Chapter 1044 A brand new form of network?

For the technicians who can participate in the KJ-2000 project, they may feel at a loss when faced with a large amount of information at first. But now, Chang Haonan has divided the data types that need to be processed into two major categories, and naturally someone soon figured out the connection:

"Mr. Chang, for the first type of data you mentioned, which requires timeliness, perhaps the UDP protocol can be used in the Ethernet transport layer of the communication subsystem to achieve as much and as fast information transmission as possible. As for the second type, I think it can be improved on the basis of the TCP protocol to make it adaptable to mixed wired and wireless network media."

This proposal quickly received some support:

"I remember that a few years ago, some scholars proposed to set up a small relay station between wired and wireless transmissions. Its function is simply to copy packets in both directions between the connections. In this way, if a transmission timeout occurs on the first connection from the sender to the relay station, the sender will slow down the data transmission speed. If a transmission timeout occurs on the second connection from the relay station to the receiver, the retransmission will be accelerated."

"..."

However, some people have raised objections:

"But the UDP protocol is not reliable in principle. Is it really suitable for transmitting military information?"

"But there doesn't seem to be any better choice..."

"..."

For a moment, the entire conference room was plunged into a heated debate, and no one could convince the other group of people in a short period of time.

However, Chang Haonan did not join in immediately. Instead, he returned to his seat and leisurely drank two sips of water.

Given his current position, although he is not the person in charge of the KJ-2000 project, as long as he makes a statement, it will most likely become the final decision.

After all, both chief engineers Guo Lin and Liang Shaoxiu have close relationships with him.

Therefore, he is now gradually inclined to provide some guidance but not to express his opinion immediately.

At least give others enough space to communicate.

However, before Chang Haonan could speak again, Guo Lin, who had been invisible just now, stood up:

"I think UDP is not a problem, because the transport layer and control layer of the combat system of early warning aircraft and fighter jets are separated. It is entirely possible to use UDP protocol at the transport layer and use the return verification method at the application layer to ensure the reliability of data."

In any case, Guo Lin is the chief designer in charge of the KJ-2000's onboard equipment in the literal sense.

Therefore, the surroundings became quiet.

Guo Lin looked in the direction of Chang Haonan, and after confirming that the latter had no intention of adding anything, he continued:

"The problem still lies in TCP. How to optimize the performance of TCP for different network forms such as HF/VHF/UHF communication links and satellite communication links?"

"Just now, a comrade did mention an idea. I'm sorry I didn't notice who it was, but I guess he's not from our Electronics Technology Group..."

When he said this, there was a hint of smile in his tone;

"Because the communication method you mentioned does not conform to the semantics of TCP, it will be very complicated to implement at the engineering level..."

On the other side of the conference room, someone had a slightly embarrassed expression.

He was indeed an engineer from Gaofei Group and was not from the communications field. He just chose to speak up because he suddenly remembered a paper he had read in the past.

As a result, something funny happened.

Fortunately, this is something outside the professional field, so no one will pay too much attention to it.

Since Guo Lin's words had already served to bring the discussion to a close, Chang Haonan stopped remaining silent and stood up again and returned to the front of the meeting room:

"In fact... people have overestimated the difficulty of optimizing the TCP protocol."

He tapped the blackboard gently and said:

"The relay platform we use is certainly not the relay between wired and wireless as the comrade just mentioned, but the relay in the wireless communication process. In the future, it will be a satellite, and in the current test, it is a stratospheric airship. This platform has passed the performance test."

"Because it uses a brand new network coding technology, it can ensure extremely low bit error rates and delays even in complex electromagnetic environments. Although it cannot be said to be at the same level as wired networks, it is at least much more reliable than the past... or existing wireless transmission technologies."

These words confused many people present.

Weren't you the first person who suggested improving the TCP protocol?

Why are you now saying that we overestimated the difficulty of this matter?

Faced with more than a dozen confused faces, Chang Haonan did not keep them in suspense. Instead, he erased the small triangle and circle he had just drawn on the blackboard and added a new diagram.

While drawing, he explained:

"In fact, the idea of ​​setting up a relay point between wired and wireless transmission is not a problem."

"Ok?"

Guo Lin couldn't hold it in any longer.

He had just criticized someone with great confidence, and if he found out in the end that he was wrong, all the boomerangs would hit his head.

Chang Haonan paused for a moment, then added:

"Of course, Mr. Guo's subsequent criticism is not wrong."

"call--"

Guo Lin finally breathed a sigh of relief, and his blood pressure, which had almost exploded, also dropped.

"We can add a sniffing agent to sniff and cache TCP segments sent to mobile hosts, as well as the acknowledgments sent back."

Chang Haonan drew a reverse dotted arrow on the part representing the wired data segment:

"When the probing agent finds that a TCP segment stops before the timer expires, but no confirmation is received from there, or a duplicate confirmation is received, it means that there is a problem in the wired part and the original data segment needs to be discarded and resent..."

"And if the reply confirms that everything is normal, but the entire transmission still has a transmission failure, then it means that the problem lies in the wireless part..."

"The specific implementation strategy still needs some testing, but because the bit error rate is not high, the path MTU minus the TCP and IP header size can generally achieve the best performance..."

"..."

The improvement plan proposed by Chang Haonan is not complicated. It will only take a few weeks, or one to two months at most to complete. It is basically software-level, or some small-scale hardware-level measures, which can be easily operated in the cabin and will not even affect the assembly progress of the KJ-2000 prototype.

As for the audience below, they looked at each other in bewilderment.

Although no one said it directly, the expressions on their faces clearly revealed the same meaning -

Is that all?

When Guo Lin first discovered this problem, and when Chang Haonan brought them into this lounge just now, the solemn atmosphere made people think that the entire KJ-2000 project would be delayed because of this.

Is this the result?

And Chang Haonan naturally saw through everyone's psychology.

"So far, we have only solved the optimization of the TCP protocol for several traditional communication links, so it is indeed relatively simple."

He took a step forward and said slowly:

"But don't forget the informationized command system I just mentioned..."

This time, the confusion on everyone's faces not only did not decrease, but increased.

After a while, someone finally asked tentatively:

"But... doesn't that command system also rely on these communication links to issue commands?"

"not quite."

Chang Haonan waved his hand:

"All the scenarios we have envisioned so far have assumed a center for the entire communication system, which is the AWACS. Of course, it could also be a higher-level command center, the same principle."

"But in reality, for example, when troops are rapidly deployed and advancing on the battlefield, the communication center cannot move forward synchronously with the front-line units for security or tactical considerations. In this case, the mobile network with a center cannot meet the needs. Instead, a new communication link that can quickly, temporarily, and automatically form a network is needed..."

“Distributed communications?”

Soon, someone threw out the key words.

"Yes."

Chang Haonan affirmed:

"Thanks to our newly developed relay technology, wireless communications can also be forwarded in multiple hops with low loss and low latency. So we can take advantage of this and design a completely new form of communication network."

When he said "brand new", there were several obvious inhalations from below.

No one thought that Mr. Chang would actually go so far as to make a move like this——

So far, the cornerstones of human modern communication technology, such as network structure and communication protocols, are still built by developed countries.

For the sake of standardization and globalization, China also chose to adopt most of its contents.

After all, it's a wheel-building process.

It will probably take a few more years before we can challenge developed countries in this regard.

But obviously, Chang Haonan no longer wanted to wait.

Therefore, he plans to test the waters in the military field, which is not so keen on global integration.

Chang Haonan gave everyone a few seconds to recover from their shock.

Then, the real heavy stuff begins:

"In addition to multi-hop forwarding, this network should also have the following features."

“First, there is no center. All nodes have equal status in the system. It is a peer-to-peer network. Any node can join and leave the network at any time. Even if a node fails, it will not affect the operation of the entire network.”

"Second, it is self-organizing. The deployment or expansion of the network does not need to rely on any preset network facilities. The nodes coordinate their behaviors through layered protocols and distributed algorithms. After the nodes are turned on, they can quickly and automatically form an independent network."

"The third is dynamic topology. Considering the application background, network nodes may move anywhere and at any time, and may also be turned on and off at any time. These will change the topology of the network, but the network itself must ensure stable operation."

"Therefore, such a communication network is not only significantly different from ordinary cellular mobile communication networks and fixed communication networks in terms of architecture and network organization, but also requires corresponding adaptation of communication protocols."

He returned to his seat, took a deep breath, and looked around:

"This is the most important and most difficult task in the next TCP optimization!"