Network design - Master Tran Van Long, Master Tran Dinh Tung Compiled

wide in the future.

If classified by ability, there are 2 types:

- Passive hub: is the type that does not contain electronic components and does not process data signals, its sole function is to combine signals from several network cable segments.

- Active hub: is the type that has electronic components that can amplify and process electronic signals transmitted between network devices. The signal processing process is called signal regeneration, it makes the signal better, less susceptible to errors, so the distance between devices can increase. However, those advantages also make the price of Active Hub much higher than Passive Hub. Token ring networks tend to use Active Hub.

Basically, in Ethernet network, Hub acts as a Repeater with multiple ports.

Maybe you are interested!

3) Bridge

A bridge is a processing device used to connect two similar or different networks, it can be used with networks that have different protocols. A bridge operates at the data link layer, unlike a repeater that has to retransmit everything it receives, a bridge reads the data link layer frames in the OSI model and processes them before deciding whether to forward them or not.

When receiving frames, the Bridge selects and forwards only those frames that it deems necessary. This makes the Bridge useful when connecting several networks together and allows it to operate flexibly.

Figure 1.14. Bridge operation

To do this, in the Bridge at each end of the connection there is a table of addresses of stations connected to that side. When the bridge is operating, it examines each frame it receives by reading the addresses of the sender and receiver. Based on the address table on the side receiving the frame, it decides whether to send the frame or not and whether it needs to be added to the address table.

When reading the sender address, Bridge checks the network address table.

receives a frame with that address or not, if not, the Bridge automatically adds it to the address table (that mechanism is called bridge self-learning).

When reading the address of the recipient, Bridge checks to see if the address is in the address table of the network part receiving the frame. If so, Bridge will assume that it is an internal frame belonging to the network part where the frame arrived, so it will not forward the frame. On the contrary, Bridge will forward it to the other side. Here, a sending station does not need to forward information across the entire network, but only needs to forward it to the network part where the receiving station is.

Figure 1.15. Bridge operation in the OSI model

To evaluate a Bridge based on two concepts: Filtering and forwarding. The process of processing each frame is called filtering in which the filtering speed directly represents the Bridge's performance. The forwarding speed is expressed in frames/second.

Currently, there are two types of Bridges in use: Transport Bridges and Translation Bridges. Transport Bridges are used to connect two local networks that use the same communication protocol of the data link layer, but each network can use a different type of wire. Transport Bridges are not capable of changing the structure of the frames they receive, but are only interested in reviewing and transmitting the frames. Translation Bridges are used to connect two local networks with different protocols, and are capable of converting a frame from one network to a frame from the other network before transferring it. For example: A translation Bridge connects an Ethernet network and a Token ring network. The Bridge then acts as a Token ring node on the Token ring network and an Ethernet node on the Ethernet network. The Bridge can transfer a frame according to the standard being used on the Ethernet network to the standard being used on the Token ring network.

Figure 1.16. Bridge compilation We often use Bridge in the following cases:

- Extend the current network when the maximum distance has been reached. Because the Bridge, after processing the frame, retransmits the frame on the remaining network, the signal is better than the relay.

- Reduce network congestion when there are many stations on the same network segment. Then, we divide the network into many Bridge segments, the frames within each network segment will not be allowed to go through other network segments.

- To connect networks with different protocols.

Some bridges also have the ability to select the transport object. It can only transport frames for specific addresses. For example: In the model below, the bridge allows packets from machine A, B to go through Bridge 1, packets from machine C, D to go through Bridge 2.

Figure 1.17. Network connection using 2 Bridges

4) Switch

A switch is an evolution of a bridge, but has multiple ports and uses fast integrated circuits to reduce the latency of data frame switching.

Switch maintains a table of MAC addresses for each port and implements the Spanning-Tree protocol. Switch operates at the Data link layer and is transparent to upper layer protocols.

5) Router

A router is a device that operates at the network layer and can find the best path for packets to travel through multiple connections from the sending station on the first network to the receiving station on the last network. Routers can be used to connect multiple networks together and allow packets to travel through multiple paths to reach their destination.

Figure 1.18. Router operation in the OSI model

Unlike Bridge, Router has a separate address and it only receives and processes packets sent to it. When a station wants to send a packet through Router, it must send the packet to Router's address (the packet must contain other information about the destination) and when the packet reaches Router, Router will process and send it again.

When processing a packet, the Router must find the path of the packet through the network. To do this, the Router must find the best path in the network based on the information it has about the network. Normally, each Router has a routing table. Based on data about nearby Routers and networks in the internetwork, the Router calculates the optimal routing table based on a predetermined algorithm.

Table 1.2. Router routing table

Routers are divided into two types: protocol dependent routers and protocol independent routers based on the method of processing packets when passing through the router.

Protocol-dependent routers: Only perform path finding and packet transmission from one network to another without converting the packet packaging method, so both networks must use the same communication protocol.

Routers are protocol independent: They can connect networks using different communication protocols and can convert packets of one protocol to packets of another. Routers also accept different packet sizes (Routers can split a large packet into many small packets before transmitting it on the network).

To prevent data loss, the Router also knows which path is available and stops transmission when the path is blocked.

Reasons to use Router:

- Routers have better filtering software than Bridges because packets that want to go through the Router need to be sent directly to it, reducing the number of packets going through it. Routers are often used when connecting networks through expensive leased lines, so they do not transmit redundant messages onto the transmission line.

- Routers can be used in a network with multiple regions, each region has its own separate protocol.

- Router can determine the safest and best path in the network, so information security is more guaranteed.

- In a complex network, when packets circulating on paths can cause network congestion, routers can be installed with methods to avoid congestion.

Some main operating protocols of Router: RIP (Routing Information Protocol), OSPF (Open Shortest Path First), EIGRP (Enhanced Interior Gateway Routing Protocol).

6) Routing switch (Switch layer 3)

L3 Switch can run routing protocol at network layer of OSI model. L3 Switch can have WAN ports to connect LANs at long distance. It is essentially added feature of Router.

1.2. Ethernet technology

1.2.1. Introduction

Today, Ethernet has become a widely used local area network technology. After 30 years of its birth, Ethernet technology continues to develop new capabilities to meet new needs and become a popular and convenient network technology.

On May 22, 1973, Robert Metcalfe of the Palo Alto Research Center of Xerox – PARC, California, proposed the idea of a computer network connection system that allows computers to transmit data to each other and to laser printers. At that time, large computing systems were designed based on expensive mainframe computers. The big difference that Ethernet brought was that computers could exchange information directly with each other without going through a central computer. This new model changed the world of communication technology.

The first 10Mbps Ethernet standard was published in 1980 by a joint development of three companies: DEC, Intel, and Xerox. This standard was called DIX Ethernet (named after the first three letters of the company names).

The IEEE 802.3 committee used DIX Ethernet as the basis for development. In 1985, the first 802.3 standard was published under the name IEEE 802.3 Carrier Sense Multiple Access with Collision Detection (CSMA/CD). Although it did not use the name Ethernet, most people understood it as a standard for Ethernet technology. Today, the IEEE 802.3 standard is the official standard for Ethernet.

IEEE has developed Ethernet standards on many different transmission technologies so there are many types of Ethernet networks.

1.2.2. General characteristics of Ethernet

1) Ethernet frame structure

Ethernet standards all operate at the Data link layer in the 7-layer OSI model, so the data units that stations exchange with each other are frames. The structure of an Ethernet frame is as follows:

Figure 1.19. Ethernet frame structure

In there:

- Preamble: Marks the appearance of a bit frame, it always has the value 10101010. From this group of bits, the receiver can generate a 10 Mhz clock pulse.

- SFD (Start frame delimiter): This field actually defines the start of a frame. It always has the value 10101011.

- DA and SA fields: Carry the physical addresses of the stations receiving and sending the frame, identifying where the frame is sent from and where it will be sent to.

- Length: The value of the field indicates the size of the data portion of the frame carried.

- FCS (Frame check sequence) carries CRC (Cyclic redundancy checksum): The sender will calculate this field before transmitting the frame. The receiver will recalculate this CRC in a similar way. If the two results match, the frame is considered to be received correctly, otherwise

The frame is considered an error and is discarded.

2) Ethernet address structure

Each Ethernet network interface is uniquely identified by a 48-bit address (6 octets). This address is assigned when the device is manufactured, called the MAC address (Media Access Control Address).

MAC addresses are represented by Hexadecimal (base 16) digits. For example: 00:60:97:8F:4F:86 or 00-60-97-8F-4F-86.

The MAC address format is divided into two parts: The first 3 Octets identify the manufacturer, which is managed by the IEEE organization. The last 3 Octets are assigned by the manufacturer. Combined, we get a unique MAC address for an Ethernet network interface. The MAC address is used as the source and destination address in Ethernet frames.

3) Ethernet Frame Types

- Unicast frame: Consider the following network model:

Figure 1.20. Unicast frame

Suppose station 1 needs to transmit a frame to station 2. The Ethernet frame generated by station 1 has the addresses: Source MAC: 00-60-08-93-DB-C1, Destination MAC: 00-60-08-93-AB-12. This is

Unicast frame. This frame is transmitted to a specific station.

All stations in the above network segment receive this frame but:

+ Only station 2 sees that the destination MAC address of the frame matches the MAC address of its network interface, so it continues to process other information in the frame.

+ Other stations see that the destination MAC address of the frame does not match the MAC address of their network interface, so they will not continue processing the frame.

- Broadcast frame:

Broadcast frames have a destination MAC address of FF-FF-FF-FF-FF-FF (48 Bit 1). When receiving these frames, although they do not match the MAC address of their network interface, stations must still receive the frame and continue processing.

The ARP protocol uses Broadcast frames to find the corresponding MAC address.

with a given IP address.

Some routing protocols also use Broadcast frames for routers to exchange routing tables.

- Multicast frame:

The source station sends frames to certain stations, not all. The destination MAC address of the frame is a special address that only stations in the same group accept frames sent to this address.

4) Ethernet Operation

The CSMA/CD access control method defines the operation of the Ethernet system.

Some basic concepts related to Ethernet frame transmission:

- When the signal is being transmitted on the channel, the channel is busy and we call this state carrier.

- When the line is idle: No carrier – absence carrier.

- If two stations transmit a frame at the same time, they will detect a collision and must retransmit the frame.

- The time it takes for a network interface to recover after each frame is received is called the interframe gap – denoted by IFG. The value of IFG is 96 times the time of a bit.

Ethernet 10Mb/s: IFG = 9.6 µs Ethernet 100Mb/s: IFG = 960 ns Ethernet 1000Mb/s: IFG = 96 ns

- The way frame transmission and collision detection take place is as follows:

Step 1. When the station detects that the line is idle, it will wait for an additional period of time equal to IFG, then it will immediately transmit the frame. If multiple frames are transmitted, the distance between frames must be IFG.

Step 2. In case the line is busy, the workstation will continue to listen to the line until the line is free, then repeat step 1.

Step 3. In case the station detects a conflict while transmitting the frame, the station will have to continue transmitting 32 data bits. If the conflict is detected right at the beginning of the frame transmission, the station will have to transmit the entire Preamble field and another 32 bits. Transmitting these remaining bits (which we consider as congestion signaling bits) ensures that the signal will exist on the transmission line long enough for other stations (among the stations causing the conflict) to recognize the conflict and handle it: After transmitting all the congestion signaling bits, the station will wait for a random period of time, hoping that there will be no conflict and then retransmit the frame as in step 1. In the next frame transmission, if there is still a conflict

Network design - Master Tran Van Long, Master Tran Dinh Tung Compiled - 4

Comment