Message switching vs packet switching with respect to congestion traffic

Question

Packet switching uses store and forward technique to forward packet. And message switching is also uses store and forward technique to forward message. But in packet switching data transfer faces congestion despite large packets are divided into small packets. Congestion Control in packet-switching :"Maintains the number of packets within the network below the level at which queuing delays become excessive. When line for which packets are queuing becomes more than 80% utilized, the queue length grows at an alarming rate." I have read from this site.

But in message switching large packets treated as single unit. I have read There is a reduction in network traffic congestion because in this technique we are using store and forward property and also any switching node can store the messages till the availability of the network . I have read from this site.

My question is that what is the actual reason why the congestion traffic is less in message switching compare to packet switching?

Answer 1

This seems to assume that queues can have limited number of packets (true enough), but unlimited packet size (not true). The idea that a large message packet size uses fewer places in the queue is true, but it ignores reality that an interface can only serialize the packet bits at a certain speed (bandwidth), so that larger packets take proportionally longer to transmit, stalling the queue for a longer period of time.

The bottleneck is the bandwidth of the interface. If an interface can transmit at a limited speed (true), the number of bits in the queue will be the same, regardless of the packet size. What is true is that the larger message packets will be fewer in the queue, but will still take about the same amount of time as smaller packets to transmit the same amount of data because the interface can only serialize the bits at its fixed speed.

Transmitting 1 Gb through a 100 Mbps interface takes the same amount of time regardless of the packet size. The only thing larger packets gain is eliminating some packet header overhead, but that can be ignored unless the packets are so small that something like the IPv4 20-byte packet header is a substantial percentage of the packet size.

In any case, the real world uses packet switching. Circuit switching, such as the traditional telephone circuit-switching network, is going away in favor of packet switching because packet switching is more flexible, allowing the circuit to be shared more easily and allowing different services, e.g. voice, data, video, etc., to use the same circuit at (relatively) the same time.

Answer 2

First of all, the term network congestion is the most fuzzy term in networking. Everyone know what happened on an event known as "congestion collapse". Everyone knows that host are supposed to do congestion control, to prevent congestion collapse from happening. Noone really knows/or agrees upon what the word congestion means.

Now, let's consider message switching:

What "store-and-forward" means https://www.studytonight.com/computer-networks/messageswitched-networks in details.

1. Switch receives entire message (i think it can come in multiple chunks).
2. Switch stores entire message.
3. Switch only forwards the message to the next switch when the next switch is up, and is ready to receive the message (including has enough link bandwidth and storage capacity).

So basically, if a message cannot be forwarded because the next router is not ready it is stored. I think this will eventually mean that when the first switch of the sender is full, sender can't send messages.

What is the difference to packet switching. Packet switching does not have "store" property in a sense that it does not have 3. A packet arrives as a router. A router decides on the outgoing interface. This interface has a buffer for outgoing packets. Packet is either put in this buffer or dropped.

Here a router does not store message for the case that next router cannot process it. If at any point along the pass the router cannot process the packet, it just drops it.

So, if you have message switching you will have fewer dropped packets. I don't think it implies anything, other than this. For example, it does not mean that the sender will get fewer/more data transmitted to the receiver.

What is the difference with circuit switching:

in a circuit switching, before sending a message the sender reserves bandwidth on each switch up to a destination. this bandwidth can be used only for transmitting this message (well, data of the given connection), regardless of whether it is used.

If there isn't enough bandwidth to reserve,the connection is not established.

So, takeaway

• packet switched networks can drop packets because at the time of packet crossing the switch, the switch was overloaded with traffic and could not process the packet.
• in circuit switched networks this does not happen. this comes at a cost of under utilizing resources, i.e., resources are reserved for particular connection but are not used.
• in message switched networks this presumably does not happen. message switched networks are more efficient, because the resources are reserved only between 2 neighboring switches not on all path.

why do we use packet switching?

well, experience!!! (as in ~50years of Internet) has shown that with the number of different types of applications that are used in Internet (Web, Video, Calls, Games), packet switched network works extremely well supporting them all. For some applications circuit switched or message switched packets could be better. But it is better to have one Internet, than separate network for each .

And what about congestion?

packet switched networks can drop packets because at the time of packet crossing the switch, the switch was overloaded with traffic and could not process the packet.

this has something to do with congestion. so i would say that congestion is a term that only applies to packet switched networks.

as I said, i assume that message switched networks drop less (if any) packets. But this is the only thing that you can state with certainty. This does not imply less or more congestion, especially without defining what congestion is.

Maintains the number of packets within the network below the level at which queuing delays become excessive. When line for which packets are queuing becomes more than 80% utilized, the queue length grows at an alarming rate.

this sentence does not make any sense. i would even say that it is just false. in general do not take anything written in that page seriously.