Take the 2-minute tour ×
Server Fault is a question and answer site for system and network administrators. It's 100% free, no registration required.

I see the smallest and the cheapest price for Amazon EC2 is 0.080 per hour which gives us 0.080*24*31= $60 per month -- the price of cheapest dedicated physical server.

So, what is the essence? May be they count only for access time or something?

share|improve this question

closed as not a real question by Sven, ceejayoz, Skyhawk, EEAA, Ward May 22 '12 at 5:09

It's difficult to tell what is being asked here. This question is ambiguous, vague, incomplete, overly broad, or rhetorical and cannot be reasonably answered in its current form. For help clarifying this question so that it can be reopened, visit the help center.If this question can be reworded to fit the rules in the help center, please edit the question.

    
It's up to you to close the instances you don't need, or you'll get billed as long as you keep them online. –  Bogdacutu May 21 '12 at 18:28
    
So this means no benefits from virtuality. –  Suzan Cioc May 21 '12 at 18:31
2  
If you really need 24/7/365, you should use their "Heavy Utilization Reserved Instances". There is an upfront charge, and a reduced per-hour cost, which comes down to significantly less than "On Demand" would cost running for a whole year. –  sysadmin1138 May 21 '12 at 18:34
    
I don't understand what does "reserved" mean. In all cases they will bill me while instance is running and won't if not. –  Suzan Cioc May 21 '12 at 18:36
1  
@SuzanCioc No, they will not reject your request to provision an "on demand" instance. You are thinking of a "spot" instance. –  Skyhawk May 21 '12 at 18:53

2 Answers 2

up vote 2 down vote accepted

No, the smallest and cheapest price for Amazon EC2 is $0.020/hour, for the t1.micro instances.

You only pay when the server is turned on. If you don't need the servers at night, you can turn them on and pay $0.000/hour (yes, nothing, except for storage costs) for that time.

This can be very handy for some use cases - as an example, some people might only need two small servers during the evening, but 10 hefty servers during the day.

Things get significantly cheaper when you purchase reserved instances, as well.

share|improve this answer
1  
Sorry but it looks useless. Nobody will turn off his/her site at night. This looks like just a marketing trick. –  Suzan Cioc May 21 '12 at 18:29
1  
@SuzanCioc On the contrary. I have a distributed application that does do dynamic scaling. The web front-ends never go down, but the processing nodes come and go. When we stop the processing instances, we don't get billed for when they're offline. –  sysadmin1138 May 21 '12 at 18:31
1  
@SuzanCioc: The benefit is not the lowest price but the fact that you only pay for what you need. And ~15$/month for the smallest option isn't exactly expensive. –  Sven May 21 '12 at 18:36
4  
@SuzanCioc: Then look elsewhere. They don't claim it's a service for everyone and they don't claim to be the cheapest. And if you think that 15$/month is expensive, you better start looking for another hobby. –  Sven May 21 '12 at 18:41
1  
@SuzanCioc - "Expensive" and "Inexpensive" are completely subjective terms. I'm building a system similar to what sysadmin1138 has, and believe me, it is very inexpensive for us to use EC2 versus having to purchase physical servers to handle a very bursty load. Like others have said, perhaps EC2 isn't for you. –  EEAA May 21 '12 at 23:04

It's certainly less expensive than keeping a pool of servers on (even with standby power management). The whole point of the cloud (regardless of vendor) is not to somehow claim that spinning up a cloud based server is going to be cheaper than spinning up a single dedicated server - but rather that the entire operational cost (power, cooling, space, admins) will be cheaper and on top of that you have access to almost infinite capacity. Keeping that extra capacity is WAY cheaper for the cloud than any individual company (thus the inexpensive bit. However AWS is still the wrogn model for the most inexpensive way to scale

PAAS (Amazon Beanstalk, Google app engine, Microsoft Azure), is a far more efficient way of scaling applications. In that model applications can scal as needed and deprovision on the fly. In IAAS you the user has to somehow manage the number of instances. In the Paas scenario your bill is simply how many resources you've used - without the need to manage any instances.

share|improve this answer

Not the answer you're looking for? Browse other questions tagged or ask your own question.