5 predictions for the future of the database

Michael Stonebraker is a preacher in the database field. He has a deep professional attainments and helped develop some of the most popular database systems, such as http://www.aliyun.com/zixun/aggregation/14171.html ">postgres,ingres and Vertica." His latest study, VOLTDB, a memory-based OLTP (online transaction processing) system, claims that the research results are two orders of magnitude faster than traditional programs.

However, Stonebraker's views have been controversial for years, with strong support on the one hand and strong opposition on the other. In 2011, for example, he said that Facebook was stuck on MySQL, "Alive as Dead";

Stonebraker a visit to this week's businessesflat-out show to talk about his own database market assessment reports, including NoSQL, Oracle's fate, and of course, Facebook's MySQL problem, below are some highlights:

A pattern does not fit all

"In any vertical market that can be thought of, there will always be some more appropriate solutions compared to traditional relational database systems," Stonebraker said.

He has always preached the idea, but now it looks more reasonable than before. There are column 820.html "> Storage architectures for analysis, there are memory architectures for transactions, and of course there are nosql architectures for key-value operations and new data types, and even diagram databases are starting to be commercially available."

To prove how these new types of databases become mainstream, Stonebraker said: "Like Obamacare, for better or worse, it has been built on NoSQL"

There can be multiple winners in this field.

"There will be 3 to 6 different types of database system architectures, with 2 to 3 successful vendors at every level," Stonebraker predicts. "I think the most important thing is that the traditional relational database system will shrink slowly and this may happen within 10 years," he said.

NoSQL's strong landing

Stonebraker said: "NoSQL will mean more than just SQL, and Cassandra and MongoDB have confirmed my predictions." "For example, pure low-level languages have little value, and Stonebraker thinks the NoSQL system will embrace acid in the future, which is already happening." Stonebraker that

NoSQL non-acid's biggest supporter is Google's Jeff Dean, who is basically responsible for all of Google's database products. Recently, he wrote a system called spanner, which can be said to be a pure acid system. As a result, Google is also moving towards acid, and the NoSQL market will also be free from eventual consistency toward acid.

Oracle will feel pressure from SAP

Among other things, I think what's really interesting is that sap is not getting the attention it deserves in the database field, and SAP's customers are also Oracle's biggest customers. Among these giants, Oracle and SAP will compete face-to-face.

Stonebraker that it may be a little early to happen, and that it's not yet clear how SAP's customers will respond, but it's best to give customers a compelling reason to migrate from Oracle to Hana.

Facebook will continue to look for a replacement for MySQL, but it may be futile

Facebook has become the toughest data management problem on Earth, and MySQL has been trying for years, but so far no database system has been found to match their size.

Stonebraker points out that even if MySQL is trying to match the size of the social giants with Facebook's efforts, it is clearly not a recognition of MySQL, but a recognition of Facebook's accomplishments in the database field.

Generally speaking, Stonebraker said (back to familiar but interesting wording), "the traditional code base vendors are 25 years old, and now it is time for them to retire home."