Conflict domains and broadcast domains of network interconnection devices (1)

Conflict domains and broadcast domains of network interconnection devices (1)

A conflict domain refers to the minimum range of conflicts. When a computer and a computer are connected through a device, a channel is established. If this channel allows only one data packet to pass in an instant, if there are two or more data packets that want to pass through the packet here, there will be a conflict. The size of the conflict domain can be used to measure the performance of the device. There is only one conflict domain in the multi-port hub, that is, all the datagram files on the port must be queued for passing. The switch obviously reduces the size of the conflict domain so that each port is a conflict domain, that is, the high-speed transmission of one or more ports will not affect the transmission of other ports, because all data packets pass through the queue in order, data only to the same port needs to be queued.

If the destination address of a data packet is the broadcast address of the network segment or the MAC address of the target computer is the FF-FF-FF-FF-FF-FF, the data packet will be received and responded by all the computers in the network segment, this is called broadcast. Generally, broadcast is used for ARP addressing and other purposes. However, uncontrollable broadcast domains also have a serious impact on network health, mainly due to bandwidth and network latency. The range of broadcast coverage is called the broadcast domain. The layer-2 switch forwards the broadcast, so it cannot divide the broadcast domain, and the router generally does not forward the broadcast, therefore, you can divide or define broadcast domains.

Network interconnection devices can divide networks into different conflicting and broadcast domains. However, different network interconnection devices may work at different layers of the OSI model. Therefore, the effect of dividing conflicting domains and broadcast domains is also different. For example, a repeater works on the physical layer, a bridge and a switch work on the data link layer, a router works on the network layer, and a gateway works on the top three layers of the OSI model. Network interconnection devices at each layer need to complete different tasks according to the characteristics of different layers.

Next we will discuss the working principles of common network interconnection devices and their respective characteristics when dividing conflicting domains and broadcast domains.

1. Traditional Ethernet operations

A typical example of conventional shared Ethernet is bus-type Ethernet. In this type of Ethernet, there is only one communication channel, using the media sharing (Media contention) access method (CSMA/CD media access method described in Chapter 1st ). Before sending data, each site must first check whether the network is idle and send data if it is idle. Otherwise, continue listening until the network is idle. If the two sites detect that the media is idle and one frame of data is sent at the same time, the data frame conflicts and the data frames of both sites are damaged. At this time, the two sites will use the "Binary Index backoff" method to wait for a random period of time before listening and sending.

In Figure 1, host A only wants to send A unicast packet to host B. However, due to the broadcast nature of the traditional shared Ethernet, all hosts connected to the bus will receive this unicast packet. At the same time, if any second party, including host B, also needs to send data to the bus, it will conflict, causing both parties to send data to fail. It is said that all hosts connected to the bus constitute a conflict domain.

When host A sends A broadcast data packet destined for all hosts, all hosts on the bus must receive the broadcast data packet and check the content of the broadcast data packet, further processing is required. All hosts connected to the bus constitute a broadcast domain.

Traditional Ethernet

2. Repeater)

Repeater appears as a product mainly for two reasons:

First, expand the network distance and regenerate the attenuation signal.

Second, Ethernet Interconnection between coarse and fine coaxial cables is realized.

Although the repeater can extend the signal transmission distance and achieve the interconnection of the two network segments. However, no available bandwidth is added. As shown in figure 2, CIDR Block 1 and CIDR Block 2 form a single conflict domain and broadcast domain after being connected by a repeater.

Network connected by Relay

3. HUB)

A hub is actually a multi-port repeater. A hub usually has eight, 16, or 24 interfaces.

A hub can also extend the communication distance between networks or connect networks with different physical structures, but it is mainly used as a collection point for a host site, connect hosts connected to interfaces on the hub so that they can communicate with each other.

As shown in 3, all hosts are connected to the hub of the central node to form a physical star connection. But in fact, in the hub, each interface is connected through the backplane bus and logically forms a shared bus. Therefore, the hub and the hosts connected to all its interfaces form a conflict domain and a broadcast domain.

Network connected by the Hub