RE:Columnstore可以每秒扫描数十亿行吗? [英] RE: Can Columnstore scan billions of rows per second?

查看:73
本文介绍了RE:Columnstore可以每秒扫描数十亿行吗?的处理方法,对大家解决问题具有一定的参考价值,需要的朋友们下面随着小编来一起学习吧!

问题描述

你认为SQL可以在SQL 2016中扫描每秒数十亿行的数据吗?

Do you think SQL could do scans of 100s of billions of rows per second in SQL 2016?

我记得X-Velocity曾经吹嘘过几百万行。 SQL可以扫描数十亿行吗?

I remember X-Velocity used to brag about 100s of millions of rows. Can SQL scan 100s of billions of rows?

推荐答案

SQL
服务器内存中OLTP和列存储功能比较

技术白皮书

作者: Jos de Bruijn(微软),Sunil Agarwal(微软) ),Darmadi Komo(微软),Badal Bordia

(ASPL),Vishal Soni(ASPL)

技术评论员: Bill Ramos(Indigo Slate)

Authors: Jos de Bruijn (Microsoft), Sunil Agarwal (Microsoft), Darmadi Komo (Microsoft), Badal Bordia
(ASPL), Vishal Soni (ASPL)
Technical reviewers: Bill Ramos (Indigo Slate)

已发布: 2015年5月

适用于:
Microsoft SQL Server 2016和SQL Server 2014

Published: May 2015
Applies to: Microsoft SQL Server 2016 and SQL Server 2014

内存中的OLTP和内存中的Microsoft SQL Server列存储跳过了一代OLTP和  data
仓储解决方案,显示10-100x的典型性能改进。内存中 
OLTP,组织可以将其交易应用程序加速
最多30倍,并且可以更快地提供 
业务洞察,查询和报告速度提高100多倍。内存中列存储索引, 扫描典型行业硬件上的每秒数百亿行的
,为用户提供以思维的速度与前所未有的
数据量进行交互和探索.SQL Server提供 
具有强大数据平台架构的组织,可以通过  近乎实时交易和分析数据的能力。

In-memory OLTP and in-memory Columnstore of Microsoft SQL Server leap a generation of OLTP and data warehousing solutions to show typical performance improvements of 10–100x. With in-memory OLTP, organizations can accelerate their transactional applications by up to 30x and can deliver faster business insights with more than 100x faster queries and reports. In-memory Columnstore indexes, with scan rates of tens of billions of rows per second on typical industry hardware, provide users the ability to interact with and explore an unprecedented amount of data at the speed of thought. SQL Server provides organizations with a robust data platform architecture that can advance your business goals with the ability to transact and analyze data in near real-time.


这篇关于RE:Columnstore可以每秒扫描数十亿行吗?的文章就介绍到这了,希望我们推荐的答案对大家有所帮助,也希望大家多多支持IT屋!

查看全文
登录 关闭
扫码关注1秒登录
发送“验证码”获取 | 15天全站免登陆