It can be said that this situation of ice and fire is a reflection of the cognitive differences between theory and practice, academia and business, on the principle of LAN technology implementation.

Logically, Ethernet adopts a bus-based topology and uses CarrierSenseMultipleAccess/CollisionDetection – CSMA/CD, the bus contention technique for carrier listening multiple access/collision detection.

The process of sending and receiving between the various nodes on the network, i.e., the computer, generates a transmission unit, that is, the concept of data packets - frames.

CarrierSense - CS stands for Carrier Listening means that any computer connected to the network must listen to the network before sending a frame, and when it is confirmed that it is free, it can only be sent.

MultipleAccess means that multiple computers can access the network at the same time, and the frames sent by one computer can also be received by multiple computers, which is similar to broadcasting.

Obviously, this way of using the network will inevitably produce the phenomenon of frame collision.

Just like a two-way two-lane road, motor vehicles, non-motor vehicles, people, pets, etc., all compete to get on the road, and accidents such as scraping will inevitably occur.

There is no cure for collisions, but they can be mitigated by certain listening/sending strategies, one of which is CollisionDetection, which stands for Collision Detection.

It requires the computer to listen to the network at the same time as the frame is being sent to determine if a collision has occurred.

If a collision is detected during the sending of data. The following collision processing operations are performed:

First of all. Send a special blocking message. The special blocking information is a continuous byte of all-1 signals, which is intended to strengthen the collision so that other computers can detect the collision as soon as possible.

The image point is to break the jar and break it, so that everyone can see the traffic congestion and don't go on the road.

Second, wait for a random amount of time within a fixed period of time and send it again.

At last. If the collision is still there, the "truncated binary exponential avoidance algorithm" is used to send. That is, within 10 times, stop the random retransmission within twice the time of the previous "fixed time", and after 10 times, stop the random retransmission within the previous "fixed time". After 16 attempts, the teleport is abandoned if it still fails.

This method is like a traffic congestion metropolis, and the relevant departments arrange odd and even numbers according to the license plate number.

Obviously, the collision in the operation of Ethernet will cause a certain degree of waste of resources, and there is a lack of advantages in execution efficiency.

In stark contrast. A 4M Token Ring network is comparable to a 10M Ethernet data transfer rate. The data transfer rate of a 16M Token Ring network is close to that of a 100M Ethernet.

The reason for such a stark gap is, of course, that the basic principles of operation are very different.

The Token Ring network uses tokens, which represent signals, to avoid conflicts in the network.

The computer that gets the token is like a police car with an alarm in its lane, and it is natural to increase the data transmission rate of the network compared to the operation mechanism of each computer on Ethernet that constantly probes to find opportunities to add congestion.

In addition, the priority of transmission can be set through tokens to meet the needs of high-level network resources.

Theoretically, the Token Ring is invulnerable, but in practice, the Token Ring is underutilized due to the non-reusability of the network.

When a computer in the network gets a token and starts using the network, no matter how much bandwidth the computer uses, even if it only uses 1M out of 4M, the other computers must wait for it to use the network and give up the token before they have the opportunity to apply for a token and use the network.

It's like driving a privileged car on a road without stopping, even if there are eight lanes, other ordinary vehicles can only stare dryly and can't get on the road.

Especially when the network becomes complex, the programs running in each computer are written by people, and the desire to monopolize the network is faintly carried, so that the theoretical circulation of the token can not be guaranteed as expected, and correspondingly, the utilization rate of network resources is not high.

Overall, the Token Ring, which is similar to the planned economy, is not as dynamic as Ethernet, which is similar to the market economy.

The benefits of spontaneous coordinated use of the network by computers over Ethernet outweigh the losses caused by collisions.

However, the correct understanding of this point is far from widespread, and large companies such as the theorists and IBM are favoring the Token Ring.

Jerry Saltzer, a well-known computer scientist, pointed out in an influential paper co-authored by others that in theory, Token Ring is superior to Ethernet.

In 1964, before Ken Thompson and Dennis Ritchie developed UNIX based on the C language, Bell Labs, the Massachusetts Institute of Technology and the General Electric Company of the United States jointly participated in the development of a multi-tasking operating system installed on a mainframe - MultiplexedInformationandComputingService, or Multics, and Jerry Saltzer was one of the team leaders at that time.

As a result, Jerry Saltzer's views are influential, and many computer manufacturers are hesitant or simply decide not to make the Ethernet interface a standard feature on their machines.

However, the advantages of Ethernet with simple structure in the actual use process can not be ignored, in the original time and space, the neglected 3Com was a blessing in disguise, and made a lot of money by selling a single Ethernet network card, and quickly grew into an international large company, and then there was an almost joking saying - "Ethernet is not suitable for theoretical research, only suitable for practical application".

However, 3Com's glory will wait until a few years later, and now the company has been around for more than a year. I'm lobbying other manufacturers everywhere. Embrace your own technology.

Up to now. Robert Metcalfe has also made some progress, such as working with DEC, Intel, and Xerox to standardize Ethernet, launch a standard, and develop a commercial version for Unix.

And Tang Huan's goal is to come up with a personal computer version with 3Com and occupy the highest point in the industry first.

Because of Simeone's relationship. Tang Huan had already had some contact with Robert Metcalfe earlier, so the process of this meeting, various verbal temptations, can be directly omitted.

He was accompanied by two other engineers in their 30s, a man and a woman: David Boggs and Radia Perlman.

Since the end of last year, the database EBase has become very popular in the market, especially after the sales of Xiaobawang game consoles exploded, the attraction of Tang Huan's company to high-end talents has increased significantly, and there is basically no need for him, the boss, to bother digging people.

Originally, David Boggs and Radia Perlman both went to DEC this year. As a result, Zheru's human resources department, which was ordered to recruit network talents for the 386 plan, was recruited into the company, and this kind of scraping ability made Tang Huan amazed.

Usually. Ethernet is thought to have been invented in 1973 when Robert Metcalfe wrote a memo to the boss of PARC about the potential of Ethernet.

However, Robert Metcalfe himself believes that Ethernet only appeared a few years later. In 1976, Robert Metcalfe and his assistant David Boggs published an article entitled "Ethernet: Distributed Packet Switching Technology for Regional Computer Networks."

Robert Metcalfe left PARC to start 3COM, and David Boggs left PARC, but with his master's degree, he stayed on the West Coast to pursue a Ph.D. at Stanford University and then to Philosopher.

As for Radia Perlman, after graduating from the Massachusetts Institute of Technology with a master's degree in mathematics, she worked on software development for network devices. I don't know how Zheru's human resources department fooled her and brought her from the East Coast to the West Coast.

Don't underestimate this, American universities are divided into circles, and what kind of university a person chooses often means his future work level and even the scope of his mate.

In 1983, Radia Perlman invented SpanningTreeAlgorithms, the STP spanning tree protocol for bridges and switch devices, which was of great significance for building large networks.

Later, Radia Pollman was dubbed the "Mother of the Internet", but she is said to have liked the title.

Radia Pollman's work is actually to develop a basic method of data transmission for the Internet.

It is no exaggeration to say that the technical people on site are experts in networking.

However, Tang Huan didn't want to enter the technical discussion session immediately, he first asked about the possibility of investing in 3COM, because Robert Metcalfe attracted a lot of venture capital at the beginning of the company, so that he had no right to speak, and there were some things that he couldn't decide at all.

One of the manifestations of this is that the board of directors of 3COM recently brought in another CEO, and Robert Metcalfe was appointed chairman of the company's board of directors and vice president of marketing and sales at the same time, which is a balance of power.

In Tang Huan's view, such operations of capitalists have their own reasons, and entrepreneurs with professional and technical backgrounds may not be able to manage their own companies.

In the original time and space, Robert Metcalfe was finally swept away in 1990, similar to the fate of the two founders of Cisco, and the time difference was not far behind, and the two companies also entered a period of rapid growth after a short period of adjustment.

It's useless to complain too much, even if the truth is so cruel, the introspection that the creators need, at least Tang Huan thinks so.

Moreover, Robert Metcalfe is a very active person, playing the role of a prophet in the online world.

He famously proposed Metcalfe's law, which is comparable to Moore's Law: "The value of a network is proportional to the square of the number of users connected to the network." ”

Various social software has verified the correctness of this statement, and a big V account with thousands of followers on Weibo. How much it's worth. Obvious.

Of course. Robert Metcalfe's big mouth wasn't always right, and it drew ridicule for himself.

Robert Metcalfe, for example, predicted in 1995 that the Internet would be devastating in 1996, and in fact everyone was fine.

Robert Metcalfe also predicted the demise of wireless networks in the mid-1990s, but the reality is that users are looking for Wi-Fi signals everywhere.

Robert Metcalfe didn't stop there, and made another "social network bubble" remark. It seems that the companies involved are still making a lot of money.

It can be seen from this that the so-called forerunners, the mythical aura is imposed by others, at least Tang Huan feels that the other party is normal when dealing with Robert Metcalfe.

For example, he could see that Robert Metcalfe was happy to see himself invest in 3COM because he wanted to have an extra partner on the company's board of directors.

Still, Robert Metcalfe, who has no decision-making power, grimaced, said, "Don't look at me as chairman of the board." But those guys toyed with each other and discussed it twice. But it's all gone. ”

"Now the situation is different, we have officially launched the second generation of the Fangyuan computer, which can be called the first 16-bit personal computer with practical value, and the support of Ethernet technology is not enough to make those people interested?"

"Even then, the board of directors will not agree to sell 20 percent of the shares, otherwise the existing power structure will be changed. Robert Metcalfe frustrated his hands helplessly, "I suggest you lower your target and use the sales performance of the more powerful second-generation Fangyuan computer as evidence of persuasion." ”

Tang Huan rubbed his eyebrows and had to accept the retreat.

The mechanical product of hard disk, no matter how large its capacity becomes and how fast it is speed, cannot get rid of its position at the end of the industrial chain.

The network is different, those venture capital are all world-class in terms of intelligence, and naturally have a plan in their hearts, and they are unwilling to give up possible future benefits in the early stage of development.

Seeing that Tang Huan was so good, Robert Metcalfe breathed a sigh of relief and quipped, "You'd better take a formal trip to 3COM and let the members of the board of directors see you, a billionaire who has made a fortune in the personal computer industry, and deter some of these guys." ”

It seems that Robert Metcalfe, who is a little angry, wants to support him for the scene.

Tang Huan smiled, "Let's not mention this first, let's start to talk about the implementation of Ethernet on the second-generation Fangyuan computer, Zheru mainly does software work, your network card, and drivers, the sooner you take it out, the more you can let the industry know the advantages of Ethernet." ”

"To be honest, Zheru's enthusiasm for LAN is beyond my expectations, can you tell me any heavyweight applications you have?" asked Robert Metcalfe tentatively.

Tang Huan glanced at Simone appreciatively, not to mention that the other party had a good relationship with Robert Metcalfe, but he was unequivocal in keeping business secrets.

"We are working on a major upgrade to the database EBase, which is to support a client-server structure. Tang Huan lit up Zheru's dry goods.

"Can I take a look at the demo?" said Robert Metcalfe, who was immediately intrigued.

Client/Server Architecture - The C/S structure is also known as the Client-server model, which is a master-slave architecture, and the characteristics of this technology can be seen only in the literal sense.

The C/S structure is not new, having been studied at PARC in the 1970s, so Robert Metcalfe is no stranger.

"Absolutely, we've built the simulation platform. Tang Huan asked Simeone to prepare for the demonstration, and he continued to discuss Ethernet cooperation with Robert Metcalfe.

Glancing at David Boggs and Radia Perlman, Robert Metcalfe said with emotion, "Don, you have a long-term vision in the technical direction, and 3COM was founded only a few months later than Zheru, but it is much worse in terms of pattern. ”

"You're just scratching the surface, I'm walking on thin ice right now. Tang Huan shook his head and loosened his tie, "After IBM's personal computer is launched, I don't know how many companies and brands will be eliminated." ”

"You can rest assured, there is basically no possibility for 3COM to enter IBM's camp, we are friends in the trenches. Robert Metcalfe laughed.

……

After the successful development of the second-generation Fangyuan computer, the work pressure brought by it was not ordinary, and Tang Huan wanted to go to Xiangjiang to relax, but he couldn't spare time.

For example, Hawkins, who had just finished inspecting the game business in Japan, had just returned to the United States and came to report to Tang Huan with a message from Lianmei.

Not long after winning the Oscar for Best Picture and Best Director, Michael Cimino's epic western "Heaven's Gate", which cost $40 million, was released on November 19.

So far, the audience has complained that the box office is dismal, and it is expected that it will only be able to recover about 1.5 million.

Lianmei, whose family background has been ruined, is panicked from top to bottom.

Even those who have made it clear that they will be loyal to EA in the future are starting to get nervous. Such a money-losing product won't scare EA away, right? (To be continued......