[Repair infrastructure ASP. NET] repair infrastructure multi-dimensional architecture (multi tiers)

ArticleDirectory

• Minute:
• Actual score:
• Shard shard transactions:
• Real-time transaction:

Starting from demand

Let's take a small example to understand how the general method is different from the multi-region method.

Hypothetical: I need to show that the sales volume of all the employees in the last three months has become the ranking

 

General Practice:

Pull a gridview and a sqldatasouce on a plane, set the required layout method after sqldatasouce, and then display it.

In this way, from the [Information Display] [business conditions and settings] [access to data sources] are handled in one aspect.

In this way, we can say that all things can be connected in the program. However, if one day I want to change my interface to Windows form, WPF, Silverlight, and PDA, we must repeat them all.

Shard practices:

If we pull the interface out, for example, using components to handle [Commercial License Agreement configuration] and [access to materials ], then it becomes as follows:

 

At this time, some of the interfaces are just being displayed, and their resources come from the middle components, in the future, if you want to change the interface to another method (Windows form, WPF, Silverlight, and PDA), then the middle part does not need to be retained, you only need to repeat the interface and then combine it with the component.

 

Further segmentation (commodity pricing and Resource Access)

The infrastructure of the hypothetical system is smaller than the infrastructure, and the amount of data is not large. Therefore, access is used as the data source in the initial development stage, as the architecture grows, there are more resources, and we want to use the ms SQL Server to handle the resources, we can be very lucky, you do not need to change the interface configurations. However, there may be a lot of memory and a lot of data access methods in the middle, so we will need to significantly change the Middle-end programming. However, if we split the [business license] and [Resource Access activity, therefore, the temporary condition does not need to be changed. You only need to modify the part of the data access. The architecture will become as follows. Under such a structure, the future information can become more dynamic than the commercial license. You only need to modify the information access quota.

Therefore, in a multi-architecture, it is divided into three parts: interface indexes, business operation indexes, and resource access indexes, generally, it is a three-tier architecture (3 tier )]. So why is it more complicated ?? In a large system, it may be divided into different parts (for example: creation, management, storage, sales, logistics, service repair ,...) . Therefore, the middleware may be different in different programs based on different features. We set each part as a phrase, these operators are equivalent, but sometimes they need to cooperate with each other. At this time, there may be many architectures, such as: During the sales period, you must check your memory and store the data. Then, you must deliver the following bio-streams. Therefore, the businesses that have been divided into different regions are as follows, not only can the operation be established, but can also be used together with other businesses to form a multi-architecture.

 

In practice, there are several ways to improve the performance of Multiple Architectures:

The operation refers to the operation of the interface operator, business operator, and resource access operator on the same host, it is the shard in the upper limit.

Because all of them are running on the same master machine, there are many simple architectures and they can pass through the class) this article describes [commodity pricing] and [data access policies]

 

Actual score:

However, the real-world score may be different on different hosts, and even in business scenarios, there may be multiple hosts, handle their respective business regions separately (for example, manufacturing, management, storage, marketing, logistics, service repair ,... Each has its own host)

In terms of how to implement the transaction, the actual score is much higher than the actual score. Because the host machine is different, we have to cooperate with each other to handle [distributed transactions ]. Currently, Xiao Miao knows that this architecture can be realized through [COM +] and [WCF ].

 

Note the integrity of the transaction (transaction]

Normally, when the system is not split, when processing the data, because the program and the legal method are connected together, therefore, transaction processing is not difficult to handle.. net, or even in stored procedure. However, after segmentation, it is easy to ignore this part, resulting in a problem that occurs in half of the data records, there is no way to roll back all the data, resulting in frequent information.

Shard shard transactions:

Because the system operates on the same host, there is no problem of scattered transactions. Therefore, you can pass through transactionscope (please take a look at this article. [transactionscope) indicates the initial part 1 (note scope. location of the complete) to solve this problem.

Real-time transaction:

Meow said that the actual architecture can be implemented through [COM +] and [WCF. However, some may ask: does WebService support transaction? When WebService calls WebService, it is impossible to package two WebServices into a complete transaction (transaction). Therefore, it is impossible to use WebService in a single region.

 

Summary

the multi-region system architecture enables us to make our systems more efficient, and allows interfaces and materials to be more dynamic. In addition, different segments are split during the development process. You can also allow different people to share different regions, and multiple people can share the same development. Some of them attempt to deal with the design, while some users attempt to access their business, while others attempt to access their resources. In business models, we can also develop their own parts by developers of different systems, let each other inject their own domain knowhow to develop the commercial license component, and then end the operation. For the development of large systems, it is recommended that you use Multiple Architectures for development.