become deprecated employed the use of a token passing between devices that are connected to a network. The token bus is a logical topology that is best employed in a ring physical topology. Devices in the network would have a token circulating through the network with the destination address of the recipient device, and with the frames that need to be delivered.
Each device in the network would receive the token, check the address and decide whether it matched or not. In case the token had the device address as its destination then it would read it and receive the frames, otherwise it would pass the token to the next device. The Token once it has delivered the attached frame would then be released and free to be picked by another device that has a send request. Thus in the token bus topology, the bus line is always active and there are no idle states upon the bus line.
Thus in the token bus protocol, the sending device has control of the token which carries on frame a time, before being available again for any other device to use it. Negative impacts of collision and broadcast on the throughput of an Ethernet. Using bus topology: In a bus topology, the backbone line of data transmission is only available when no device is transmitting. This ideally limits the use of the main resource to only two devices that have taken hold of the resource.
Thus in the incident where the network address of recipient is multiple, delivery of the message won’t be instant since only two devices can communicate at a time. The overall speed of the network is greatly reduced and any other device cannot send data frames until the sender releases the bus after transmitting to all its recipient. Also, since the bus topology has only a single bus to be shared by all devices that have been connected to the network, a wrong address would mean that the bus won’t be released thereby blocking other devices from communication.
As a result the network would be suspended indefinitely and thus lowering the overall throughput of the network. The idle time also when no device has data to send reduces the efficiency of the network rendering it cost ineffective. Repeaters: