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

 
 

OLTP + OLAP  

2010-09-28 10:13:45|  分类: 默认分类 |  标签: |举报 |字号 订阅

  下载LOFTER 我的照片书  |

OLTP and OLAP workloads are typically handled separately by two systems with different architectures – RDBMS for OLTP and data warehousing system for OLAP.  Periodically, data in RDBMS are extracted, transformed and loaded (aka. ETL) into the data warehouse. The system-level separation was motivated by the facts that OLAP is computationally expensive and its execution on a separate system will not compete for resources with the response-critical OLTP operations, and snapshot-based results are generally sufficient for decision making. Although this system-level separation provides flexibility and efficiency, it also results in several inherent limitations, for example,

lack of data freshness in OLAP, redundancy of data storage, as well as high startup investment and high maintenance cost.

Unlike the situation with OLTP and OLAP workloads, the divergence between Web 2.0 application hosting and web data analysis is mainly by design. The storage layer and processing layer are loosely coupled so that the processing layer can read data in any format in bulk and perform the necessary processing to produce the indexes or views required by the applications. The frequency at which an analytical or bulk processing task is invoked is a business decision, and its data freshness is therefore determined based on needs. However, such design causes applications to rely heavily on periodically generated meta data (e.g., indexes) due its lack of OLTP support and transaction management. Further, due to design by choice, these systems do not support indexing mechanisms that facilitate ad-hoc query processing, and today, most of them cannot support real time search as they have not been designed to index the data as they are generated.

The consequence of not being able to support real-time search is that many applications that require real-time updates and search cannot be supported.  With the fusion of information and technology, the ability of being able to simultaneously supporting OLAP and OLTP is important.

Most database systems today support hot backup, which allow users to access the database while it is being backuped.  It is granted that the interaction between OLTP and OLAP is more complex than the OLTP and backup since backup is essentially disk based activity. The technology could form the basis for supporting both OLAP and OLTP within the same instance, assuming the hardware platform has sufficient capacity to support both loads.

 

  评论这张
 
阅读(1652)| 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