I have a local web service which is using 60 running node.js servers. Now I want web service to drag over on Amazon EC2.
Which the best plan to choose to run 60 concurrent node.js servers?
As I understand it, for optimal performance need to run one node.js server for one CPU core. If I use the plan Cluster Compute Eight Extra Large Instance, where 88 EC2 Compute Units, will be issue in performance?
This would totally depend on what the performance constraints of you apps are (i.e. are you memory-bound, CPU-bound, or IO-bound)?
You best starting point is to just look at your current server and find an equivalent EC2 instance type. Of course you should consider that if you current server uses higher end hardware at all, that the performance would not be as good on the virtualized EC2 which runs on commodity hardware, so perhaps you want to go with something with a little higher RAM, CPU, etc. than what you currently have.
Of course this above advice is based on the use of a single server. IMO, that is really not the use case that EC2 was created to resolve. To truly take advantage of the EC2 infrastructure, you should think about how you can horizontally scale your services. In some cases it can be more cost-effective to have a fleet of lower-price instances perforing your work rather than a single monolithic larger instance.