简体   繁体   English

高性能大文本存储解决方案

[英]High performance large text storage solution

Here is the scenario: 这是场景:

1) All the data currently is stored in SQL Server 1)当前所有数据都存储在SQL Server中

2) There is one table X that stores the most important and highly accessed data 2)有一个表X存储最重要和最易访问的数据

3) Each record in the table X has one column Y that stores up to 2GB of text 3)表X中的每条记录都有一个Y列,最多可存储2GB的文本

4) Number of records is growing (> 10 million and growing) 4)记录数量正在增长(> 1000万并且还在增长)

5) The column Y with max of 2GB of text per row takes about 90% of space of table X 5)每行最多2GB文本的Y列占用表X的大约90%的空间

Problem: database becomes unmanageable and we need to store data from Column Y somewhere else (outside of SQL Server) 问题:数据库变得无法管理,我们需要将列Y中的数据存储在其他地方 (SQL Server外部)

Question: 题:

What is the best solutions for storing large text? 存储大文本的最佳解决方案是什么?

I am looking for a solutions with least amount of overhead : 我正在寻找开销最少的解决方案:

  • fast reads 快速阅读
  • fast writes 快速写
  • scalable 可扩展性
  • can handle concurrency well 可以很好地处理并发

I don't care about full text index or any extra overhead. 我不在乎全文索引或任何额外的开销。 Just plain storage and access. 只是普通的存储和访问。

PS PS

Some ideas that I have so far: 到目前为止,我有一些想法:

  • very simple NOSQL database 非常简单的NOSQL数据库
  • flat file - based database 平面文件数据库
  • custom solution that I can write from scratch using some library 我可以使用某些库从头开始编写的自定义解决方案

From what I read about NoSQL technologies, Riak may be a great fit, however, there's certainly overhead with this. 从我读到的有关NoSQL技术的内容来看,Riak可能非常适合,但是这样做肯定会产生开销。 Flat files sound promising, but highly OS and patch dependent - you don't want to be stuck explaining to a customer that Windows Updates caused performance problems. 平面文件听起来很有前途,但是高度依赖于操作系统和补丁程序-您不想被困于向客户解释Windows Updates导致性能问题。

声明:本站的技术帖子网页,遵循CC BY-SA 4.0协议,如果您需要转载,请注明本站网址或者原文地址。任何问题请咨询:yoyou2525@163.com.

 
粤ICP备18138465号  © 2020-2024 STACKOOM.COM