The richest man starts with a blind box

Chapter 298 Jiangnan Server

"Chairman, the second batch of 300 servers have all passed the verification. After 30 days of actual operation, the general failure rate is 17%, the serious failure rate is 4%, and the stability rate is 97.79%!" In the Jiangnan Network Center in Cao County, The person in charge here reports to Huang Hehui.

"Yes, you have worked hard!" Huang He nodded, and was very satisfied with the result.

Although the failure rate of 7% sounds a bit high, and there is a serious failure rate of 1.8%, which means that 100 out of every 2 servers have serious failures, and it will not be repaired in ten days and half a month , or even the kind that needs to be scrapped directly.

If the server purchased from outside had such a high failure rate, then Huang He would have been clamoring to return it.

But if these servers are all produced by Jiangnan Group itself, then it's a different matter.

Yes, this batch of 300 servers is actually Jiangnan Group's own products, and it is the second batch of servers produced. The first batch has been sent to the Jiangnan Network Center in Wenzhou. There are a total of 2 servers in one south and one north. Jiangnan Group's network service provides guarantee.

Because it is the first batch of products produced by the company, it is a product of crossing the river by feeling the stones, and the high failure rate is a matter of course. If you want to produce amazing quality as soon as it is produced, this is only the case in novels.

When any new product is just developed, it is accompanied by various faults. Only by constantly searching for faults during use, accumulating experience, improving production technology, and avoiding those places that are prone to problems can the quality of the server be gradually improved. .

In fact, the failure rate of the first batch was even higher, reaching more than 1%, so Huang He is already very satisfied with the failure rate of the second batch, and it only needs to keep improving.

At the same time, it must be noted that although the failure rate is very high, the stability rate is also high.

The stability rate of 97.79% is almost comparable to the stability rate of servers of international first-line brands, such as the HP servers purchased by Jiangnan Group.

So the question is, why the quality of these servers is so bad, and the stability rate is still amazing?

That's because the problem of a single server failure was considered at the beginning of the design of these servers, so a whole set of linkage system was designed for these servers. The 300 servers form a whole, and if any server fails, then this The tasks of one server will be quickly transferred to other servers.

At the same time, because these 300 servers are actually responsible for the work of only about 200 servers, even if dozens of servers are disconnected instantly, they can continue to operate without any impact.

The reason why there is still a 2% instability effect this time is because there was a large-scale regional power outage in one of them, and there were two local network failures. A fault occurred, resulting in data loss and network fluctuations.

All in all, these computing servers actually maintain 100% stable operation efficiency.

Of course, one thing must be admitted, compared with the world's first-rate server manufacturers... No, in fact, even compared with the world's second-rate server manufacturers, Jiangnan Group's servers are still very different.

First of all, the amount of data like Jiangnan Group, in fact, can be maintained normally by purchasing 50 HP servers, but Huang He has invested 600 servers in the two network centers in the north and the south, so 200 of them are In order to maintain stability, there are 400 servers that are actually working.

In other words, the efficiency of one HP server is equivalent to the efficiency of eight Jiangnan servers.

However, there is no way around this. After all, the chips used by Jiangnan servers are all mid-range chips designed by themselves. The computing core is two grades behind foreign servers. The design is still too immature, so the current level that these Jiangnan servers can achieve is already the highest level that Jiangnan Group can manufacture. If you want to catch up with the international high-end level, there is still a long way to go.

Of course, Jiangnan Group's server also has its own advantages, that is, the price is quite cheap.

For example, an HP server at this time requires a purchase price of 120 million US dollars, which is the cheapest among all high-end servers.

But what about the price of the Jiangnan server?
In terms of pricing, there is no public pricing because it has not been sold to the outside world yet, but the estimated price of Huanghe is 50 yuan per unit.

Yes, you can buy an enterprise-level server for only 50 RMB. If converted into US dollars, it only costs a little over 6 US dollars, which is equivalent to 1/20 of the HP server.

The price of one HP server can buy 20 Jiangnan servers, and the calculation effect of 20 Jiangnan service areas is equivalent to the efficiency of 2.5 HP servers. Such a cost account, everyone can still calculate it clearly.

But this is Huang He's external selling price, so what about the actual production cost?

If the cost of purchasing equipment is not considered, then the production cost of a Jiangnan server is only about 20 yuan, of which 10 yuan is used to purchase various parts that Jiangnan Group cannot produce, and the remaining 10 yuan is Almost all of them are used to purchase chips and labor costs.

However, the chip cost is a game of left-handed and right-handed, because Nikon's ten lithography machines have all been in place, and now all the chips used by the Jiangnan server can be self-sufficient.

For example, in the second batch of Jiangnan servers for experiments, 2 of them used chips produced by Jiangnan Group itself.After that, Jiangnan Group will completely stop purchasing chips from outside, and the server groups will all use chips developed and produced by themselves.

At that time, the production cost of Jiangnan server can be further reduced, and a profit of more than 30 yuan can be made by selling a server. This is indeed a very amazing industry.

But considering that HP bought it for $120 million, you can imagine how much money they made.

And if Brother Xiao Ma can come to this base to visit, then most of the doubts in his heart can be answered.Because the Jiangnan Group has invested a large number of servers at a relatively low cost to maintain the operation of Facebook.

And if Tenda wants to maintain a server group of the same scale and computing power, it will probably cost about 2 million US dollars, and Tenda does not have that much money.

So this is a fundamental gap in strength. Jiangnan and Tenda are no longer companies in the same dimension. Tenda is still just a pure network company, but Jiangnan has evolved to the point of network operator + network hardware manufacturer. Servers can be produced, which is a problem that Tenda cannot solve at all.

It is also worth mentioning that although the computing servers of the Jiangnan Group are at the third-rate level in the world, the storage servers of the Jiangnan Group have reached the top international level.

This is mainly due to Tong He, and his graphene.

The black crystal that made the R&D center buzz before is actually just the second graphene material with high practical value developed by Tonghe. The first graphene material with high practical value before is actually a kind of storage particle.

The storage particles made of this special material can store 20 times the data of the best storage particles in the world at the same time under the same size.

That's right, it's 5 times the data, so even if the design of other aspects of the storage server is far behind the international advanced level, the storage server produced by Jiangnan Group can reach the best storage server in the world. double the effect.

Whether it is capacity, read/write speed, or volume, there is a gap of more than 5 times, which is roughly equivalent to the level of storage servers in the future around 2008.

At this time, there are about 50 storage servers incorporated into the server group, which can store hundreds of millions of terabytes of data. For at least two years, Jiangnan Group has no need for storage server expansion.

At the same time, Jiangnan Server is also planning to use the storage server as their first product in the international market. After all, the computing server at this time is really not available.

So let's get down to business, as for another question in my mind, it is the problem of the Internet.

One is because Jiangnan has set up two network centers in the north and the south, which can share all network speed requirements and improve network efficiency.

But this is not the main reason, the main reason is because of Facebook's own design optimization.

In fact, it only took more than two months for Facebook from research and development to the birth of the first version, and the next four months were all used to continue to polish Facebook.

Because in the process of actual testing, the Jiangnan team found that the network that Facebook needs to consume is too huge.Therefore, after 4 months of optimization, Facebook has stored almost all of the pictures in the offline center in the part of the most network-consuming space decoration.

In this way, apart from opening the space for the first time, every time you open the web page afterwards, you can directly call the image data from the offline center, and you don’t need to connect to the network to obtain the data at all.

The second is the various news circles carefully designed by Facebook. In the news circles, the recommended space dynamics will be directly published on the circle page. You need to click on the space to read directly, which also saves a lot of calling space For the data step, only a part of the text data or picture data is needed, which naturally greatly saves the network speed.

In addition, there are dozens of carefully polished designs that can save traffic in details. These are the results of 4 months of research and development.

As for Tenda, there are only 15 days in total, so the QQ space is basically based on the OO space, which was directly sentenced to death as soon as it was born, and copied without any polished samples.

Although the code has been changed, the functional design is essentially the same, so if you want to see the dynamics of Qzone, you must first enter Qzone, read a large amount of spatial data from the server, and then read the dynamic data , and finally the content of the space can be presented.

These steps greatly consume the speed of the network, which naturally leads to the failure to get up.

In fact, Brother Xiao Ma also discovered the problems later, but all these problems need time to polish, but Brother Xiao Ma has no time.

(End of this chapter)