注册 登录  
 加关注
   显示下一条  |  关闭
温馨提示!由于新浪微博认证机制调整,您的新浪微博帐号绑定已过期,请重新绑定!立即重新绑定新浪微博》  |  关闭

Ooi Beng Chin 黄铭钧

Databases, Machine Learning and Systems

 
 
 

日志

 
 

New Architecture for DBMS?  

2011-10-18 10:14:40|  分类: 默认分类 |  标签: |举报 |字号 订阅

  下载LOFTER 我的照片书  |
DBMS system architecture is 30 year old.  The system is very efficient for processing set oriented operations and was designed to perform some of these operations, such as join, very efficiently.  The technology is matured -- its query processing strategies, indexing methods such as the B+-tree and R-tree, concurrency control, recovery, etc have not changed much in the last 10 years or so.  In recent years, there has been renewed interest in providing column-based storage and processing.  Is it a paradigm shift or just a re-invention and re-use of old technologies in the context of vertically partitioned data? The same argument applies to the log structure only system, where data are appended without in-place updates.  The reconsideration is generally driven by the need for supporting huge amount of user generated data.  On the other hand, the MapReduce framework has generated a huge amount of interest in the last five years, and database researchers, including myself, have been trying to make it database-centric even when the orginal design was meant for different applications and settings.  Have we learned anything from it?  Can we really have a paradigm shift when most database operations and designs are constrained by the hardware architecture?
  评论这张
 
阅读(887)| 评论(0)
推荐 转载

历史上的今天

在LOFTER的更多文章

评论

<#--最新日志,群博日志--> <#--推荐日志--> <#--引用记录--> <#--博主推荐--> <#--随机阅读--> <#--首页推荐--> <#--历史上的今天--> <#--被推荐日志--> <#--上一篇,下一篇--> <#-- 热度 --> <#-- 网易新闻广告 --> <#--右边模块结构--> <#--评论模块结构--> <#--引用模块结构--> <#--博主发起的投票-->
 
 
 
 
 
 
 
 
 
 
 
 
 
 

页脚

网易公司版权所有 ©1997-2017