Are there any good sources of hardware sizing guidelines for servers running SSAS Tabular? Something along the lines of "if your models are estimated to be this big (in terms of memory) and you're expecting and average x-number of simultaneous connections...here's what you should start out with".

obviously they need more memory than most other systems to hold the data...but I'd also expect them to be able to take advantage of more cores and higher clock speeds...once it's in memory, its really just a matter of ripping through it and crunching the numbers. More cores to handle more concurrency, higher clock-speed for faster "crunching".

有帮助吗?

解决方案

you're supposed to aim for 40% of the memory size of the original database. 1tb data = 400gb ram for SSAS Tabular.

其他提示

FWIW, Microsoft just released a whitepaper (Jan 2013) that addresses this material

Hardware Sizing a Tabular Solution (SQL Server Analysis Services)

许可以下: CC-BY-SA归因
不隶属于 StackOverflow
scroll top