Monday, November 3, 2014

2.1.3 TCP/IP Operation

Figure 2.7 shows TCP/IP operation of accessing a website. After a user click on the website link, the application uses DNS client software to request DNS Server to send an IP address of the destination host. The DNS client software uses UDP and IP to build a DNS name query message to request the IP address of the host that the user wants to connect to. Link layer builds an ethernet frame and pass it to the network operating system to be sent to the DNS Server. DNS Server sends back the DNS name query response message and a target host IP address. It comes in link layer, goes through IP layer where the address information is stripped off and goes back through UDP where the packet integrity is checked. Then it goes up to the DNS client software and passes the IP address to the application.

Figure 2.7: TCP/IP operation of accessing a website
The application then instructs TCP to create a connection. After a TCP connection is established, the application requests a file from the target host. The target host uses TCP to segment the file user wanted into packet and send it to the client. Each packet comes through link layer, through IP and passes it up to TCP which verify the segment integrity and hold until all files segments had arrived. It also acknowledges receiving of the packet to the sender. This cycle repeated until the whole files had been reassembled.

If the packet arrive does not pass the integrity check, TCP discarded it and not send acknowledgement to the sender. If the last packet was not acknowledged within a lot of time, the sender retransmit it. When the last packet arrives, the files are reassembled and acknowledged to the sender. The client then closed TCP connection. The completed files passed to the application where users can access it.

No comments:

Post a Comment