Portal field news

Portal field news

in ,

📱 | Japan Acer "Chromebook 314/712" review, will it be a new option in the cloud era?


Japan Acer "Chromebook 314/712" review, will it be a new option in the cloud era?

If you write the contents roughly
A multifunctional USB hub and dock with HDMI output and wired LAN will also be convenient.

● The red circle is soaring in Japan, and the cheap and crispy Chromebook is getting more attention in Japan ... → Continue reading

 My Navi News

Mynavi News is a comprehensive information news site with more than 100 categories updated daily. It consists of the latest news and analysis reports and columns from a wide range of genres such as business, digital, lifestyle, and entertainment to niche categories such as management, technology, railways, and cars.

Wikipedia related words

If there is no explanation, there is no corresponding item on Wikipedia.

USB hub

USB hubWhat is (USB Hub)?USBFor connecting multiple devicesHub.


Generally, the USB on the computer itselfportIs about 2-4, but todaykeyboard-mouse-printerBegins withExternal HDD,USB memoryからWebcam-Desk fanUntil then, the number of USB-compatible devices has increased significantly. A USB hub is used to connect these peripherals at the same time.

What is commonly known as a USB hub is a 4-port standalone product (see photo), but some are built into other peripherals.For exampleliquid crystaldisplayThere are about 3 ports on the side of the, and the USB hub function is installed on the USB-connected keyboard.

The shape of the housing is often a thin rectangle or square, but the arrangement of ports varies depending on the product and its application. In addition to products in which 4 pieces are lined up horizontally and products that are grouped in a 2x2 square, 2 pieces are placed on each side and 1 piece is placed on each side to avoid physical interference between connectors on adjacent ports. Products that have been made are often seen.In addition, there are products in which ports are lined up on the top surface of the hub body instead of the side surface, and products that can be connected from any direction by having ports at the ends of cables extending from the body body.Some have as many as 16 ports, but they are more expensive than those with 4 ports (the reason is).Technical explanationSee).

Most of the size fits in the palm of your hand, making it convenient to carry.This is because of the small number of parts required.Various products from each company that are particular about shape and designVCCIProducts with the mark are on sale.

Technical explanation

The specifications of the USB hub are part of the standard of the USB main unit (USB 2.0 Specification , Chapter 11).

USB hub is based on USB standardHost controllerHas an important role to work between and the device.The main roles are as follows

  • Device connection detection
  • Device communication speed detection
  • Powering and managing devices
  • Signal distribution
  • Communication speed conversion (USB 2.0)

Some USB hubs can turn on / off the power supply (VBUS) to their downstream ports.This ON / OFF can be commanded from the host side. The specification also specifies a protocol that notifies the host whether the USB hub has this function.In reality, there are many hubs that violate the specifications that report that they have this function but remain ON because the external electric circuit is omitted.

According to the specifications, the number of downstream ports that a USB hub can have is 255.However, it is less than that in a real USB hub implementation.During bus-powered operation, a maximum of 500mA can be taken out from the upstream, but since it is necessary to have the ability to supply at least 1mA per downstream port, adding four of these will result in a total of 100mA, and if the power used by the hub itself is combined, the maximum is four. Will be.Therefore, hubs with 4 or more downstream ports must always be self-powered.

A USB hub that supports the USB 2.0 standard has a speed conversion function for high speed transfer introduced in USB 2.0.If you have a device that is connected to the upstream at High Speed ​​and is communicating at Full Speed ​​or Low Speed ​​on the downstream port, buffer Full / Low Speed ​​communication once in the USB hub and then send it again at High Speed. It is a function.The part that executes this conversion is called TT (Transaction Translator).There are two types of hubs: one that has only one TT for the entire hub (only one port can be buffered at a given time), and one that has multiple hubs.The former is called Single TT and the latter is called Multiple TT.These two can be distinguished by the descriptor notified from the hub to the host.

A USB hub is itself a USB device and has a USB address (which isEthernetHub is on the networknodeIn contrast to not being considered).Therefore, a driver is required for the hub as well, but most USB-compatible systems have a hub driver, so the user does not need to be aware of it.

The USB host controller has a built-in circuit equivalent to a hub called a root hub, and can usually take out multiple ports (usually two in UHCI).

Compound device

Some USB devices have a built-in hub.Such a deviceCompound device (compound device) と呼ぶ[1][2]。コンパウンドデバイスでは、デバイス部分は内部的にハブの下に接続される形になっている。例として、2ポートのハブを内蔵したUSBキーボードの場合、内部的には、3ポートのハブと、そのうち1つのポートに接続されたキーボードの組み合わせとなっている。

A device with multiple functions (such as a keyboard with a built-in pointing device)Composite device It's called a (composite device), but it's not a compound device because it has nothing to do with the hub.The composite device isComposite deviceAlso called.


USB Type-Cの登場以降、USB Type-CからUSB Type-Aや、カードスロット、映像出力端子への変換を兼ねた機器が登場した。これらはポートリプリケータと呼ぶべき機器だが、ドック、USB-Cハブと呼ばれる場合もある。詳しくは、ドック(コンピュータ)See.

Related item



Ethernet (Ethernet) is the mainstream wired network technology used in homes, offices, data centers, businesses and WANs.Ethernet has evolved to support data rates up to 100 Gbps and distances from meters to tens of kilometers, making it essential for organizations of all sizes to support personal computers, workstations, servers, and high-capacity data storage devices. It has become a thing[1].Computer networkWired LAN (which is one of the standards of the company and is commonly used in offices and homes.Local Area Network) In the technical standard,OSI reference modelThe bottom two layers ofPhysical layerとData link layerIs stipulated regardingTCP / IPOften used in combination with.


With the progress of technology, new standards with higher communication speeds continue to appear, and the physical layer is divided into various standards depending on the difference in transmission speed and physical specifications, but the data link layer is old and new. Between the standards ofcompatibilityThere is.

The initial communication speed is 10 Mbps (Bit per second)of10BASE TSince, it has a transmission capacity of 10 Mbps, which is 100 times that.100BASE-TXHas become widespread, and today it is 1 Gbps1000BASE TIs becoming widespread.In addition, as a new standard, 10GBASE-T (UTPby10 Gigabit Ethernet<< 10GbE >>) The standard has been decided.As an even faster standard, 40 Gigabit Ethernet (40GbE) and100 Gigabit Ethernet (100GbE) is an organization that discusses international communication standards.IEEEIs in the adjustment stage.

Continuing to evolve from the early coaxial cable LANs, many LANs around the world now use Ethernet, which is a wider range of networks.MAN,WANBut some technologies are beginning to embrace the Ethernet standard under the name "Wide Area Ethernet."

The name "ether" is a virtual substance that was thought to fill every corner of the universe as a medium of light in the era of classical physics.ether”(Ether, Aether)[3].

In Japan, "Ethernet" and "Ethernet" areFuji Xerox 商標I have registered.[4]


The origin of the idea of ​​Ethernet was developed by Professor Norman Ebramson of the University of Hawaii.ALOHA systemIt is said.It was a system that connects the islands of the Hawaiian Islands with a wireless network of 4,800 bits / second.[2].

The first Ethernet is based on the idea of ​​the ALOHA system1972年 - 1973年To riceXerox OfPalo Alto Research Center In (PARC)Robert MetcalfeWas developed around. 19735/22,PatentThis day is considered to be Ethernet's birthday because it was registered as.The transmission speed at the time of invention was 2.94 Mbps, which is the computer Xerox under development at that time.AltoBaseク ロ ッ クIt is said that this is because it was adjusted to 5.88MHz.Xerox then released the patent to make it an open standard.IntelとDECIn addition to development,1979年, Establish DIX specifications by taking the acronyms of the three companies.The transmission speed was 3 Mbps.Next year1980年To this specificationIEEE Submitted and published as "Ethernet 802 standard" to the 1.0 committee.Companies and engineers from all over the world participated in the formulation of technical specifications and product development for this open standard, and various products were created.Metcalfe's own company left Xerox and the United States3ComHe founded the company and led the competition for network product development. In the 1980s, riceIBMThe company saidToken ring", RiceApple computer AppleTalk"Local Talk" was strongly promoted as a network product, but in the end, Ethernet, which released the standard and gained many supporters, survived.[5].

Ethernet, which is currently in widespread use,1982å¹´Based on the "Ethernet 2.0 standard" proposed in1983å¹´IEEE 802.3 CSMA / CDIt is a specification formulated as.

In the early days of Ethernet, 10BASE2 / 5 / -T, network support on the OS side was limited, and on PCs.NovellCompanyNetWare,Microsoft OfLAN ManagerI couldn't even get basic functions such as file sharing without purchasing special software such as. (As a workstation, for example, there was Sun's NFS (Network File System).)

On PCs from the 1980s to the 1990sNetwork interface card (NIC) or Ethernet cardPlug into a motherboard calledISA/EISA/NESA/PCIFormalDaughter cardWas often sold as an option, but in the early 2000sChipsetFrom the beginning, a part of the circuit was included, and an Ethernet jack was installed on the motherboard.By this time, the implementation of LAN functions by Ethernet became commonplace, and it became rare to use the term Ethernet itself (often referred to simply as a network interface). As of 2015, motherboards with two network ports from the beginning are readily available for both home and business use.

Communication technology

Ethernet isOSI reference modelInPhysical layeras well as the Data link layerIt regulatesIEEEByIEEE 802.3And its extended version, the specifications are published[6].HistoryAs mentioned in the column, the prototype was developed in 1970, submitted and published to the IEEE in 1980, and specified as IEEE 1983 in 802.3. Ethernet uses a 50Ω coaxial single-core cable and is a bus-type topology. It is a network with, and achieved 10 Mbps by half-duplex communication.It uses a stiff and heavy cable with a diameter of 1 cm10BASE5Then, the diameter of the coaxial cable was reduced to make it lightweight and easy to route, 10BASE2 Thin Ethernet cable, 10BROAD36 75Ω coaxial cable,FOIRLMultimode optical cables came to be used in, and then adopted a pair of unshielded twisted copper wires used for telephone wiring.1BASE5Established with (StarLAN-1) as the prototype10BASE TSince then, UTP cables have been used, and star-shaped configurations have been adopted for physical configurations.On the other handData link layerWill be described laterJumbo frame,VLANAlthough it is extended by, it is basically signal compatible, and frames can be exchanged with each other by connecting each standard using a media converter or the like.

Physical layer

OSI reference modelLayer 1 (physical layer) in is basically a bus type connection with a 50Ω coaxial cable.Even if it physically has a star-type configuration, it is logically a bus-type configuration.Currently, with the extension of IEEE 802.3, it is possible to use various infrastructures such as UTP cables and optical cables.

In Ethernet, when transmitting signalsmodulationIs performed.Baseband modulationIn the nameBASE,Broadband modulationIn the nameBROADIt is a rule to put on.

Manchester code .PNG

For baseband modulation, at 10BASExManchester codeWas used.The Manchester code is always Lo at the center of the signal indicating each bitHi and HiBy changing the signal level to Lo, it was possible to superimpose the clock signal, which is the reference for transmission, on the data signal.other,100BASE-TXThen 8B-6T,1000BASE TThen, suitable modulation such as 8B / 1Q4 (4D-PAM5) is used.

Early Ethernet had a bus-type configuration, both logically and physically, with multiple terminals connected to a single coaxial cable.When a large number of terminals are connected, "one-to-one" exclusive communication between any terminal A and B is impossible, and the data sent from terminal A is connected to the same Ethernet wiring. It is a "one-to-all" communication method that is delivered to all terminals. Since it is a "one-to-all" communication, if terminal C wants to newly transmit while terminals A and B are already communicating, it is necessary to wait for a free transmission line.If multiple terminals are connected and transmissions are performed at approximately the same time, collisions may occur, in which case data will be lost.thiscollisionIt is called (collision), and the countermeasures will be described later.CSMA / CD (Carrier Sense Multiple Access / Collision Detection). CSMA / CD supports up to Gigabit Ethernet (IEEE 802.3ab, etc.).

In Ethernet, the original communication data to be transmitted is first divided into a block of data having a fixed format of a certain length or less.Each of these chunks of data is called a MAC frame (Media Access Control Frame), or simply a frame.In Ethernet, data always flows through the transmission line in the form of MAC frames.Since the original data to be communicated is divided into frames, the network device only needs to handle the data of the frame of a certain length or less at one time, so all the processing related to information transfer is very simple. It comes down to the repetition of work.

Physical composition

The physical configuration of Ethernet is assembled with network devices (nodes) such as PCs and routers and cables.Since Ethernet is a logical bus type configuration, one logical bus range can be used.Collision segment(Or simply segment, collision domain, etc.).The nodes in the collision segment are each electrically equivalent, and the network interfaces of all the nodes receive all frames.The network interface of each nodeMAC addressAnd discard the received frames that are not addressed to you (it is also possible to listen to the communication contents of others by a device that monitors packets placed in the collision segment).Also,Switching hubEtc., when directly connected to a device that supports a layer 2 or higher networkFull dual communicationIt can be performed.

The network on which the same data arrives is called a "collision segment" or simply a "segment", and there is a maximum transmission line length due to the physical restrictions of collision detection, which is defined as the Ethernet standard for the physical layer.
If there are too many nodes in a collision segment, see below.collisionThe frequency of occurrence of is increasing at an accelerating rate, and the band is saturated when the threshold value is exceeded.
Also, if the distance between devices is longer than specified, it may not be possible to establish a data link.
If you need a transmission line longer than specifiedrepeaterOr the number of repeaters has increasedRepeater hubIt is possible to extend by.furtherbridgeAnd the bridge has become multi-portSwitching hubCan be used to extend the segment.
Full-duplex communication and half-duplex communication
"Full-duplex communication" is a communication technology that enables simultaneous transmission and reception at all times between terminals on one transmission line.
"Half-duplex communication" is a communication technology in which each terminal can perform one-way communication of either transmission or reception only by switching as necessary.In half-duplex communication, waste may occur when switching between transmission and reception.Telephones are full-duplex communication, and CSMA / CD is half-duplex communication.
As long as the terminal knows the signal emitted by itself, it is possible to obtain only the signal (and noise) transmitted from the received signal, and by the echo canceller technology that erases the echo component transmitted through the transmission line. Full-duplex communication is now possible.
Generally, between devices such as PCscrossWhen connecting with a cable orSwitching hubBy interposing a device that supports the data link layer or higher, full-duplex communication can be performed.Many switching hubs have a function that can automatically detect and switch between full-duplex and half-duplex (auto-negotiation).
MAC address
Only one unique 1-bit worldwide allocated during manufacturing to distinguish each terminal on EthernetMAC addresshave.The first 2 bits are the address type (unicast / multicast, global / local), and the following upper 22 bits are vendor eigenvalues ​​(in the case of global addresses. In addition, the first 2 bits are often represented by 24 bits. ), The lower 24 bits are serial numbers freely assigned by the vendor.It is a communication standard located above Ethernet.IPProtocol (IPv4) 32-bitIP addressIs different from.


What characterizes Ethernet CSMA / CD(Carrier Sense Multiple Access / Collision Detection; Carrier Sense Multiple Access / Collision Detection), which defines the means of sending signals on the communication path.Other physically similar methods,Token ringIn the case of, between the nodes on the routetokenA special packet called is sent, and only the receiving node can send a frame, but in Ethernet, each node can send a signal freely.On the other hand, of the signalcollisionIs generated, so constantly monitor the communication path (Career sense), Collision detected (Collision detection), After waiting for a while, the frame is retransmitted.Waiting time isTBEB algorithmTo decide with.The disadvantage is that even if there are many terminals waiting for transmission due to congestion, there will always be a vacancy in the communication path, resulting in wasted bandwidth.Also, as with other communication methods, the transmission speed represented by the standard is the communication capacity for all information that can be sent to the communication path, such as the header attached to the beginning of the frame.PreambleSince the inter-frame gap also consumes communication capacity, the transmission speed of user data is always below the standard speed to some extent. If TCP or IP is used, the header etc. will consume more communication capacity.

Career sense
All terminals constantly receive and monitor signals on the communication path.This function is called "Carrier Sense".
96-bit time from the end of the last frame if no frame has been sent from any other device after sending the frame[7]After waiting for the above, it is allowed to send the next frame to the communication path.This waiting timeInter-frame gapThe function in which multiple terminals share one transmission line in this way is called "Multiple Access".
Collision detection
If it happens that multiple terminals send out frames at almost the same time, a "collision", that is, a collision occurs.When signal disturbance due to collision is detected, the frame being received is discarded and a special signal called a jam signal is transmitted to the transmission line.If the transmitting terminal detects the disturbance of its own signal due to collision or detects a jam signal, the transmitting is immediately stopped and the frame being transmitted is returned to the state before transmission.This function is called "Collision Detection".
Collisions need to be detected before the sender finishes sending frames.Therefore, the total length of the communication path connecting two terminals connected on the same network and the minimum size of the frame are limited.In other wordsrepeaterThe total length of the cable through such as should be shorter than the length that can be reciprocated between the two terminals before the minimum size frame is sent.Basically, the minimum frame length in Ethernet is 2 bits = 512 bytes, and if the data to be sent is small and the frame length is less than 64 bytes, the shortage will be taken.0Fill with 64 bytes.thisPaddingCalled. In the case of communication of 1000Mbit / s (1Gbps) or more, the minimum frame length is 4096 bits = 512 bytes, and dummy data to make up for the shortage is provided.Carrier extensionCalled (described later).
Determining retransmission time
The terminal that stopped sending the frame is a pseudorandom numberAfter waiting for a random time by, if the communication path is free, you can send your own frame.If a collision occurs again at that time, 2nIt is randomly determined from the waiting time candidates.This method is called the "TBEB algorithm" (Truncated binary exponential backoff algorithm).

Data link layer

Layer 2 of the Ethernet standard, that isData link layerDefines the work related to creating a MAC frame to send and interpreting a received MAC frame.

Consider the case of transmission.Ethernet communication equipment that is a network terminalhostIn order to send the information from the computer to the communication path, first, if the received original data is long, it is divided into several chunks.This chunk of data is 46-1500 octets in size.Add the following additional information to this to complete the MAC frame.In the following explanation, 8-bit is not called a byte but is expressed as an octet, which is a term in the communication field.This is because there was a system in which 1 byte was not always 8 bits.

  • MAC frame configuration
    • Destination MAC address: 6 octets
    • Source MAC address: 6 octets
    • (VLAN: 4 octets)
    • Length (length / type): 2 octets
    • Data body: 46 --1500 octets (4 --42 octets if containing 1500 octets of VLAN)
    • Frame check sequence for error detectionFrame check sequence, FCS) 4 octets

There is a preamble of 8 octets in front of the destination MAC address, which is layer 96 as well as an interframe gap of 1 bit hours or more.Physical layerIt is not included in the standard of MAC frame which is layer 2 layer "data link layer" because it is automatically inserted by "".

The 8 octets of the preamble in the DIX standard are actually10101010It consists of a 7-octet preamble and 1 octet of Start frame delimiter (SFD).

destinationMAC addressAnd the 6 octets of the source MAC address have exactly the same configuration, and the first 2 bits indicate the communication type.Multicast,broadcast,UnicastAnd so on.The vendor-specific number purchased from the IEEE by the vendor of the network device that uses the MAC address in the following 22 bits is entered.The last 24 bits contain a number freely assigned by the vendor, and a serial number is usually used. The MAC address is the only unique number in the world (actually, there is equipment with a function to fake the MAC address for the purpose of taking turns in response to server failure repair replacement etc. To do).

IEEE 802.1QWhen using the optional VLAN in the standard, 4 octets are added before "length / type".フ レ ー ムThe whole is also 4 octets longer.If 4 octets for a virtual LAN called this VLAN are added, the maximum frame length will change from 1518 octets to 1522 octets, but the network device must handle it correctly.While some early network devices could not handle frames larger than 1518 octets correctly, modern network devices can handle frames up to 1522 octets correctly.

The type in the DIX standard and the length / type in the IEEE 802.3 standard may be mixed.If the value of this 2-octet field is 46 to 1500, it represents the length in the IEEE 802.3 standard, and if it is 1501 or more, it represents the type in the DIX standard.

The FCS determines whether the information in the four areas of destination MAC address, source MAC address, type, and data is correct.The error detection method for judgment is CRC (Cyclic redundancy check,Cyclic redundancy check) Use the law.

There is no signal indicating the end of the frame, and the receiving side determines the end of the frame when the last FCS signal is cut off.It's a rough method, but it allows data and FCS to contain completely free binary information, which simplifies frames and simplifies the processing of network devices.

Inter-frame gap
A signal-free inter-frame gap is secured between each frame for 96 bit hours or more.
This MAC frame is passed to layer 1, that is, the physical layer, and transmitted with a free transmission line.
Reception is the opposite, receiving received data through layer 1 / physical layer, and if your MAC address is not in the "destination MAC address", discard it as it is.If it is addressed to you, if the "length" is valid, the "length" is calculated back from the length of the entire MAC frame if it is not valid, and the length is calculated.Cut out the "data body" and "FCS" according to the length, check if there is a transmission error from these two, and if there is an error, it is usually the upper layer, layer 2, network layer and layer 3, transformer Report to the port layer, come down to layer 4-4-3 at the direction of layer 2 and send a retransmission request, or information is lost without being resent.If there is no transmission error, the cut out data is passed to layer 1 and the reception work of one MAC frame is completed.Layers 3 and 1 mentioned here are outside the Ethernet standard and are areas where the TCP / IP standard is used in many networks.That is, the Ethernet standard does not include retransmission processing or destination specification.

Layer 2 information is on the LAN if requested by the "destination MAC address"Switching hubIs transmitted across segments.In other words, in layer 1 alone, the switching hub is located at the boundary of the segment, and the transmitted signal does not cross the segment, but inside the switching hub, it is interpreted once up to layer 2 and the destination MAC address. If it is the next segment or the segment after that, it is transferred to another segment, so it is possible to cross the segment.For this reason, the MAC addresses of the terminals belonging to each connected segment are held as a list inside the switching hub, and the transfer destination is determined by comparing at high speed each time a MAC frame is received. ..The behavior of these Layer 2 switching hubs is the same for all speeds and formats of Ethernet standards.

About frame length regulation
For Ethernet, the shortest frame size is 64 octets and the longest frame size is 1518 octets.
CSMA / CDAtcollisionTo detect correctly, before the transmitted signal finishes transmittingTransmission lineThe transmitter must be able to receive the result of the collision above and the signal disturbance.

For this purpose, the transmitted signal must have a certain length, and the length must be calculated assuming that the collision detection takes the longest time.It takes the longest time for a signal transmitted from one end of a transmission line to start another transmission the moment it reaches the other end, so this time the first signal propagates from end to end It is the sum of the time it takes to do this and the time it takes for the collision signal to return on the same path, which is simply the time it takes for the signal to propagate twice as long as the transmission line.

The calculation of the round-trip radio wave delay time for detecting CSMA / CD collision is 2,500 μSec in total, which is the delay time of about 2 μSec for the cable for the round-trip 5,000 m × 26 20 m and the delay time of about 46.38 μSec for the repeater and transceiver. Is calculated.

This length corresponds to 10 bit times in 464BASE-x.

It is the first full-scale Ethernet standard from these calculations.10 megabit ethernetThen, the shortest frame size was set to 512 bits (64 octets) time with a margin.
There is no clear reason to decide that the maximum frame size is 1518 octets, and if it is too long, the packet buffer for transmission and reception of the device that performs the transfer process.memoryIf the capacity is too short, only the frame header will flow on the network and communication efficiency will be poor. Therefore, considering the hardware environment at that time, the longest data length should be used. I decided to use 1,500 octets.The length of the frame header was added to this, and the longest frame size was decided to be 1518 octets.VLANIf you use, a VLAN tag (4 octets) will be added, for a total of 1522 octets.

Individual Ethernet standard name

同軸,Twisted pair,Optical fiberVarious standards are specified due to differences in communication paths such as, and transmission speeds such as 10M / 100M / 1G / 10Gbps.The general method of assigning Ethernet standard names is shown below.

In "10BASE-T", "10" is a transfer rate of 10Mbps, and "BASE" is not modulated.Baseband transfer, "T" is unsealed in the communication pathTwisted pair cable It means to use (Twisted pair cable).Of course, not all combinations of the above are available.

Equipment and cables

The equipment and cables for configuring Ethernet will be described.


It feeds the physical layer of Ethernet, relays signals, configures MAC frames, and sends and receives frames according to CSMA / CD.This section describes devices that support the physical layer and data link layer specified by Ethernet.

Repeater, bridge

A repeater is a device that relays and reproduces electrical signals between networks to extend the network.Supports the physical layer.Currently as a multi-interface repeaterHubIs often used.
The bridge has features supported by the data link layer.


Hubs (dam hubs, repeater hubs, cascade hubs, etc.)
A hub is a device that plays a central role in a star-type configuration, connects a plurality of terminals, and relays and reproduces signals as repeaters. A device that transmits signals in the first layer (physical layer) of the OSI reference model.
Switching hub
A switching hub is also called a layer 2 switch, and as the name implies, it is a device that transfers information in the second layer "data link layer" of the OSI reference model, and determines the transfer destination of the MAC frame of the Internet by the MAC address. A "layer 2 switch" (L2 switch) is the most typical Ethernet network device, also called a "LAN switch" or "switching hub".Originally, there was a relay amplifier on the transmission line called a repeater hub in the early days of Ethernet, and a switching hub, that is, a layer 2 switch, was a major development of that function.


coaxial cable

Both 10BASE2 and 10BASE5 use a 50Ω coaxial cable. 10BASE5 uses a so-called Thick cable (or yellow cable) with a diameter of 12 mm.10BASE2 uses a so-called Thin cable, which is about half the size of a 10BASE5 cable and has a diameter of 5 mm.

Fiber optic cable

Optical cablereference. Used with 10BASE-F, 100BASE-FX, 1000BASE-SX / LX, etc.Generally, multimode fiber (MMF) and single mode fiber (SMF) are used.Generally, MMF has a thick core wire, is strong against bending, and is easy to handle, but the communication distance is short and the speed is slow.Suitable for indoor wiring, etc. SMF has a thin core wire, is vulnerable to bending and is expensive, but has a small transmission loss and is suitable for long-distance communication.

Twisted pair cable

Of the cables used for Ethernet, the most commonly used is Twisted Pair Cable.Twisted pair cables can be broadly divided into "UTP" and "STP".

  • UTP is an abbreviation for Unshielded twisted pair cable, which is an unshielded, twisted pair cable.UTP is a category 6 or lower grade.
  • STP is an abbreviation for Shielded twisted pair cable, which is a shielded, twisted pair cable.Augmented Category 6 or higher grade is STP[Needs verification].

Cables that satisfy the frequency characteristics corresponding to each of these transfer speeds are called "LAN cables" and "Ethernet cables" and are sold.Cable grades are categorized by the number of categories according to each upper limit frequency.Most have RJ-45 plugs on both ends already connected, but you can easily connect the plugs to a cable of your desired length using a special tool.The plug attached to the cable is male at both ends, and the network device side is always a female socket.The cable that is normally used is called a "straight cable", in which the same pin numbers of the RJ-45 plugs on both ends are connected straight, but when connecting multiple old hubs in cascade, or when using Ethernet on a PC -When connecting ports on a one-to-one basis, use a "cross cable" with pin numbers 1-1, 1-3, 2-6, and 4-7.All connection ports, including cascade ports on many switching hub products currently on the marketAutoMDI / MDI-XIn most cases, there is no need for a cross cable because it has an automatic discrimination function between MDI and MDI-X called.

The number of categories and applicable transfer rate standards are shown below. STP has high frequency characteristics due to shielding and other devices.

  • UTP
    • Category 3: ~ 16MHz: 10BASE-T
    • Category 5: ~ 100MHz: 100BASE-TX
    • Enhanced Category 5: ~ 100MHz: 100BASE-TX, 1000BASE-T, 2.5GBASE-T
    • Category 6: ~ 250MHz: 1000BASE-TX, 5GBASE-T
    • Enhanced category 6: ~ 500MHz: 10GBASE-T[Needs verification]
  • STP
    • Augmented Category 6: ~ 500MHz: 10GBASE-T etc.[Needs verification]
    • Category 7: ~ 600MHz: 10GBASE-T, etc.

For UTPs of category 6 or lower, there is no problem even if a high-grade cable with upward compatibility and excellent frequency characteristics is used for connection at a lower transmission speed.The abbreviations for the number of categories are displayed on the cables sold as "Cat5" and "Cat5-e".

Cable standard list[8][9][10][11][12][13][14][15][16][17]

CategoryTransmission speedUsable distanceTransmission frequency bandNoise shieldOpposite lineConnector shapeRemarks
CAT120kbps4 cores 2 pairsRJ-11Used for voice phones (modular jacks), etc.
CAT24Mbps1MHzUTP8 cores 4 pairsRJ-45Used for ISDN, etc.
CAT310Mbps100m16MHzUTP8 cores 4 pairsRJ-45
CAT410Mbps100m20MHzUTP8 cores 4 pairsRJ-45
CAT5100Mbps100m100MHzUTP8 cores 4 pairsRJ-45The transmission speed in the ATM standard is 155 Mbps, and the Ethernet standard 100BASE-TX uses only 4 pairs of 2 cores.Since the current optical communication is 1 Gbps, it is often cited as a cause when the communication speed is slow.
CAT5e1Gbps100m100MHzUTP8 cores 4 pairsRJ-45Current general specifications that support up to 1 Gbps.

Affordable price and enough specs for everyday use.

CAT61Gbps100m250MHzUTP8 cores 4 pairsRJ-45Transmission speed in ATM standard is 1.2Gbps
10Gbps55m250MHzUTP8 cores 4 pairsRJ-45
CAT6a10Gbps100m500MHzUTP / STP8 cores 4 pairsRJ-45
CAT710Gbps100m600MHzSTP8 cores 4 pairsGG45 / TERAIt is the fastest but expensive, so it is often used for corporate servers.
CAT7a40Gbps50m1000MHzSTP8 cores 4 pairsGG45 / TERA
100Gbps15m1000MHzSTP8 cores 4 pairsGG45 / TERA
CAT840Gbps30m2000MHzSTP8 cores 4 pairsGG45 / TERA

Ethernet specifications

Standard namealiascommunication speedStandardization standardCable used距離
1BASE51MbpsIEEE 802.3eUTP (2 pairs)250m
10BASE2Thin Ethernet10MbpsIEEE 802.3a50Ω coaxial (5mm)185m
10BASE5Thick EthernetIEEE 802.350Ω coaxial (12mm)500m
10BASE TIEEE 802.3iUTP / STP (Cat3)100m
10BROAD36IEEE 802.3b75Ω coaxial3600m
10BASE-F10BASE-FBIEEE 802.3jOptical multi-mode2000m
10BASE-FPOptical multi-mode1000m
10BASE-FLOptical multi-mode2000m
100BASE T100BASE-TXFast Ethernet100MbpsIEEE 802.3UTP (Cat5)100m
100BASE-T4UTP (4 vs. Cat3)100m
100BASE-T2IEEE 802.3yUTP (2 vs. Cat3)100m
100BASE-F100BASE-FXIEEE 802.3Optical multi-mode / single mode2000m / 20km
1000BASE T1000BASE TGigabit Ethernet1000MbpsIEEE 802.3abUTP (4 vs Cat5e)100m
1000BASE-TXTIA-EIA / -854UTP (4 vs. Cat6)100m
1000BASE-X1000BASE-SXIEEE 802.3zOptical multi-mode550m
1000BASE-LXOptical multi-mode / single mode550m / 5000m
1000BASE-CXCoaxial cable (2-core parallel)25m
2.5GBASE-T2.5GbpsIEEE 802.3bzUTP (4 vs Cat5e)100m
5GBASE-T5GbpsIEEE 802.3bzUTP (4 vs. Cat6)100m
10GBASE-T10GbpsIEEE 802.3anUTP (4 vs. Cat6)100m
10GBASE-R10GBASE-SRIEEE 802.3aeOptical multi-mode300m
10GBASE-LROptical single mode10km
10GBASE-EROptical single mode40km
10GBASE-ZROptical single modeAssuming 40km or more
10GBASE-W10GBASE-SWOptical multi-mode300m
10GBASE-LWOptical single mode10km
10GBASE-EWOptical single mode40km
10GBASE-X10GBASE-LX4Optical single mode10km
10GBASE-CXIEEE 802.3ak4-to-2 core copper wire (CX4) coaxial15m
25GBASE-T25GbpsIEEE P802.3bqUTP (4 vs. Cat7)30m
40GBASE-T40GbpsIEEE P802.3bqUTP (4 vs. Cat7)30m
40GBASE-R40GBASE-KR440GbpsIEEE 802.3ba
100GBASE-R100GBASE-CR4100GbpsIEEE 802.3bj
100GBASE-CR10IEEE 802.3ba


  • Nikkei Network February 2002 "First Gigabit Ethernet"
  • Nikkei Network July 2003 "Mechanism of Communication in Layers"
  • Nikkei Network November 2005 Issue "The Essence of Ethernet Revealed in the Development Process"
  • Nikkei Network December 2005 Issue "The Meaning and Significance of CSMA / CD"
  • Network Magazine Editorial Department "Switch & Router Expansion / New Edition Starting from Zero" ASCII, September 2007.ISBN 978-4-7561-5004-2.
  1. ^ Stallings, William ,. Foundations of modern networking: SDN, NFV, QoE, IoT, and Cloud. Agboma, Florence ,, Jelassi, Sofiene ,. Indianapolis, Indiana. ISBN 978-0-13-417547-8. OCLC 927715441. https://www.worldcat.org/oclc/927715441 
  2. ^ a b Nikkei Network October 2005 Issue "Inherited, Created"
  3. ^ Ethernet 40 years technology - ITpro(December 2013, 12). Retrieved December 9, 2013.
  4. ^ Japanese trademarkPatent information platformYou can search with.Some trademarks including "Ethernet" are registered by other companies.
  5. ^ Nikkei Network October 2005 Issue "Ethernet Technology Reader" p10
  6. ^ IEEE 802.3 ETHERNET (English)
  7. ^ イーサネット上で1ビットの情報をやり取りする時間を示す。例えば10Mbpsでは1ビット時間は10M分の1の100ナノ秒となり1Gbpsでは1G分の1の1ナノ秒になる。
  8. ^ LAN cable standards and categories Wiring-less cube log
  9. ^ List of longest cable distances by standard Aim Electronics Co., Ltd.
  10. ^ Differences in LAN cables and how to choose them Web convenient notebook
  11. ^ [TECH blog] About CAT cable. 40GBASE-T! PGD Blog
  12. ^ What is the difference between Cat5e, CAT6, CAT6A, CAT6E and CAT7 LAN cables? TeraDas
  13. ^ CAT3 and CAT4 Black Box Glossary
  14. ^ ASCII.jp: Can Gigabit Ethernet be used with old LAN cables | Touching network contraindications
  15. ^ [10GBASE-T is finally popular? ] Even faster! How many years will the "40G-25GBASE-T", which can handle 40Gbps, become widespread! [Expected new Internet technology] --INTERNET Watch
  16. ^ Ethernet is a LAN cable!Explaining cable types and how to choose | WiFi style
  17. ^ LAN cable type | Category and selection method

Related item

外部 リンク



[You can easily write your impressions using Twitter! ]
Just press the button below and write after "Impression of the article:" and tweet! In that case, please do not delete the input from the beginning, please put a space after the impression so that sentences will not be connected to the URL.
The official account of PORTALFIELD News will spread (follow, like, retweet) your impressions. After tweeting, you are free to spread or delete it yourself, so feel free to send us your thoughts! Thank you for your cooperation.
You can see the impression tweet to this article sent to Twitter.
Back to Top