Azure分析和数据仓库成本计算 [英] Azure Analysis and Data Warehouse costings

问题描述

我在Azure上花费AS的时间很难.我们目前有一个解决方案，即我们使用SSAS在数据仓库上通宵处理销售多维数据集，但我发现几乎不可能 在Azure上花费等效或至少使其具有成本效益！< o:p></o:p>

首先，对于数据仓库，您是为数据库在线还是每小时为其付费?被AS访问.同样，对于AS，您只需要支付其实际处理时间 立方体?如果没有，并且您在一个730个小时的月内有效地为这两者支付了费用，那么我无法确定在我们的情况下它如何能够为SMB节省成本，我确定这不是所有的独特之处吗?

我知道您可以暂停Analysis Service，但可以根据时间自动进行工作运行?

任何帮助将不胜感激.

谢谢

解决方案

回复:DW
如果使用Azure DW，则可以暂停计算节点.

回复:AzureAS
只要您需要Azure AS实例可用于查询(或处理)，就可以运行它.

有两种方法可以自动化(例如按计划)两个平台的暂停/恢复.

对于您来说，最好使用Azure DW + Azure AS架构.这假定您所有的报表需求都可以通过Azure AS模型得到满足.

用于每晚处理...

1. 恢复Azure DW计算(节点)
2. 运行一些ETL以更新Azure DW数据库(例如从本地源?)
3. 处理Azure AS模型
4. 暂停Azure DW计算

或者，您可以根据延迟要求在一天中多次运行上述ETL模式.有很多选择，有效地使用它们是优化PaaS惊人的价格标签的要求.

Hi,

I'm having a hard time costing AS in Azure. We currently have a solution on premise where we use SSAS to process a sales cube overnight on a Data Warehouse but i'm finding it almost impossible to cost the equivalent in Azure or at least make it cost efficient!<o:p></o:p>

Firstly for the Data Warehouse, do you pay for every hour that the DB is online or only when its being accessed by AS. Likewise with AS do you only pay for the actual time its processing the cube? If not and you effectively pay for both over the course of a 730 hour month then i cant see how an it can work out cost efficient for a SMB in our scenario which i'm sure isnt all that unique?

I'm aware that you can pause the Analysis Service but can that be automated based on the time the job runs?

Any help would be much appreciated.

Thanks

 

解决方案

Re: DW
If you use Azure DW, then you can pause the compute node(s).

Re: AzureAS
The Azure AS instance needs to be running anytime you need it to be available for queries (or processing).

There are ways to automate (e.g. on a schedule) the pause/resume for both platforms.

In your case, it may make sense to go with an Azure DW + Azure AS architecture. This assumes all of your reporting needs can be met via the Azure AS model.

For nightly processing...

1. Resume Azure DW compute (nodes)
2. Run some ETL to update the Azure DW db (e.g. from on-prem sources?)
3. Process Azure AS models
4. Pause Azure DW compute

Alternatively, you could run the above ETL pattern several times throughout the day depending on latency requirements. There are a lot of options and using them efficiently is a requirement to optimizing the sticker-shocking price tag of PaaS 

这篇关于Azure分析和数据仓库成本计算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！