1. Disk I/O, 硬盘IO速度
硬盘的IO速度一直都是数据库的瓶颈，所以有条件的情况下尽可能的使用高IO的磁盘。
可以使用微软的工具SQLIO测试磁盘的IOPS
2. CPU的主频，
DW和传统的OLTP数据库在使用场景上不一样。
传统的OLTP数据库具有[事务小][并发多]的特点；而DW的数据库相比较具有[事务大][并发少]的特点。
所以对比起来，传统的OLTP数据库可以使用[低主频][多核]的硬件架构，而DW建议使用[高主频][少核]方案。
上述都是相对情况，对于不差钱的土豪，高主频，多核当然是更好的选择。
我们可以计算的是要满足具体的业务需求，需要多少CPU(Core)，多少内存。
MCR，Maximum Consumption Rate，这是一个Core的吞吐量指标
3. 计算MCR
可以使用下面的脚本计算出当前计算机的MCR
USE master;-- Create a database for benchmark queriesIF EXISTS (SELECT * FROM sys.sysdatabases WHERE name = 'BenchmarkDB')DROP DATABASE BenchMarkDB;GOCREATE DATABASE BenchMarkDB;GOUSE BenchMarkDB;-- Include a heap and a table with a clustered indexCREATE TABLE heap_table(col1 integer identity,col2 integer,col3 varchar(50));CREATE TABLE clust_table(col1 integer identity PRIMARY KEY CLUSTERED,col2 integer,col3 varchar(50));-- Insert 100 rows to start withDECLARE @i integer = 0;WHILE @i < 101BEGINSET @i = @i + 1INSERT INTO heap_table VALUES (@i, CAST(@i%5 AS varchar))INSERT INTO clust_table VALUES (@i, CAST(@i%5 AS varchar))END;-- Now keep reinserting exponentially until the tables each contain 2 million rowsWHILE (SELECT COUNT(*) FROM clust_table) < 2000000BEGININSERT INTO heap_tableSELECT col2, col3 FROM clust_table;INSERT INTO clust_tableSELECT col2, col3 FROM clust_table;END;
USE BenchmarkDBGOSELECT SUM(Col2) FROM heap_table WHERE col1 % 3 = 1GROUP BY col3;SELECT SUM(Col2) FROM clust_table WHERE col1 % 3 = 1GROUP BY col3;SET STATISTICS IO ON;SET STATISTICS TIME ON;-- run these muliple times and take an average of the logical reads and CPU timeSELECT SUM(Col2) FROM heap_table WHERE col1 % 3 = 1GROUP BY col3OPTION (MAXDOP 1);SELECT SUM(Col2) FROM clust_table WHERE col1 % 3 = 1GROUP BY col3OPTION (MAXDOP 1);/* Max Consumption Rate (MCR) is the average of (logical reads / CPU time in seconds) * 8 / 1024(or put another way, the size of the table in MB / CPU time in seconds)This gives us the throughput of a coreTo estimate the no. of cores required, use the following formula:(Amount of data scanned in an average query / MCR) * Concurrent Sessions / Target response timeFor example:(18000 MB/200 MBs) * 10 users / 60s response time = 15 cores (round up to 16) */
4. Memory内存需求
最少1核对应4G内存，或者对每组CPU给64-128G内存
