10 Tips of Industrial Ethernet Technology (Tip 5 - 10)
5. Speed and Distance
Discuss the distance of the Shared Ethernet, and you can't ignore the Collision Domain concept.
Media access to a Shared Ethernet or half duplex Ethernet is determined by the CSMA (Carrier Sense Multiple Access)/ CD (collision detection). In a halfduplex mode of communication, the sending and receiving cannot occur simultaneously, otherwise the data will collide. Before the site is sent, it depends on whether there is a free channel. When sent, the site will be tuned for a period of time to ensure that no other site is being delivered synchronously and eventually the site is successful. Conversely, a collision occurs when the source site sends a blocking signal to reinforce the collision. Retry after the competition site is delayed (the delay time is determined by the algorithm, it is random). Under this mechanism, all sites and all hubs must be within the same collision domain.
For industrial Ethernet, 10Mbps and 100Mbps are the most commonly used. In 10Mbps, there are two concepts associated with distance, namely the Segment (Segment) and Network Diameter. The former refers to the distance between two devices (hub, switch, or host). The rest one refers to the distance between the two farthest devices in the network. Whether it's 10Mbps or 100Mbps, the maximum distance from the segment can't exceed 100 meters. The most useful rule for extending the network is the 5-4-3 rule (only for the 10Mbps repeater). The rules are as follows: a network with up to five web segments, four Repeaters, and no more than three hybrid segments. The hybrid segment refers to the coaxial bus network segment (obsolete). The maximum distance from the double stranded wire segment is 100 meters, and the maximum network (network range) is 500 meters.
The longest distance between the fiber network segment can be up to 2 km, but to IEEE802.3 standard regulation, cascade cannot exceed the maximum number of three in the use of fiber, the end of the network needs to use twisted pair cable, and in the middle of the two uses the fiber network segment to ensure that each segment is less than 1 km. Thus, the length of the entire fiber network is limited to two kilometers.
As the 5-4-3 rule does not apply to 100Mbps, a 100Mbps switch is recommended.
6. Hub and Switch
The hub is the basic equipment that makes up the Ethernet topology. Most of the hubs in current market are multi-port equipments, four ports, eight ports and twelve ports etc., which can form a decentralized star topology.
The hub meets the requirements of the IEEE802.3 relay unit, which include preamble generation, symmetry and amplitude compensation. The repeater must be timing the signal so that the signal jitter caused by the transceiver and cable will not accumulate when the multi-mesh segment is propagated. These devices can detect incomplete packets and conflicts and produce a blocking signal. They also automatically isolate problematic ports to keep the network working.
Another series of products are interface mediators, sometimes called transceivers. The use of time is turning one medium into another. The most important conversion is the twin-strand to optical fiber. As many hubs have no optical fibre ports, the interface mediators are used to support the application of optical fibers in the network. These devices are independent in the network. Ports do not store frames or detect collisions, but instead converting one medium to another.
Switch hubs can replace repeater hubs and improve network performance. Unlike the physical layer device (the repeater hub), the switch hub is actually the bridge that connects two data links, namely that the collision domain ends on each switch port. Therefore, the increase of the switch is to expand the geographic scope of the network. What's more, the cascade switch can achieve the network extension on a large scale.
The switch is more complex than the repeater hub. Double stranded port is automatically negotiated with the attached port (10Mbps or 100Mbps). Traffic control functions are also negotiated. PAUSE is adopted in full duplex segment, while the backpressure is usually adopted in the semi-duplex segment. The switch reads a full frame and looks at its source address to find out the port location of the Ethernet device. And the switch then generates a port address table and maintains the contents of the table. From this point on, network communication is limited to ports associated with this transmission. The throughput of the network has been increased because the synchronous transfer does not require any operations to be implemented on these ports. The contents of the table are automatically refreshed according to the change of the connection information. If the information received from a port needs to be broadcast, group sent or sent to an unknown address, the switch would automatically send the message to all ports.
Differrent from repeater hub, there are multiple collision domains existing, and each collision domain must comply with the above rules.
The repeater hub can be connected to port of switch. If the network is full of switches, the segment of double stranded wire is maintained 100 meters, but the cascade is not limited. Before using the optical fiber, must be marked half or full duplex.
Comparison between repeater hub and switch hub
Obviously, the performance of the switch is higher than the hub, but the advantage of the hub is that it is easy to understand that data communication can be observed on any port through the network analyzer. While switches must be broadcast on a port to be measured. As a bridge, the switch stores and forwards the entire data frame and causes the delay of the data. The hub receives network signals without delay of data. Switch cascades also increase latency, so hubs and switches have applications in industrial Ethernet.
7. Half Duplex / Full Duplex
Half duplex means that the same media is sent and received asynchronously. Full duplex, on the other hand, has a separate transmit and receive path. Full duplex link is the key to expanding fast Ethernet (100Mbps). The device quantity limit of full duplex link network segment is less than 2 sets. The device could be either the network card or the port of switch. Note: it is not a repeater hub port, and the hub does not have full duplex mode due to the hub is part of the collision domain and it enhances the collisions that other ports receive. It can only achieve full duplex communication with two network cards, if the quantity is over than that, it is only achieved by switch. There are separate transmission and receiving channels for 10base-T and 10base-FL, with which it can perform full duplex ccording to the complexity of network card or switch ports. If these ports are configured in a half duplex mode, the synchronous detection of receiving and sending would trigger collision detection. While the same port is set up for full duplex, and the collision detection will be banned due to the full duplex does not follow the Shared CSMA/CD rule.
The full duplex link configuration should be correct. When the site is configured in full duplex mode, the site or the switch hub port would send the frame by ignoring the CSMA/CD protocol. If the other end is set in a half-duplex mode, it detects the collision and would cause other problems, such as CRC error, the speed drop of the network , and the advantage disappear of fast Ethernet.
As mentioned earlier, the network scope of 100Mbps has been reduced due to the collision. Therefore, to the double-stranded network segment and the switching port, the maximum distance of the segment is 100 meters (in the collision domain). The problem is that for multimode fiber, the length of the segment is 2 kilometers. And for single-mode fiber, it's 15 kilometers. In the half-duplex mode, the collision domain is restricted, and the network distance is 412 meters. Therefore, only in the full-duplex mode (CSMA/CA is ignored), it can reach the extension limit of the fiber-optic network.
In conclusion, switch technology is recommended for fast Ethernet. And the fast Ethernet fiber ports are recommended to be under full duplex mode.
With fast Ethernet popular adopted, the traditional Ethernet port can work with other fast Ethernet ports due to fast Ethernet connection rules similar to traditional and auto configuration for fast Ethernet suggested by IEEE802.3u. This configuration protocol is based on National Semiconductor's NWay standard. The double strand link is automatically matched to facilitate data communication and the scheme is applicable to double stranded links. While it is different to fiber optics. Although optical fiber has a very important position in the development history of Ethernet., the speed of two optical fiber devices cannot be automatically negotiated, because the 10base-fl equipment works at 850nm and 100base-fx equipment works at 1300nm. The two cannot interoperate. However, for automatic negotiation protocols, self-negotiation between two fiber-optic devices is feasible (if there is no problem with communication). Aware of this, the new 100base-sx standard allows 850nm fiber to work under 10Mbps or 100Mbps. The net segment distance under 100Mbps is 300 meters. Therefore, please note that the rate of optical fiber is usually fixed and non-negotiable. And the automatic negotiation protocol is successful in the double strand link. The advantage of automatic negotiation is that it enables users to set their own technical standards by themselves without manual setting.
The level is as follows:
1000BASE-T Full duplex, high
100BASE-T2 Full duplex
100BASE-TX Full duplex
10BASE-T Full duplex
The lowest level is 10base-t (half duplex, Shared Ethernet), highest level is 1000base-t full duplex. This is a complete priority level solution, but it doesn't mean that a network card can handle all of these technologies. In fact, there are some technologies that are not commercially implemented, but they are all consistent with the IEEE802.3 standard. Each port examines its technical performance and determines the final rate (lower rate). For example, if the network card supports 10base-t and the switch port capability is 10base-t or 10base-tx, then the final choice is 10base-t. And if a network card is 10base-t, and the other is 100base-tx, the two cannot communicate because they are incompatible.
9. Transmission Protocol
The original design didn't involve a reliable end-to-end message transmission. The obligation of network interconnection (two networks to communicate with each other) is on the third layer, the network layer.
Transport and interconnection are part of the protocol stack, and TCP/IP and SPX/IPX are commonly used. These two protocols do not operate on each other, so the Ethernet nodes must use compatible protocols. As TCP/IP is used on the Internet and industrial network , it becomes the main protocol. In fact, TCP/IP is a set of RFC defined protocols (request for comments) for years. In addition, to Ethernet, TCP/IP also works with other data link technologies, which are located on the physical layer/data link layer. On the transport layer, there are two important protocols: TCP and UDP. The former confirms the information received. Both are useful.
At the top of the protocol stack, there are several useful application layer protocols used in industrial Ethernet. Addressing is an important topic for users. The IP protocol is responsible for routing between sites across different networks. Each site has a unique 32-bit address (the network address, host address, respectively). The address is represented in decimal four bytes. 126.96.36.199 is an effective address, but it is impossible to determine what a host is and what the network is. The address is divided into five categories, and the address is divided into A ~ E class. You can sort by looking at the first byte.
IP allocation is not simple, usually distributed by administrator. Once distributed, it must be applied to every workstation in the network. There are two kinds of IP address static and dynamic allocation. Dynamic allocation is done by the server and static allocation is done by configuration. The following address is a private address and cannot be allocated on the router. Therefore, they are not available on the Internet.
10.0.0.0 ~ 10.255.255.255
The IP address and the Ethernet MAC address are different and cannot be confused.The MAC address is distributed by device manufacturers, so it is the only one in the world.The IP address is allocated at the time of installation and redistributed as needed.
10. Application Layer Protocol
After know above parts, now you need to consider the compatibility of OSI executives. The industrial automation protocols recommended here include Ethernet/IP, iDA, PROFInet and Modbus/TCP. This does not include traditional Internet application-ftp, SNMP, SMTP, and TELNET. And some Devices in current market may not support these protocols, so you need to know the compatibility of your system.