IT History: History of LAN [2004-11-5] LAN Development History - Network Engineer Alliance (CCNU) In the last 25 years, Ethernet has developed from 4800bps disputeless radio transmission system to the most popular local network standard And can transmit 100 megabits of information per second on the unsatched twisted pair. The history of Ethernet is so attractive, so that countless technologies, prides and famous companies, are worshiped under her pomegranate skirt. People have seen technology prospects and tempting wealth from its development history. In fact, the entire industry will take out the concept of different computer equipment. Ethernet origin: Aloha Radio System (1968--1972) Ethernet's core thinking is the use of shared public transport channels. The idea of sharing the data transfer channel is from the University of Hawaii. In the 1960s, Norman Abramson and colleagues of the school have developed a radio network called an Aloha system. This ground radio broadcast system is developed in order to connect the IBM360 host in the campus on Oahu Island and the readers distributed on other islands and marine vessels and terminals. The initial speed of the system is 4800 bps, and finally upgrades to 96O0 BPS. The uniqueness of the system is to use "Inbound" and "OutboundL" radio channel for two data transfer. Outbound radio channel (from the mainframe to the island) is quite brighter, as long as the end address is placed in the transferred content title, then the corresponding receiving station is decoded. The entry radio channel (from the island or ship to the host) is more complicated, but it is very interesting. It uses a random retransmission method: sub-station (station on the island) issued it after the operator tap the return button. Wentine or packet, then the station waits for the primary station to send back confirmation; if there is a certain time limit (200 to 1500 nine seconds), the remote channel is not returned on the exit channel, the remote station (deputy station I will think that two stations are attempting to transmit simultaneously, thus have collision conflicts, so that transmission data is damaged. At this moment, the two stations will once again choose a random time, try to return their packets, at this time, the successful grasp is very Big this type of network is called a contention network because different stations are playing the same channel. This contention network has two meanings: This mode allows multiple nodes to accurately transmit on the same - channel channel with simple and smart methods. The clocks used to use the channel, the higher the chance, the higher the chance, resulting in increased transmission delay and low information flow. Norman Abramson published a series of articles on the theoretical and application of Aloha systems, in which an article in 1970 explained in detail the mathematical model of the theoretical capacity of the Aloha system. Now this model is known from the classic Aloha model, and it evaluates the theoretical capacity of the Aloha system to reach 17%. In 1972, Aloha improved to a time slot Aloha group broadcast system by simultaneously accessing the efficiency more than doubled. The development of Abramson and its colleagues has become the basis for most of the currently used packet broadcasting systems, including Ethernet and multiple satellite transmission systems. In March 1995, Abramson won the Kobayashi Award for IEEE because of its groundbreaking research in the dispute system. Xerox Parc created the first Ethernet (1972-1977) Today we know the Ethernet in 1972, when Bob Metcalfe came to the Xerox Palo ALTO Research Center (PARC) computer science laboratory, Xerox is a famous world Research institutions. In 1972, Parc researchers have invented the world's first laser printer and the first PC called ALTO-called graphical user interface.
At that time, Metcalfe has been hired by Xerox as a PARC, his first job is to connect Xerox Alto Computers to Arpanet (Arpanet is the formerity of the Internet). In the fall of 1972, Metcalfe is visiting the administrator of the Arpanet program in Washington, DC, and accidentally discovered Abramson's research results on the ALOHA system. When reading Abramson's 1970 papers on the Aloha model, Metcalfe realized that although Abramson has made some questions, it can increase the efficiency of the Aloha system to nearly 100% by optimization. Finally, Metcalfe received a Ph.D. in Harvard University because of his packet-based transmission theory. At the end of 1972, Metcalfe and David Boggs designed a network that connects different Alto machines and then connects the NOVA computer to the EARS laser printer. During the development process, Metcalfe named his work name from Alto Aloha network because the network was based on an Aloha system, while a large number of Alto Computers. The first personal computer local area network in this world is the first time to run on May 22, 1973. On this day, MctCalfe wrote a memorandum, saying that he has renamed the network as an Ethernet, which is inspired by "electromagnetic radiation is the idea that can be spread by glowing." The initial experimental PARC Ethernet runs at 2.94 Mbps (megabit) speed, this speed value is a bit too zero, why the first Ethernet interface timer adopts the ALTO system clock, meaning every 340 femto The second pulse is sent, resulting in 2.94 Mbps, of course, the Ethernet has a huge improvement than the initial Aloha network, because Ethernet is characterized by carrier audition, that is, each station is to transfer its own data stream It has been to listen to the movement of the network, so an improved retransmission program enables the utilization rate of the network to increase by nearly 100%. By 1976, in the PARC's real-test Ethernet has evolved to 100 nodes, which have been run on a coaxial cable for 1000 meters. Xeror is eager to convert Ethernet into a product, so the Ethernet is renamed Xerox Wire. However, in 1979, DEC, Intel and Xerox could standardize this network, and the network recovered this name. In June 1976, Metcalfe and Boggs were published in: "Ethernet: LAN's Distributed Packet Exchange", the end of 1977, Metcalfe and his three partners have obtained multi-point data with conflict detection. The patent of the communication system, the multi-point transmission system is called CSMA / CD (carrier monitor multi-channel access and conflict detection). Since then, Ethernet is officially born. Dec, INTEI and Xerox will standardize Ethernet (1979-1983) in the late 1970s, dozens of local area network technology have emerged, and Ethernet is one of them. In addition to Ethernet, the most famous networks at that time were: Data Universal Company's MCA, network system company's HyperChannel, Data 'Point's ARCNET and Corvus's Omninet. The Ethernet finally sits on the LAN treasure, but the Ethernet of the Metcalfe version has become an industrial standard. At the beginning of 1979, he returned to Xerox Parc's Metcalfe to the phone of Gordon Bell at DEC.
Bell wants to discuss Dec and Xerox to build Ethernet LAN's ideas, Metcalfe believes that the idea of developing Ethernet in Ethernet with different manufacturers, but Metcalfe is a bit of itself, because Xerox wants to protect its patents, limit Metcalfe to DEC . Therefore, Metcalfe recommends that DEC directly discusses the Xerox main management plan to transform Ethernet into an industrial standard, and finally Xerox has taken this step. One of the obstacles that make DEC and Xerox cooperate in industrial standards is an anti-laratue. Metcalfe's friend HoWard Charney lawyer, suggested that he transferred true Ethernet technology to a standardized organization (shortly Charney as one of the founders of 3COM). When Metaclfe visited the US Standardization Bureau (NBS) in Washington DC, MetaClfe met an engineer in NBS working in the Washington DC, which is looking for new applications for his advanced 25MHz VLSI NMOS integrated circuit processing technology. The advantage of Zhu Lianbi is obvious: XEROX provides technology, DEC has strong technical strength, and is a powerful supplier of Ethernet hardware, Intel offers Ethernet chip members. Soon, Metcalfe leaving Xerox into entrepreneurs and brokers. In July 1979, DEC, Intel and Xerox preparations were held in July, in 1979, in 1979, it was officially held for the first three conferences. On September 30, 1980, DEC, Intel and Xerox announced the third draft "Ethernet, a local area network: data link layer and physical layer specification, 1.0 version", this is the famous Ethernet Blue Book, Also known as DIX (taken by the first letter of three companies name) Ethernet 1.0 specification. As mentioned earlier, the initial experimental Ethernet works in 2.94 Mbps, while the DIX starts to run under 20Mbps and finally falls to 10Mbps. DIX redefines the standard in the next two years and published the Ethernet version 2.0 specification as an end in 1982. While DIX works on Ethernet standardization, the World Professional Organization IEEE forms a committee defined and promoting the LAN standards of the Industry, and is the main goal of the office environment, the committee named 802 project. Although the DIX Group has launched an Ethernet specification, it is not an internationally recognized standard, so in June 1981, the IEEE802 project decided to form an 802.3 subcommittee to produce an internationally recognized standard based on DIX work results. After a year and a half, that is, 1982 From 12:00, 19 companies announced a new IEEE802.3 draft standard. In 1983, the draft was eventually used in IEEE10 Base5. (Selecting the quotation 10Base5 is because the standard specifies the 10Mbps transfer rate using the baseband and the distance between the nodes is 50 meters, 802.3 and DIX Ethernet 2.0 are technically different, but this is very different.) Today Ethernet and 802.3 can be considered synonymous. During this time, Xerox has transferred its 4 Ethernet patents to IEEE, so anyone can now use $ 1,000 from IEEE to obtain an Ethernet license. The 1984 US federal government adopted 802.3 standards in the name of FIPS PUB107. In 1989 ISO adopted 802.3 Ethernet standards by standard number IS88023. At this point, IEEE standard 8O2.3 officially recognized internationally. 3COMs will be Ethernet product (1980-1982) in DEC, Intel, Xerox engineers still in the final processing for Ethernet specification, Metcalfe has undergone other commercial interests, and the wellsterer will recommend him to participate in Apple Computer Company. Development network recommendations.
In June 1979, Bob Metcalfe, Howard Charney, Ron Crane, Greg Shaw and Bill Kraus form a computer communication and compatibility company, is now a famous 3COM company. In August 1980, 3COM announced its first product, which was used in UNIX commercial Edition TCP / IP, and was officially launched in December 1980, and a grand business plan was developed in February 1981. 3Com has received a large risk fund. In March 1981, the 3CM company has put its first product (3C100 transceiver) in the first batch of 802 standard products (3C100 transceiver). At the end of 1981, the company began selling the DEC PDP / 11 series and VAX series transceivers and cards, and also sells Integrat Multibus and Sun Microsystems. The initial business plan for Metcalfe is to invest in 1980 to develop Ethernet adapters for new PC, because new personal computers have just happened around the world. In 1981 Metcalfe and all large-scale PCs (including IBM and Apple) discuss planning to build an Ethernet adapter. Steve Jobs working at Apple immediately agreed that 3Com's first batch of Ethernet products were placed in the market for Apple machine. This Ethernet device named Apple Boxes is a clumsy chassis that connects to Appleii parallel port, which ends in the market with failure. The original IBM PC has also announced the initial IBM PC at the time, but not working with 3COM, because IBM is busy inventing its own token ring network. However, 3COM decided to advance its own plan without IBM cooperation, began to develop EtherLink ISA adapters. After 18 months, on September 29, 1982, the first Etherlink was put on the market and randomly configure the corresponding DOS driver software. The first Etherlink has a technical breakthrough in many ways: EtherLink network interface card can be achieved by silicon semiconductor integrated processes. In 1983, 3COM became a partner of the new SEEQ technology company. SEEQ promise that a silicon film can contain most discrete controller functions in its VLSI technology, thereby reducing the number of components on the printed board and its cost, and leave enough space to make the transceiver can be assembled in a piece of print. Plate. In 1982, Etherlink became the first network interface card (NIC) of a Ethernet VLSI controller silicon wafer - SEEQ8001. More importantly, Etherlink became the first Ethernet ISA bus adapter of IBM PC, which is a milestone in the history of Ethernet. Due to the low price of the SEEQ silicon, 3COM can sell Etherlink at $ 950, which is much cheaper than other cards and previously sold. Before the EtherLink adapter is launched, all Ethernet devices are characterized by an external MAU transceiver, which connects it on the whate cable of the Ethernet. Since the large amount of space is saved with a large scale integrated circuit chip, the transceiver can be integrated on the plugin card. Since there are various disadvantages of traditional coarse coaxial cables, 3COM companies also adopt new cable wiring methods. This basic idea of this name is the basic idea of Etherlink designer Ron Crane, and soon becomes in fact standard. This cable Ethernet has many advantages: do not require additional transceivers and transceiver cables, cheaper, because the thin-shaft cable is easy to install and use, making the network more friendly with users. Metcalfe decided to make 3COM companies as a goal of IBM PC. At that time, IBM Design IBM PC would like to use the machine mainly to make a home computer; however, it is a large number of purchases of the PC, not family users.
In 1982, the demand for PC has exceeded the predicted value. IBM sold 200,000 PCs in one month, which is more than doubled than the original forecast of the company, making IBM's factory overtime, using a year to produce for two years The output can be completed to meet market demand. In early 1981, IBM XT was listed. At this time, IBM has 75% of the PC commercial market, but unfortunately IBM didn't recognize their personal computer at the time. By 1983, Etherlink's business is hot, and the 3COM stock began to go public in 1984. In the same year, ICL (International Computer Co., Ltd.), HP submitted the concept of the cable Ethernet to the IEEE, and IEEE admitted that it is official standard with L0Base2. Since the distance to the node is shortened to 200 meters, the standard is called 10Base2; and there is also because it uses a cheaper thin-shaft cable, it is also known as the cheapernet. Starlan: Great thinking, but the speed is not good (1984--1987) The cable Ethernet is excellent in most aspects, and the cable Ethernet is used to replace expensive yellow with cheap flexible thin-axis cable. Coaxial cable. In addition, most of the cable Ethernet network interface card (NIC) has an internal transceiver that makes it easy to install and reduce costs. However, the cable Ethernet still has some main disadvantages, such as the same shaft cable to break due to certain accidents of accidents or users (this kind of matter often occurs), will make the entire network. In addition, it is required to be correctly terminated on both ends of the network, and network reconstruction is a problem - if the user performs the movement of the entity, the network cable must re-wiring accordingly, which is often neither convenient, and easier . At the end of 1983, Bob Galin, which came from Intel, started working with AT & T and NCR to study the Ethernet on the unsaled twisted pair (UTP) telephone cable. NCR recommends using a total cable structure similar to cable Ethernet, while AT & T Phone is keen on a roof structure similar to the current telephone wiring structure. The advantages of UTP star configuration are multifaceted: easy to install, configure, manage, and find faults, and low cost; this star is a breakthrough because it allows a structured wiring system, which uses a single line alone Connect each node to a central hub, which is clearly an obvious advantage for installation, fault search, and reconfiguration, which can greatly reduce the cost of the entire network. In the early 1984, there were 14 companies participated in the research activities of UTP Ethernet. They have been discussed, mainly around how to make fast Ethernet can run on the UTP line. They confirmed that the low speed Ethernet (L-2Mbps) can run on Category3 and meet the limitations of electromagnetic interference requirements and crosstalk. However, some dealers strongly oppose 10% of the speed dropping speed to conventional Ethernet speed, so that many people lose interest, including the two leaders of the Ethernet 3COM and DEC, and other participants believe 1Mbps is fast enough to configure the IBM PC and the XT machine. After-all-fierce technical discussions, the Group's vote passed the Ethernet to 1 Mbps. 10 companies decided to perform LMBPS Ethernet and discuss with IEEE. The IEEE 802 team entrusts standardization of Starlan Task Force headed by Galin. In 1956, 1BASE5 as IEEE802.3 was approved (Starian supported from the hub to the node to a distance of 250 meters, and 5 in 1BASE5 represents a distance of 500 meters from the node to the node).
STARLAN Tuning In 1984, the STARLAN hub network interface card in 1984 pushed the Starlan Hub Network Interface Card to the Market. In the 1980s, Starian completed millions of connections, but many dealers, including 3COM and DBC, have already identified that 1Mbps is too slow - traditions that have been doubled in the computer industry, some customers and The dealer regards LMBFS Ethernet as a backward behavior. (In 1984 IBM has announced the PC AT based on Intel80286 microprocessor, two years later, this year, Intel launched an 80386 microprocessor, this 32-bit CPU is more than its previous generation. 80286 strongly many times.) Therefore, Starlan is no longer possible to get from the industry and market support. Finally, in 1987, the company launched the Lattisnet and submitted a product of 10Mbps Ethernet performance on conventional telephone lines. Soon, the Lat Tisnet was standardized by IEEE in accordance with twisted pair Ethernet, which was named 10BASE-T, so that Starlan and Galin's death is still countless, but as unscadless buttons and star line Ethernet The pioneer, its achievements are unsteady. 10Base-T and Structured Wiring History (1986--1990) In the mid-1980s, the PC revolutionary wave was unstoppable. In 1986, the personal computer was boosted under the driving of the application. Lotus L-2-3 has become a powerful opponent for IBM PC AT - each business is less. Apple's Macintosh takes off from its non-parallel graphical user interface in 1986. At the same time, people want to share expensive laser printers to print their spreadsheets and desktop printing publications, and the online sales are also very popular. There are two major events that make Ethernet once again set off a climax: First, 1985 Novell began to submit NetWork, this is a high-performance operating system for IBM compatible personal computer network, and the other is 10Base-T, one can be twisted without shielded twisted Ethernet running full speed 10Mbps on the telephone line. The first Ethernet of the Optical Cable Ethernet and UTP Ethernet uses a coarse coaxial cable. After a few years, Metcalfe and Eric Rawson prove that the CSMA signal can run on the cable. In the early 1980s, the cable caused a sensation. Xerox decides to run Ethernet on the cable. Er IC Rawson was appointed the person in charge of the cable Ethernet project, and the Ron Schmidt also participated in. Rawson and Schmidt find that Ethernet can indeed run on the cable, but can only be star structure, not a typical Ethernet bus topology. In 1985, Schmidt changed the cable Ethernet hardware to run on shielded twisted pair (STP), however, since the price of STP cables is expensive and cumbersome; so he has done some experiments again, prove that Ethernet can be regular No masked twisted pair (UTP) running. Structured wiring: Starlan and Token Ring In 1985, IBM started to launch its 4-mbps token Ring Lan - this has been 6 years later in Metcalfe to discuss the Ethernet adapter used by IBM PC, is also almost the first Three years after ISA Etherlink listed. Although the Token Ring is almost half a 10Mbps Ethernet, it has a major advantage than Ethernet - it is based on structured wiring system, which shields the central concentrator or hub. (STP) is connected to the node.
By 1986, STARLAN, which was used in structured wiring systems, also started, but unfortunately the speed of Starlan only 10% of the Ethernet speed, that is, LMBPS, and thus unable to replace conventional 10Mbps Ethernet or 4Mbps token ring network. However, the emergence of Starlan and token rings makes the future use twisted pair wiring and centralized wiring hubs. The birth of the Synoptics Communications returned to 1983, Schmidt has begun to seek a business unit within Xerox to produce a cable Ethernet, although he has not been given, but it has a big harvest on the other hand - he found a in Xerox. The business plan named Andy Lud Wig, two people very speculative, after repeated negotiations with Xerox, by the summer of 1985 finally set up their own company, and Xerox is just a small shareholder of their company. After obtaining a risk fund, Schmidt and Ludwig led 8 Xerox employees from Xerox, starting from Xerox, in 1985, starting with Astra Communications. The company's goal is to sell structured wiring cables and STP Ethernet hubs. (Astra has not last last, because NEC has been registered for the name, and threatens to accuse Schmidt's new company violation of the trademark law. A chairman of the new company accidentally found the word synoptis when browsing the dictionary, Synoptics communication The company's name is born.) 10Base-T approved the IEEE Standard In 1986, Synoptics started research on 10Mbps Ethernet on the UTP telephone line. The first Synoptics product named Lattis Net officially puts the market on August 17, 1987. On the same day, IE rose IEEE 802.3 Working Group to discuss the best way to implement 10Mbps Ethernet on UTP, and later was named 10BASE-T. In addition to the Synoptics Lattisent program, many competitive proposals have also flied to IEEE, the most famous of which is 3COM / DEC and HP proposals. After three years, engineers around the world periodically gathered in the IEEE802.3, and explored the best solutions to run 10Mbps Ethernet on UTP. Finally, IEEE agrees to standardize the 1990 autumn in the autumn of HP multi-port repeater and improved Synoptics Lattisnet technology, and the new 802.3i / 10Base-T standard is officially passed. The sales volume of Ethernet of the second year is approaching, and its attractive is the new 10 base-T repeater, twisted pair dielectric accessory (MAU) and NIC network interface card. The appearance of the star wiring structure is the great milestone in the history of Ethernet development. First, the Ethernet begins the increasing thing, and there is a private switch in the wiring room and a dedicated line to each node. Second, IBM Token Ring has lost its two advantages - structured wiring methods and use twisted pair wiring. Novell NetWare: Network "Authoritative" in early 1980, a small software company called Novell has developed a network operating system named NetWare. However, if you want to run the NetWare operating system, you need to purchase the related network hardware, Novell seizes this opportunity, start selling network interface cards, and the interface card is only available for Ethernet. Almost a day night, NetWare became the "authority" of office applications, which allows individual PC access shared printers, send electronic letters, exchange files, and access central databases. The huge success of NetWare has driven a lot of demand for Ethernet adapters, making Ethernet a leader in the network market.
In the future, NetWare is modified to be available for ArcNet and Token Ring. At this point, Ethernet overwhelves all other LAN technology. In 1980, Novell sells its network interface card business, but at the same time, it starts to sell licenses to any people designed to design it, and there is a huge NE2000 imitation business. Just like the IBM PC imitation industry is booming. Suddenly, some foreign lines can enter the network interface card business by purchasing the NS Semiconductor's chip and get name, design, and corresponding software from NovCLL. Some companies even don't even care about the design license, as long as they purchased 8390 wafers from NS and submit without software, it is marvelous with "NE2000 Compatibility". Due to the fierce competition between NE2000 imitation and 3CO, the price has fallen sharply. By 1988, there were dozens of NE2000 card manufacturers in business. As long as they spend $ 200, they can purchase Ethernet network interface cards, while IBM's Token Ring card has more than 1,000 dollars. By 1990, Western Digitai (later replaced to SMC) and some NE2000 dealers, including some Taiwan companies, even began to imitate nationa1 8390 Ethernet chip. The result is that the Ethernet business is rising due to continuous reduction, and the distributors that can be supplied are more increasing, and the technological innovation is constantly emerging due to competitive pressure. The demise of Token Ring was until 1992, many industrial and market research analysts mistakenly predicted token Ring to overwhelm the Ethernet, but the fact is Synoptics in December 1987 to Texas Instruments (TI) and Boeing Airlines. When submitting its first Ethernet hub, give the future of Token Ring with a fatal blow. At the end of the 1980s, even IBM didn't dare to see the existence of Ethernet: An example is IBM to sell microchannel Ethernet adapters for its personal computers and AS / 400 mini machines to meet its customers for ten years. On the request provided, another example is IBM to introduce RS / 6O00 engineering technology workstation product line in order to adapt to the world's trend. The leader of the workstation - SUN microsystems have long adopted Ethernet as standards, and the Ethernet standard component is built into its machine. In 1991, IBM Austin, Texas Workstation Division took out a sparet-end-adopted Ethernet as its standard. From this RS / 6000, the built-in INTE132-bit Ethernet association processor is started. Only in the 1990s, the IBM network segment began to realize that Token Ring is unable to replace Ethernet, its network strategy has failed. At this time, the ratio of the sales volume of Ethernet and the sales volume of Token Ring is 3: 1, and this overpower is still accelerating. At the last minute of the online stage, IBM still tries to copy the Novell Ne2000 / NationA18390 imitation strategy to stop the market share of Ethernet. In 1992, IBM transfers its license of the Token Ring to the National Semiconductor Company (NS) to produce a large and economical imitation industry, but by 1992, the ending of Tocken Ring completely failed has not been saved. At this time, Ethernet has become a factual standard for local area networks in the world, medium and small companies. In the past four years, the sales volume of Ethernet increased to ten times, from 1 million in 1988 to M1 million in 1992. This is a very eye-catching growth speed! Exchange and full-duplex Ethernet The appearance (1990--1994) In the late 1980s, several market factors caused the need for fast network infrastructure: independent PCs connected to existing networks, resulting in higher information traffic.
There are more and more new PCs in the market, and it also leads to higher information traffic. With the graphical user interface, the powerful PC has more and more, causing a higher graphic-based network load. Multiple LANs are connected together, sharing Ethernet depends on a single connection of all users shared media, in which only one website at any location is available at a single connection. Due to the limited part of the same bandwidth, these different local area networks are connected together, resulting in the smart rate of information traffic. The two-port bridge (only two LANs only connected) has a history as long as the Ethernet itself. At this time it is a best-selling goods on the connection LAN, because it enables information traffic controlled. By the end of the 1980s, a new network bridge-intelligent multi-port bridge began to appear. Alantec, SynerNetics, Raca1-Milgo, Clear Point, etc. have sold intelligent multi-port bridges, but in 1990, a completely different bridge - Kalpana etherswitch EPS-700 is available. EtherSwitch is very different from the vast majority of the bridges at the time: the network switch (Switch) is a system structure that provides both multiple data transmission paths, and the telephone switch is very similar to the overall throughput. Significantly increased. EtherSwitch uses a new bridge technology called "cut-through" (Cut-THROUGH) (and regular bridges are used in storage, forwarding technologies). This network switch reduces the delay time to reduce a quantity. It eventually makes Kalpana a brand-name, the company's founder Vinod Bhardwaj and Larry Blair have a non-fond of distribution methods to sell the product. EtherSwitch is sold in a network switch that improves LAN performance, rather than selling bridges with different LANs. This difference is very microseconds, but almost a day night, EtherSwitch has opened up a new market area - network switch (or network switch). The full duplex Ethernet was created another breakthrough in 1993, Kalpana created another breakthrough. Conventional shared media Ethernet works only in half-duplex mode, and the network either sends data at the same time or receives data without simultaneously transmitting and receiving data. For all users, sharing Ethernet relying on a single shared medium, so it is not possible to send and receive simultaneously. The advantages of the full duplex are obvious - simultaneously send and receive, which in theory can turn the transmission speed. Kalpana adds this feature to its hub, the full duplex is likely to be the standard of future exchange hubs and server network interface cards. Fast Ethernet appearance (1992-1995) Although the network switch is a best device for reducing network communication, each Ethernet switch can only provide 10Mbps maximum flow through each port. For applications requiring 10Mbps or more, the only qualified competitor is once a fiber optic cable distributed data interface (FDDI), which is an expensive 100Mbps-based LAN. Managers of large networks have been connected to the FDDI main network and FDDI server, and in some cases, they even connect the client or workstation to the FDDI ring. In the 1980s, brand-name companies such as DEC, AMD, NS, and IBM put millions of dollars in FDDI semiconductor mills and product development. In 1991, Sun Microsystems even intended to connect FDDI to each-stage Sparc Station (all Suns ParcStation have built-in 10Mbps Ethernet connection) E unfortunately due to high cost and complex FDDI Not a mainstream technology. At the same time, the price of Ethernet has fallen sharply due to the surge in sales. Some networks began to build a swap hub containing high-speed FDDI and Ethernet ports. Crescendo Communication Company is one of them, which built a working group switch with FDDI and swap 10Base-T port.
However, many users are increasingly considering the long-term future of Ethernet, because this technology has been a 10-year history, which is recognized in the computer world. In August 1991, the general engineer of the inventor David Boggs, the first Etherlink network adapter of the inventor David Boggs, the first Etherlink network adapter from 3COM, the first Etherlink network adapter, the first Etherlink network adapter, the first EtherLink network adapter. Larry Birenbaum with Vice President, set up a new company. They intended to start from the sales network testing device, but in the discussion process, Birenbaum suddenly proposed whether the Ethernet is run 10 times according to the original speed, and Crane who has been studied immediately answered "not a problem", and Boggs also agreed. . The following history is to establish a Grand Junction network company on February 28, 1992, engaged in design, construction and distribution of high-speed Ethernet devices. The company's directors have Metcalfe, Chamey and Trinity Venturers. The company is located in Fremont, California, only 20 minutes away from Silicon Valley. After the founding of Grand Junction, the development of 100 Mbps Ethernet immediately puts it. In the second half of 1992, Grand Junction developed 100Mbps Ethernet's news, so the company decided to officially released the news of 100Mbps Ethernet in advance in September 1992. IEEE 802.3 100Mbps Standard Wars In 1992, IEEE802 Engineering Group held a plenary meeting, one of the agendas is high-speed Ethernet, and two technical programs are proposed by the Grand Junction Network. The program recommends retaining the current Ethernet protocol, which is recommended by 3COM, Sun Microsystems and Synoptics. The second program comes from HP, which is recommended to use a full new Mac (Media Access) protocol for 100Mbps transmission. This meeting marked the "Fast Ethernet War" kicked off. During 1993, the IEEE's high-speed research team continued its 100Mbps standardization. A variety of recommendations are brewing, but the main problems have not been solved - 802.3 groups are the new MAC program proposed by HP or the CSMA / CDMAC protocol to retain the original Ethernet. Most group members have retained original. Net interested, but the end reaches 75% of the 75% of the standards required, so that the debate is continuing. Grand Juncttion, Intel, Lan Media, Synoptics, Cabletron, National Semiconductor Company (NS), Standard Micro Systems (SMC), Sun Microsystems and 3COM Soon Soon the politics of this never-ending debate and IEEE standard The obstacle is bored, so, in order to speed up the standardization, the partnership is also stoveted, and the "Fast Ethernet Alliance" has been established, pulling up the banner of "100Mbps Ethernet program based on the original Ethernet standard", so that HP and AT & T are in the industry Falling into the situation of four Chu songs. Unfortunately, HP and AT & T still adhere to their priority request agreement is the best solution, refuse to retain CSMA / CD Ethernet protocols. In order to break the deadlock, IEEE is appointed a new working group called 802.12 for priority request access. The following events were then taken: October 1993, the Fea (Fast Ethernet Alliance) announced its 100BASE-X interoperability specification, and now called 100BaseitX. In the same month, Grandjunction launched the world's first fast Ethernet hub and NIC - Fastswitch 2.0 / 100 and Fastnic100.
In May 1994, INTEI and Synoptics announced and showed their fast Ethernet device. In the 1994 half years, IEEE802.3 is busy with other parts of 100Mbps Ethernet standards such as 100BASE-T4, MII, repeater and full-duplex. In the same year, FEA's members increased to 60, many companies that support new 802.12 technology at the beginning of the beginning of the company also poured, or both technologies were supported. At the end of 1994, INTE1, Sun Microsystems and NetWorth have begun to launch 100BASE-TX-type compatible products. In the first quarter of 1995, Cogent, 3COM, DEC, SMC, ACCTION, Synoptics / BayNetworks, etc. also launched their own 100BASE-TX products. March 1995. The IEEE802.3u specification is adopted by its members and committees. So fast Ethernet is announced! After a few months, the Fea (Fast Ethernet Alliance) disbanded, and its work also ended, and its standard has been completed. At this time, it is a pre-sale of the Ethernet product. Until 1995, each manufacturer's revival continued to launch a new fast Ethernet product, and quickly Ethernet reached its top age. Industrial Trends (1995) 1995 was the year of the industry and jointly provided complete sets of products. At the end of 1995, 3COM, BayNetworks and Cisco have become a real network company that provides complete sets of products, providing all products from routers, chassis hubs, and workgroup hubs to network management and remote access to all products. Another trend is that many manufacturers realize that ATM is a revolutionary technology, which is difficult to immediately replace the local area network shared by the media. Users and manufacturers have been more committed to some revolutionary technologies such as Ethernet Switching and 100Mbps Fast Ethernet. In 1995, it is committed to the banner of bankers who are committed to high-tech undertakings. In 1994, Wellfleet and Synoptics were combined into BayNetworks, thus set off a wave of industrial mergers. BayNetworks continued to perform this extension strategy, in the spring of 1995, it announced the merged Centillion plan, which is a pioneer company engaged in token Ring Exchange Technology. In 1995, the largest high-tech merger was 3COM purchased Chipcom Company engaged in chassis hub, which obtained the advanced chassis hub technology and favorable map of IBM's OEM contract, this event is worth one It is because ChipCom is the only primary chassis hub company that can provide 100Base-T and 100VG-Anylan products. With this mergers of 3COM, the VG program disappeared. For us, this is the most interested in the year is Cisco Systems, Grandjunction Networks. For those in the industry, Grandjunction Networks is a fast Ethernet company that provides the first 10Mbps working group switch, and first provides various network interface cards, transponders, and exchanters of fast-type Ethernet. 3COM and BayNetworks are a member of the Fast Ethernet Alliance from the beginning, and Cisco will participate in the fast Ethernet business, and thus, whether in exchange-type Ethernet, FDDI, 100VG or 100BASE-T, or waiting for ATM aspects. I can't clarify my own direction. At the beginning of 100 base-T, the lack of Cisco's router is the main obstacle to the success of fast Ethernet. With Cisco, in 1995, the pioneer Kalpana, which has been exchanged in 1995, has changed.
Grandjunction has special decisive significance for fast Ethernet, with this mergers, exchange and fast Ethernet clearly become the high-speed networking technology selected in the next few years. In 1995, some of the most interested incidents of the Internet include: Cabletron (it is Synoptics's opponent's opponent's opponent) without a flight of exchanging and fast Ethernet. By the end of 1995, Cabletron no longer watched, it purchased the SMC Ethernet switching section, which is the first time the mergers, eventually makes Cabletron into the exchange and fast-type Ethernet market. Another meaningful merger action is a company-leading company in ATM technology that matches Ethernet exchange company Alantec. For many years, ForeSystems has always advassed that do not use Ethernet, ingenuity, in ATM, this merger indicates that it has recognized: frame-based technologies (such as Ethernet and token rings) are still core technology in many years. The world's largest PC and server manufacturers Compaq computer companies enter the network market through two network companies that purchase NetWorth and Thomas-Conrad. In fast Ethernet adapter production business, Intel is already a leading position and has 40% of market share. Jozhao Ethernet (1999-?) In January 1995, the IEEE802.3 Standards Committee established a new "High-Speed Study Group" to study the Ethernet of 1 dry rate per second. . In March 1996, IEEE formed a new 802.3z working group, responsible for researching dry gigabit Ethernet, formulating corresponding standards. Soon, some fast Ethernet supporters and some new initiators have formed "GEA Ethernet Alliance (GEA)", including 11 companies, they are 3Com, BayNetworks, Cisco, Compaq, Granite Systems , Intel, LSI Logic, Packet Engines, Sun, Microsystem, Q, UB Networks, and VLSI Technology. One month later, 28 companies also joined the factory, including Hewlett-Packard (HP). The key to the Cashpin Ethernet is to use exchanging full-duplex operation to build a backbone network and a super server and workstation, which may also use the cable with existing technologies (such as cable channel front technology) (some people have already Gigabit Ethernet is called 1000BASE-F, indicating the transmission rate of 1000 mbps on the optical cable medium). Semi-duplex / forwarding LAN and copper core cables are also required to be 20 to 25 meters (some people advocate modifying the CSMA / CD access method to increase the network diameter). If the Gigabit Ethernet has succeeded, it is clear that the Ethernet frame will remain the standard of data communication in the future for at least 10 years. LAN Basic Knowledge ----- What is LAN -------------------------------------------------------------------------------------------------------------------------------------------------------------------- ---------------------------------------------- One. What is LANs must use two ways to complete the definition of the LAN: one is a functional definition, and the other is technical definition. The previous LAN is defined as a set of desktop computers and other devices, not far from each other, to allow users to communicate and share computing resources such as printers and storage devices, are interconnected together. This definition is suitable for LANs used in the office environment. For technical definitions of LAN, it is defined as a computer that is interconnected from a particular type of transport medium such as cable, fiber optic cable and wireless media and network adapter (also known as NIC), and is subject to network operating system system.
The difference between functionality and technical definitions is obvious. The functional definition emphasizes the external behavior and service; technical definition emphasizes the material foundation and configuration of the LAN. The name of the local area network (LAN) is implied in this network geographic area. Due to the limitations of smaller geographic ranges. Due to the smaller geographic range, the LAN is usually higher than the wide area network (WAN), for example, the current transmission rate of 10Mb / s, the transfer rate of FDD is 100Mb / s, while the main line rate of the WAN During the country is currently only 64kbps or 2.048Mbps, the end user's online speed is usually 14.4 kbps. The topology of the LAN is currently common to bus type and loop. This is determined by the limited geographic range. These two structures are rarely used in a wide area network environment. LAN also has a variety of features such as high reliability, easy expansion, and easy management and security. II. Network Operating System Network Operating System (NOS) is the heart and soul of the network. It is a special operating system that provides services to the network computer. It works under the computer operating system to increase the capacity required for the computer operating system. For example, when you talk about it, when you use a word handler on a LAN, your PC operating system behavior is like when the LAN is not constituted, this is the LAN operating system software to manage your access to the word handler. . The network operating system is running on a computer called a server, and is shared by a connected computer user, which is called a customer. NOS and single-user operating system or multi-user operating system running on a workstation are different from those of the provided service types. Under normal circumstances, NOS is to optimize network related features. Such as shared data files, software applications, and shared hard drives, printers, modems, scanners, and fax machines, etc. General computer operating systems, such as DOS and OS / 2, etc., the purpose is to optimize the interaction between users and systems and applications running on this operating system. To prevent access from more than one user from one or more users, the general network operating system has a file lock function. If there is no such function, it will not work properly. File lock features track each file in use and make sure it can only be edited once a user. The file can also be locked by the user to maintain the dedication of the dedicated file. NOS is also responsible for managing connections between LAN users and LAN printers. NOS always tracks each of the printers that can be used and print requests for each user, and manages how to meet these requests, allowing the user's operating system to feel that the desired printer is as directly connected to its computer. NOS also manages communication between each network device, which is implemented by media access methods in NOS. The various security features of NOS can be used to manage access rights to each user to ensure security confidentiality of critical data. Therefore, NOS is fundamentally a manager for managing connection, resources, and traffic flow. Third. The history of LAN is in the early stages of computer applications, and people use large and medium-sized computers, usually referred to as the host. People who need to use a computer must submit a request to the computer operator, and after receiving the machine, you must wait for a few hours or days to get the results. Later, with the development of electronic technology, through the terminal, it is connected to the host, and there is no need to enter the computer room, and only the request can be submitted from the office of the office. Later, a small and medium computer appeared, and the operating system also appeared. At this time, the user has been able to submit a request to the center machine in an interaction manner. However, the popularity of the computer is only achieved after the Personal Computer (PC) in the 1970s. The processing capabilities and storage capacity of IBM PCs in 1981 have been comparable to large machines in the morning. As the PC is put into the market, people find that each PC is configured to configure a disk drive and printer, which is really unbearable at the time. Then there is a way of resource sharing: disk servers and shared printers. This is a combination of hardware and software that makes several PC users to share access to the public hard drive. The first disk server is running under the CP / M operating system. Early LANs, users who share the hard drive are implemented by a computer connected to the shared drive. Software in the computer divides the public disk drive into the area called "Volume", one for each user. In the user's view, the "volume" of the user is like his own dedicated disk drive. Hard drives typically include public volumes, user-use. In the current LAN, the disk server has been replaced by the file server.
The file server is better than the disk server in terms of user-sharing files or helping users track their files. Some LANs can support multiple file servers, each of which has multiple hard drives, thereby making LANs easy. In addition to the hard drive for PC users, the second device shared by PC is a printer. Currently, each LAN can have this capability, and in most cases, the print server has become part of the entire LAN package, not a separate computer. With the LAN print server, users can only use printers connected to a certain file server, or use printers connected to any user workstation on the network. The LAN manager can limit access to certain printers. Users can also send several files to the same printer. These features and other features depends on the LAN software features used. Other types of servers have also appeared, such as communication servers, database servers, etc., will be described in the next topic. It is emphasized that the LAN is implemented by connecting a set of PCs to the machine specified as a server, and the connection media can have a variety of connecting media, such as shaft cables. IV. The basic composition of the LAN must form a LAN, and there must be its basic components. Since LAN is a computer network, it is naturally less than computer, especially personal computers (PCs). There is almost no network only consists of mainframes or small machines. Therefore, for LAN, the personal computer is an essential component. The computer is interconnected, and of course no transmission of the media, this medium can be coaxial cable, twisted pair, optical cable or radiological media. The third component is a network card that is usually not equipped with any independent computer, also known as a network adapter, but is an indispensable component when constituting the LAN. The fourth component is a variety of connecting devices such as DB-15 plug, RJ-45 plug seats such as DB-15 plug seats, RJ-45 plug seats. With the above four network components, it can be formed into a basic LAN hardware platform, as shown. With the LAN hardware environment, you need to control and manage the normal operation of LAN, that is, NOS is to increase the functionality required to increase the network on each PC machine. For example, when you need to use a word handler on a LAN, the user's feeling is like not forming a LAN, which is the management of the LAN operation to play the management of the word handler. In the case of the LAN, a copy of the word handler is typically saved in the file server and shared by any of the LANs. As can be seen from the above, it is known that the composition of the following five basic structures: 1 computer (especially PC); 2 transmission media; 3 network adapter; 4 network connection equipment; 5 network operating system. V. Network Topology Network Topology refers to the physical layout of various devices with transport media interconnects. There are several ways to participate in various equipment media involved in LAN work, in fact, only a few ways to fit the work of LAN. If a network is only connected to several devices, the easiest way is to connect them directly, and this connection is called a point-to-point connection. The network formed in this way is called a full interconnect network, as shown in Figure 1. There are 6 devices in the figure, and 15 transmission lines are required in full interconnection. If the device has n, the required line will reach n (n-1) / 2! Obviously, this approach is only possible to use the possibility that there is little geographical range and fewer devices. Even in this environment, it is not used in LAN technology. This topology is given here because it is possible to encounter interconnect technology such as a wide area network (WAN) when multiple LANs need to be interconnected by interconnect (such as routers). At present, there are three topologies used in most LANs: 1 Star Topology 2 Ring Topology 3 Bus Topology 1. Star Topology Star Structure is the oldest connection, everyone uses the phone every day It belongs to this structure, as shown in Figure 3. Among them, the star structure for the telephone network, for the current use of the most common Ethernet star structure, the network device in the center position is called hub, and the English is called HUB. The star line structure of the telephone network, which is conveniently controlled by HUB-centric structures, because communication between terminal users must pass through the center station. Because of this feature, it also brings advantages for easy maintenance and security. Due to the shutdown, the end user equipment does not affect the communication between other end users, but this structure is very disadvantageous, the central system must have extremely reliable, because the central system is damaged, the entire system tends to be paralyzed. .
