Big Data is the key, Oracle, SQL Server into a thing of the past?

Primer

Someone in a dedicated SQL number of public comments as follows:

Big Data is the key, Oracle, SQL Server into a thing of the past?

The message touched a very sensitive issue: engage in a relational database there a future? Now in 2020, the block chain is hot, AI has talent "surplus" big data have become commonplace turnip cabbage, you are still engaged in SQL Server? Are you still engage in SQL?

the reason

The last 10 years, changes in the field of data management can be described as brilliant. What is Hadoop, Spark, MemcacheDB, Redis, Vertica, HANA, Cassandra, MongoDB, InfluxDB, TiDB and so forth. All kinds of high-end summit, the forum really saw AI, big data, rare figure of a traditional relational database.

Preliminary analysis down to four reasons:

One
large-scale application of cloud computing with the rise of open-source database, greatly reducing the cost calculation of power and data storage.

Leader Oracle to traditional relational data, for example, with the rise of the Internet, a massive user-generated huge amounts of data, but when large-scale deployment of Oracle, Oracle not only the software itself is expensive, with the support of IBM's minicomputer + EMC storage spending is also quite large. For example, Oracle's data clusters moved from a room with a data center in another room, toll must be hundreds of thousands. Hardware and software maintenance costs, including the requirements for administrators is also very high.

Further Oracle performance itself has an upper limit, followed by a black box is closed source. For the scene did not come across, no matter how hard, but also unpredictable problems that may arise.

Chinese Internet companies mostly wilderness origin, highly valued for the price. Then boss of American Internet company Yahoo began pioneered the use of MySQL database, Chinese Internet companies follow suit.

The most typical example is Ali Baba. Alibaba was once known as Oracle's "star customer", "Whampoa Military Academy." But with the surge in the number of registered users, data generated by more and more users, Taobao is enabled Asia's largest Oracle RAC cluster, the amount of data Alibaba B2B Chinese station is also due to large volumes of data and business requirements, the annual 08:00 am -09: CPU 30 maintained 98% utilization, server load is also high, even with cold isolated manner can not solve a large capacity and a large data concurrency problem, replacement of a storage device in such a situation will soon reappear .

Alibaba forced to "save the" a technical route to the IOE: "low-cost, linear control, decentralized (ie distributed): go to IBM, PC Sever replace minicomputers; to Oracle, replaced with MySQL; to EMC , with low-end storage. "

In the explosive growth thereafter, including 2014, 15, 16 two-eleven years as the representative of the business peak, Ali engage in out based on open source version of MySQL branch extremely well to support demanding business requirements, and ultimately very convenient cloud oriented, while the open source MySQL in the domestic Internet companies gradually began to replace Oracle's status to become the first choice of many Chinese Internet companies.

The second
reduction cost general purpose hardware. With Moore's Law, even the HANA ,, Vertica, InfluxDB and so a lot of memory or distributed-memory database.

So now with cloud storage cloud computing in the end how cheap it today to see advertising a site just taken, right?:

Big Data is the key, Oracle, SQL Server into a thing of the past?

Big Data is the key, Oracle, SQL Server into a thing of the past?
Cloud Hosting a website advertising

This market-oriented large-scale sales of the product, can be used as an inexpensive PC Server, can be truly cabbage price, this is simply unimaginable 10 years ago.

The third
addition MySQL with the code document is public, so the corresponding requirements for database development and maintenance personnel will be greatly reduced.

Finally,
from the development of SQL to NoSQL, support for semi-structured data, databases shine unstructured data, unstructured data can not handle this kind of problem before been resolved, unstructured data processing become the standard configuration, seems to occupy the market hot spot. The NewSQL still under development.

development of

The direction of development of the database generally includes the following aspects:

  • To thoroughly distributed development, Oracle and other database has been implemented by single cluster

  • Integration with the OLAP OLTP, this feature has been implemented very early

  • Integration of structured unstructured, currently Oracle, MS SQL Server, also has been achieved

  • Integration with disk memory database, currently Oracle, MS SQL Server, also has been achieved

  • Integration of rows, columns, graphics, objects, documents, content, timing, search engines, navigation category, and so different types of databases, we have a small preliminary study found partially achieved

  • A little early analysis, data modeling tool supports the above features are cashing in, but there is a great distance away from the perfect support

In summary point of view, the future longitudinal database personally think that pictures, audio and video storage will continue to be a hot spot, will be further integration of the lateral development, such as audio and video integration with structured data storage and so on.

Conclusion
It is not a traditional relational database completely become a thing of it? If you think so, you would be wrong! Let's look authoritative data ranking:

Big Data is the key, Oracle, SQL Server into a thing of the past?

In recent years, the popularity of data trends

Big Data is the key, Oracle, SQL Server into a thing of the past?

December 2019 database popularity ranking

This shows that unstructured data, while hot, but relational databases still sit back!

The reason where is it? Database development colorful, but in the end it is still the key to good business scenarios. Different types of database adaptation business scenarios. Structured data in the real world there are a lot, from relational databases to handle them is relatively mature, after decades of development and a variety of stringent tests, all aspects of the unique advantages. Now fires its focus on big data effectively solve semi-structured and unstructured data previously been unable to deal with this long-standing problem. Unstructured play allows database relational database task, is tantamount to talk to a large excavator truck race up the goods.

Before the relational database does occupied the absolute mainstream. Only in recent years with the development of 3G, 4G, images, audio and video content increased slightly. With the later development of 5G, AR \ VR, and so things new technologies, storage has become expensive, block chain with quantum computing will bring new changes to the world ...... but no matter how developed, I think relational database over the next decade will remain the mainstream, then far in the future it will have an important role. Because in the foreseeable future, structured data is still humanity's most important digital assets.

I am concerned, the original data related to dry week.

Guess you like

Origin blog.51cto.com/14658150/2464104