Cyq. Data V5 from getting started to abandoning the ORM series: the benefits of the framework

Objective:

After the framework open source, learning to use more and more people, so I also more actively in the code to respond.

After the framework is complete: the database reads and writes the detach function and the distributed cache function:

After three days of sleepless nights, I finally finished the third heavyweight function of the framework: automated distributed caching.

Source code has been submitted, the source address see: Finally wait for you: Cyq.data V5 series (ORM Data layer) The latest version of Open source

Remember many years ago, about 2010 years or so, Cyq. When the data framework enters the fast-updating version phase,

At that time I will often in some technical groups consciously or unconsciously speaking of their own framework, and then there are always some self-conceited people will say:

What are the advantages of your framework? Why should I use yours? You don't want to use the name of this frame.

may be not good at expressing ...

May be disdain to answer ...

There may be no merit.

Suddenly there is an egg of sadness ...

Later on the change of mentality, transfer fees, small public on the small, tough life, do not need to explain ...

Later in the years:

Entrepreneurial experience, learned from product thinking to make the framework of the use of simplification.

The technology of precipitation, thinking of the framework and the design of distributed thinking.

The process of the project has led to the enrichment and automation of the framework application scenarios.

Advantages of the framework:

If someone asks me the advantages of the framework and other ORM, I can only say, can you? Because it's not already on a level.

But since you sincerely asked, I will be merciful to say, ha:

5 years ago V4 Series version diagram:

But I do not want to consider these functions as the advantages of the framework, because I think the technology precipitated a certain years, as long as the effort, can be made in time.

Today's V5 series version map:

V5 a lot of features, here only focus on the five points, because the framework can integrate the five points, the current market only this one to do, there is no branch.

1: Multi-database Interoperability support:

V4 Multi-database support and other ORM support a number of libraries similar to support a variety of databases, usually only need to complete the interface can be;

V5 to do multi-database interoperability, the entire universe only this one.

To achieve multi-database interoperability, it is necessary to study various database data type differences, function differences, need to be abstract, need to be compared to the need for strategic thinking, a large number of scenes to collect.

And Dbimport is a polishing this function of the best scene, before and after a few years of time, before grinding to the current smooth stable stage.

Sometimes I think that dbimport overseas should have a market, after all, the world on this software (so cheap) can be done in such a variety of databases between the arbitrary interoperability.

2: Automation Framework Programming:

The ORM on the market is a single assignment, with a single row of data saved.

The evolutionary history of the framework: Single assignment = "Single UI interaction =" Bulk UI interaction = "bulk data interaction =" automated Interaction

For a related introduction, see: http://www.cnblogs.com/cyq1162/p/5634414.html

About the Development framework: (ORM as Data tier + turnkey solution)

The two development of the currently discovered development framework is stuck in the stage of generating a bunch of code with the code generator ...

It's hard to write code without a code generator, the project is tough, or the development cycle is infinitely long ....

I can only say: once I have also experienced, and later found that there is improved space.

The V5-based development framework: the ASP. NET Aries lets you discover an excellent automation development framework that does not require any code generators at all.

For a related introduction, see: http://www.cnblogs.com/cyq1162/p/5487796.html

3: Distributed Database (read/write separation)

The ORM on the market does not offer this feature:

One of the reasons: poor realization;

The second reason: data consistency does not know how to solve;

Reason three: think the data volume is big, the company has the architect, the architect will take the business aspect to solve this problem.

Due to the large environmental problems. NET market, there are few architects, and there is a lack of architects.

While the V5 framework not only accomplishes this function, it is important to:

The transition from a single database to a distributed database, at the code level, requires only a database link, and no other framework is currently available.

4: Distributed cache:

There are many many distributed cache components available, but the only downside is that you need to reference the components and modify the code.

But the V5 framework, from single-machine filtering to distributed, only needs to configure the server link, and there is no other framework to do so.

5: Distributed Automation cache:

And the 4th difference is, the 4th is the manual file, here is the automatic file.

Easy to understand, say something like NHibernate's level two cache.

It has been said that many ORM implementations cache the entity, and will make the cache rate a selling point.

But the reality is that the orm+ executes SQL, so the framework's level two cache is useless because the policy invalidation policy and the data timeliness problem cannot be handled at the SQL execution end.

V5 is not the same, because of the innate support for SQL statements, so the cache policy can be global analysis, coupled with the distributed cache enable, can make the automation cache immediately b into a distributed automation cache.

Review

Well, the advantages of the introduction here, has been invincible in the universe, and then b down, angry Green will not stand it.

A point of view to spit out the trough:

Most people think that from WebForm to MVC is a kind of progress, actually I think is a kind of retrogression.

From the modular development mode, return to the original manual file mode.

Let WebForm carry the sink and scold sound, the surface is ViewState and server control, the actual and cruel truth is: Most people write bad data layer .

If everyone goes back to handwriting SqlConnection, SqlCommand, SqlDataReader, and SqlDataAdapter, MVC is the same: it's slow and hard to use ...

If there was a V5 frame, the spring of WebForm should last a long, long time ...

Now, although WebForm is drifting away, at least it does not have a slow reason, because of the V5.

Summarize:

Over the years, the framework has been accompanied by personal growth, and has evolved from a niche orm to a big, noisy ORM:

From the original as long as the database can operate, to product thinking, to the automated thinking, to the distributed architecture thinking, the next stage, what will be the thinking?

When you start a project, developers are weak, and it helps you solve efficiency and performance issues.

When your project is big, there is no architect, it helps you solve the problem of distributed architecture;

And all of this, you are now free to get ... (Yes, my blog installed a reward plugin, ha)

All in all, words and in short, always say, such a good framework, mud, even open-source ...

Cyq. Data V5 from getting started to abandoning the ORM series: the benefits of the framework