问题
I'm considering using Fat Entities from Lokad.Cloud in a new Azure based application. At the moment I'm still trying to decide if I should go with Table Storage or SQL - I know SQL better, but TS has the potential to perform better if designed correctly. When I was reading the wiki from Lokad.Cloud I came across this page on AutoScaling. Under Limitations it says
Lokad.Cloud only supports a single VM size for now: VM referred small in Windows Azure.
Does this mean that I won't see any benefit when the number of worker roles are increased?
Any feedback from about this product would be greatly appreciated. Everything else I've read has been good - just worried about perfoamnce.
Mark
回答1:
Does this mean that I won't see any benefit when the number of worker roles are increased?
No. You missunderstood the statement. What they mean is that they only support Small compute instance at the moment. When you build for Windows Azure, your application is being deployed on a Virtual Machine (VM). You can choose the size of Virtual Machine for each Azure Role you define. And you can choose how many VMs to be running your code (know as Role Instance).
If your application architecture is Cloud Aware (designed to run in multiple instances) you will have benefits of running it on more than one Instance. Please note that here the correct term is Worker Role Instances, not Worker Roles.
Of course you may have more than one worker role, but this would mean more than one application. The whole point about scaling with Cloud services is that your single application is scaled out. In Windows Azure Terms, this scaling out is achieved trough increasing the number of Role Instances (also often referenced as "workers").
来源：https://stackoverflow.com/questions/8063098/fat-entities-performance-query