1.1 OSI and TCP / IP protocol framework

OSI is a network of community law, the main purpose of implementation of various manufacturers are compatible with equipment operation, TCP / IP protocol are the mainstream of the Internet. Figure 1 is the OSI and TCP / IP protocol model comparison.

1.2 OSI functions and features of each

1, physical layer: its role is to transmit BIT signal, representative of the typical equipment such as HUB (hub).

2, data link layer: including LLC and MAC sublayer, LLC is responsible for communications with the network layer, network layer protocol consultations. MAC is responsible for the physical layer control. Typical equipment in this layer are SWITCH (switch).

3, network layer: This layer is responsible for routing table set up and maintenance, data packets transmitted. Typical equipment in this layer are ROUTER (router).

4, Transport Layer: This layer application data subparagraph of paragraph ended to set up virtual connection, to provide a reliable or unreliable delivery.

5, Session Layer: This layer is responsible for the conversation between the two application management and maintenance.

6, express layer: This layer solution data indicated that the conversion problem is the coordination of human-machine communication between persons, such as binary and ASCII code conversion.

7, Application Layer: This layer is the man-machine communications interface. Typical applications such as FTP, HTTP and so on.

1.3 OSI package solution packages as well as PDU

1.3.1 Packaging

Packaging refers to the so-called sender happened in the top-down process

In the application data for each layer to add on a particular head / tail of information (PDU, Protocol Data Unit, Protocol Data Unit)

Application (application) → segment (data segment) → packet (packet) → frame (data frame) → bit (bit, binary-bit)

1.3.2 Packaging Solutions

The so-called solution package is the recipient of the bottom-up processes happen

Layers of information to remove the head and tail

1.4 IP packet structure

One of the important fields include:

TTL (Time To Live, survival time): After the router every time, this value minus one. If the value is 0 router will not forward this packet.

Protocol (protocol): network layer and transport layer of communication between the interface and used to identify the transport layer of the transfer protocol.

Identification (serial number): for each packet sent to a number.

Flag (offset mark), Frag.Offset (offset): used to receive data packets will carry out the sub-tablets.

1.5 IP packet transmission

IP packet transmission process includes:

1. Host sends packet to default gateway (the host data packet sent to default gateway)

2. Packet placed in frame (data packet is encapsulated into the frame)

3. Router receives frame (router received frame)

4. Router finds destination network in route table (router in the routing table found in the target network)

5. Router chooses next hop toward destination (a router to choose a target closer to the next hop)

6. MAC address of next hop determined (next hop’s MAC address was identified)

7. Packet placed in frame (data packet is encapsulated into the frame)

8. Repeats steps 2 through 7 as necessary (If necessary, repeat steps 2 ~ 7)

9. Router receives frame (router received frame)

10. Router finds network directly connected (Direct Connect network router found)

11. MAC address of end host determined (the final host MAC address was identified)

12. Packet placed in frame to final destination (frame in the packet is sent to the final host)

IP packet transmission the whole process.

In the end-to-end packet transmission process, the logic does not always happen to change addresses, and MAC address with the specific link vary.

Entry in a router interface to receive data frames, the first detect whether the destination is its own. If so, then to the upper handle, otherwise they will cache the contents of data packets, and then search according to destination address routing table to find the relevant table entry, the NEXT HOP and the interface MAC address, use of these two addresses as a new purpose and source MAC address prior package cache packets, and then forward, this process is known as frame rewrite (REWRITE).

1.6 IP-related agreements

1.6.1 ARP

ARP (Address Resolution Protocol, ARP) has the following characteristics:

ARP by ARP Request (Broadcast) and the ARP Reply (unicast) component.

Only when the sender that the target host with its own in the same logical network (same network segment), ARP Request will be issued.

Belong to the local MAP (different FR MAP) IP address and MAC address are the same equipment (interface).

1.6.2 ICMP

IP protocol is an unreliable protocol, error control can not be carried out. However, IP protocol can make use of other agreements to achieve this function, such as ICMP. As shown in Figure 5.

ICMP protocol allows the host or router report errors and provide information on the report of unusual circumstances.

Generally speaking, ICMP packet network layer provides error diagnosis, congestion control, path control and inquiry services to the four major functions. Such as, when a packet can not reach our destination site or TTL timeout, the router will discard the packet to the source site to return to a site can not be the purpose of the ICMP packet arrival.

1.7 Transport Layer functions

Transport layer features include:

Session Multiplexing (Multiplexer): a number of application session reuse in the same end-to-end connection (with a source IP and destination IP of the) above, through the port number identified.

Segmentation (sub-): chunks of application data will be separated into more suitable for transmission of the paragraph.

Flow Control (Flow Control): Software flow control to prevent network congestion when the packet loss to reduce the possibility of network congestion.

Buffering (buffer): acceptance into the buffer cache and sent to temporary data.

1.8 TCP and UDP contrast

Are TCP and UDP packet header format.

Their difference is:

TCP

Connection-oriented (synchronous, confirmed that the window)

Provide reliable transmission service

High reliability

UDP

Connectionless

Provide best (Best-Effort) service

High efficiency

1.9 TCP sequence number and acknowledgment number

TCP sequence number and acknowledgment number used to receive each other’s data packets to confirm. Serial number and confirmation number is the number of bytes as a unit, and confirmation number equal sender serial number plus 1. Shown in Figure 8.

1.10 TCP sliding window

Sliding window used to achieve flow control. It is used to prevent the sending end to send excessive data will be slow to receive the receiving end submerged, resulting in the receiver buffer overflow.

Window packet size is the number of bytes as a unit rather than the number of packet units, the TCP sliding window flow control belong to one of the ways.