Database veteran: The Big Data age NoSQL is not disruptive technology

A few years ago, when people talked about the emerging NoSQL database technology, a considerable part of the view was that it was only a matter of time before NoSQL replaced traditional relational databases in large data markets. Now, this prophecy has not materialized, Mitchell Kertzman's general manager Hummer Winblad that in most cases, NoSQL did not show the so-called revolutionary.

As a database veteran, here are excerpts from some of Kertzman's views in this week's video interview:

What people need is actually SQL.

If the cloud's SQL database is resilient and scalable enough to handle the amount of data needed, people will never need nosql.

See NoSQL Hummer Winblad himself is also http://www.aliyun.com/zixun/aggregation/13374.html ">newsql start-up company NuoDB Investors, NuoDB's main competitor in the Newsql market is geniedb or even vmwaresqlfire. In the competition with NoSQL database start-up company, one of the important reasons for NuoDB to play the song of Triumph is:

NuoDB has always believed that customers who run critical business applications actually want a better relational database.

"I think NoSQL is a new market opportunity," Winblad said, "but its market size has yet to be validated." ”

Relational databases are a large 35 billion dollar market, not that people like SQL query language, but because the acid properties of relational databases can guarantee the security of transactions. If you miss a status update or a "praise" failure on Facebook, it's not a big deal, but if a bank loses a deposit, even a small deposit, it's a serious accident.

NoSQL is not disruptive technology.

Kertzman pointed out that the IT market has experienced a lot of disruptive technology, such as relational database is to network database subversion, cloud computing on the client/server subversion ... People like subversion because disruptive technology creates many new opportunities, but NoSQL does not look like a disruptive technology that people expect.

Kertzman the object database movement in the 80 's, for example, "That's not disruptive technology, because the surviving databases such as Sybase and Oracle are simply adding object database technology to the customer's needs, rather than the object database as a religion." ”

Kertzman that MongoDB as a representative of the NoSQL database and won orders from companies that did not care about the acid characteristics and did not want to pay a large price for the Oracle database. But kertzman that MongoDB's turf is threatened by Teradata, Memsql and Heroku, "we all like to invest in companies that can use new, reliable technologies to solve complex problems, and I am not referring to the ownership of intellectual property, but to the very high technical threshold." MongoDB's success so far has not been technically a high threshold. ”

So what is the disruptive technology of the big Data age, Kertzman thinks Hadoop is a real opportunity, but the problem with Hadoop is that enterprise deployments are slow, while Sql-on-hadoop is the technology path for large data vendors to speed up the choice of Hadoop deployments.