TSN Protocol Interpretation Series | (2) Qbv: The Road to Knowledge in the Workshop

At night in spring, the wind is clear and the moon is bright, and the cycle seems to have turned to the past again. An old man who didn't care about his length, he hadn't brushed his teeth yet, but he was already lying on the bed with his head tilted. The word Victory could be vaguely seen on the dimly lit screen beside the bed, and the screen turned off after a while.

This is exactly our hero today, Qbv.

▎I am a workshop worker

Qbv, Distinguished Professor of 802.1 University, likes to play games, asks for help, famous Gate Control (Gate Control) theory proposer, founder of the check-in system that data packets hate, honorary captain of the cafeteria window defense team, etc. Of course, he was also a workshop worker.

When he was young, Qbv worked in the factory workshop, but suddenly it was like an apple hitting his head. He watched all kinds of workpieces reciprocating on the conveyor belt, but with an empty head, he wondered how the workpieces were transported, processed, and assembled into precise The device is curious. And the spark between him and the dispatching thought is also bursting out from this.
Qbv's journey of seeking knowledge is not smooth sailing. Without a teacher, he can only grope in the dark, and the hardships need not be described too much.

As for the job-shop scheduling problem he studied, it originated in the 1950s and has been widely studied, including job shops, replacement flow shops, mixed flow shops, flexible job shop scheduling, and so on. Qbv, who understood the research, seized an opportunity to conduct a pilot in the workshop. The effect was good, and he became a workshop dispatcher as a matter of course, and he will be responsible for the improvement of the process in the future.

insert image description here

After becoming a workshop dispatcher, because of his good eloquence and wide range of friends, he quickly got along with the people in the workshop. He had nothing to do on weekdays, and he fell in love with playing games. What he plays most is simulation business games. He arranges priorities, divides working hours, or builds production lines. in it.

▎ But I want to look for some changes

The transformation of Qbv thinking started from the day when the TTE (Time-triggered Ethernet) team arrived.

The inspection team sent from the SAE Association was well-dressed, and Qbv communicated with them as a representative of the workshop. They talked many times about whether it is possible to determine the time when each workpiece enters the process, whether the process has a fault-tolerant mechanism, etc. Qbv Back to the words, yearning for their life from the bottom of my heart.

At the time when Ethernet was booming, TTE came to this survey to draw some inspiration from workshop scheduling to solve a problem that was a headache for Ethernet management at that time, that is, the bounded latency of traffic. )transmission. At the end of the research, Qbv approached TTE after thinking deeply, expressing their hope to participate in their research. But TTE had a good impression of this person who is familiar with workshop scheduling, and in this way, Qbv obtained the status of an assistant researcher.

Then came the long journey.

In the end, they integrated the clock synchronization solution AS6802 into the scheduling, and through time division multiplexing and reasonable time planning to ensure that critical traffic and non-critical traffic do not conflict with each other when they are transmitted on the same link. The dramatic thing is that Qbv's time planning experience when playing games really played a big role here.

This should be a happy ending, if the four characters of "proprietary technology" are not engraved in the bones of TTE. Manufacturers use thick barriers made of patents to isolate TTE technology, and Qbv doesn't like it.

▎I seem to have become a legend

The 802.1 University gave an opportunity. TSN, as the school committee, took the initiative to throw an olive branch to Qbv. In 2012, Qbv officially joined as a specially-appointed professor to lead the research of a more open bounded low-latency project. In the following four years, The gating mechanism is proposed, and three types of state machines are implemented to guarantee the gating mechanism, namely, the updated List Config state machine that manages the scheduling configuration, the Cycle Timer state machine that initializes the gating list, executes and maintains the gating cycle, and, according to A List Execute state machine that executes gate operations sequentially and maintains a certain time slot between gate operations.

insert image description here

Based on the synchronization time support provided by the 802.1AS laboratory, Qbv took the lead in carrying out pilot work in the campus, and the TSN school committee divided an entire campus for Qbv to use. Qbv has set up a check-in device at each node, and divides different time windows for them to swipe the card according to the 8 priorities of the data packets. The data packets in the Gate-closed time period are not allowed to pass, only to be switched Only when it becomes Gate-open can it pass through. The three types of state machines are running rumblingly in the laboratory, modifying the time period data of the punch card machine in real time.

insert image description here

The experience learned from TTE also played a big role in it. Qbv also uses two methods of solver and heuristic to solve the scheduling table, and sends it to the check-in configuration through the state machine. Compared with TTE, which strictly plans where each key data packet should be at each moment, the window mode adopted by Qbv makes the time planning of data packets more flexible.

Most of the time, things do fall into place.

▎Playing games? I'll bring you

Having experienced the embarrassment of studying alone, after entering the 802.1 University, Qbv often communicates with other scholars, although sometimes he talks nonsense about whether to play games or not, which makes him have a bad impression in the eyes of some scholars.

The earliest cooperation was in 2012, and it was carried out with the old man br of 802.3 University. They jointly established a Qbu laboratory. br This old man was sent by the TSN committee, saying that the Ethernet Bureau asked them to find someone to take it with them, and just do the whole thing [br: Where is the road]. It just so happened that Qbv had just arrived, so they asked them to work together, whoever imagined that they really did a lot of work.

It is said that they found that when the gate control is switched from close to open, there are always data packets that should have been completed in the previous time period running on the road. What Qbv mentioned is that there is a margin of time it takes to leave a data packet before the door is opened. Everyone should not continue to enter, and wait for the people on the road to finish running. br The old man started thinking about jumping in line again. In the end, the two of them combined together and proposed two guard band mechanisms.

insert image description here

After this cooperation, Qbv began to make fun of old man br for naming his crappy thesis, and old man br chased after Qbv with a stick, saying that he was not polite. Playing around, but it can be seen that Mr. br is not as depressed as before.

And Qbv really knows how to make friends, not only with br, but also with Qch, Qci and others, and even became the captain of the canteen defense team in the mouth of ordinary data packets with Qci. After all, after they came, one piped out and the other piped in in the cafeteria, which almost aroused public anger.

But in all fairness, this unserious professor who came all the way from the factory floor really added a little different luster to this serious campus.

Guess you like

Origin blog.csdn.net/Hirain1234/article/details/132662694