What are the characteristics of digital servers?

The existence of the database allows us to have better results when making marketing plans or acquiring new users. Let's take a look at the characteristics of digital servers in combination with IT training in Huo Ying.

Emphasize the Internet, because the premise of this paper is the application of the Internet. Different from "traditional" applications, applications in the Internet are faced with massive data and requests every day, which requires higher reliability and response speed of the system. For the "traditional" application, I think it is simply that the amount of data is not large, the range of users is relatively small, and naturally a large number of high concurrent request scenarios almost do not exist.

After having a general understanding of Internet applications and traditional applications, let's talk about the different ways of using thematic relational databases in these two types of applications, and whether relational data is no longer the focus of attention in today's Internet applications.

Massive data. Millions or even trillions of data can no longer be stored in a data table or even in a database. Imagine that if all the data are stored in a single database and a single table, once the whole table is scanned, it will be a disaster for the response speed of the system. However, in traditional applications, a single library and a single table may be enough.

Secondly, due to the large amount of data, the storage of data on the disk is designed as a "sub-database, sub-table" mode, and a specific "routing" algorithm is used to locate the location of a data. It is precisely because of the design of "sub-database, sub-table" that the scene of "linked table query" in relational data fails, so there is almost no "foreign key" in the design of a table in Internet application, that is, linked table query has almost disappeared.

Third, a large number of requests. This is very common in Internet applications. An emergency, a star's news, will cause a large number of requests to arrive instantly. The carrying capacity of the database is limited. Once all the accesses flood in at the same time at a certain moment, it will directly lead to the database downtime, and even the whole system will be unavailable because of the database. Therefore, in today's internet applications, reading and writing data almost no longer directly operates the database, but adds a "security" barrier-cache in front of the database.

Fourth, the reliability of service. The reliability of the service, even if there is a problem with the system, should also ensure that part of it is available. Separation of reading and writing is a good solution. Read and write operations are no longer carried out in the same database, but separately. If there are a lot of write operations at this time, try not to affect the read operation, or the database is down when writing to the database, and try not to affect the read operation of the database at this time. At this time, a set of "master-slave" database is usually deployed in Internet applications, and the main database writes and reads from the database, which will lead to the problem of data synchronization or be summarized as the problem of data consistency.