Cloud computing is gradually falling behind the relational database (reprint)

Abstract: Despite the large relational databases such as Oracle's offerings, has been deployed in many data centers, but cloud computing requires a different setting to its full potential.

 

"In the cloud computing plans in the relational database will find the shadow, this is not accidental, because the relational database is not suitable for cloud computing environment" Geir Magnusson, vice president of engineering 10Gen think so. 10Gen is an on-demand platform for service providers.

Magnusson help write too Apache Geronimo application server software, speech O'Reilly Web 2.0 conference in New York this week, he said: "Cloud computing is a different technology, different enough to change the developers look at the problem and solve problems the way". "We will have to rethink the way we do things," he said.

During the speech, Magnusson listed a number are specifically developed for the new database cloud computing environment, including the Google 'the Bigtable, Amazon's SimpleDB, 10Gen own Mongo, AppJet's AppJet open source database and Oracle BerkelyDB.

Magnusson noted that none of these databases are relational (He particularly emphasized a notable exception "Drizzle", MySQL used in a hardcover edition web environment).

These databases have some common features, it is these characteristics make them particularly suitable for cloud computing service type applications. Most of them can be run in a distributed environment ---- means that they can be distributed across multiple servers in multiple locations. They are not transactional in nature, and are at the expense of some advanced query capabilities in exchange for better performance. (In many cases, these databases can be retrieved by calling the object, instead of SQL, in any case, the programmer, the former is more natural).

Despite the large relational databases such as Oracle's offerings, it has been deployed in many data centers, but cloud computing requires a different setting to its full potential. Database component of cloud computing is necessary to disperse in different locations, which is the reason for the name come. Perform complex queries across vast geographical distances can reduce the response time. In addition, the design and maintenance of data backup support different locations and at a point when the paralysis can guarantee the synchronization of system data is not easy.

"Cloud architecture where the measure has the attribute relational structure we use now," he said, the result is: "In the cloud architecture, the relationship no longer exists, people look at the data in a cluster form."

Magnusson's view was endorsed at the meeting, another spokesman AdaptiveBlue's Alex Iskold of AdaptiveBlue is a customer-oriented company, which provides a browser plug-in, according to the use history of the user, using semantic tags and Web service reflect personalized settings. The company put service built on Amazon's hosted platform services, including SimpleDB. Iskold pointed out that if AdaptiveBlue use a relational database so that services will not expand into widespread use.

Reproduced in: https: //www.cnblogs.com/licheng/archive/2010/09/09/1822074.html

Guess you like

Origin blog.csdn.net/weixin_34102807/article/details/92626949