I got basic question. Trying to understand the mechanism how network layer identify when it receive packet larger than its interface MTU configured and has to fragment it.
For example. If the data to be transmitted by application is 2000byte where as the NIC on the server is having MTU 1500. Network layer has to do fragmentation now but how it knows it has to do it. Does the Transport layer give the full 2000byte data to network layer and than network layer compares it with the interface MTU configured and decide accordingly
If the data to be transmitted by application is 2000byte
There are transport-layer (L4) protocols (most prominently TCP) that do all the chunking for an application. TCP is MTU-aware and segments its data automatically into the optimal chunks.
Other L4 protocols don't do any chunking of their own (e.g. UDP). Any application data is simply encapsulated into a datagram and passed to the network layer. If necessary, the network layer then fragments the encapsulating IP packet across several data-link-layer frames.