I am mining new and upcoming ALTcoins. Currently I am running the best droplet that DigitalOcean has which is 8CPU's at 0.238 per hour. but they throttle me around 15%. Right now I am scoring 65% CPU usage. With AWS I am using c1.xlarge bidding at .20 per hour
Both running the linux ver I mentioned earlier.
Working with CloudShards and they can offer me the OS and the Chips at 90$ a month. I think I will go for that
I wouldn't pay 20 cents per hour on the c1.xlarge, because you are not getting the best performance for your buck. I constantly use the c1.medium which is just 1/4 the size of the c1.xlarge, but I pay cheaper than what you are paying for your c1.xlarge (if you compare the pricing on 4 instances medium vs 1 xl)
This is what you should do.
The c1.medium is actually much better because as I can guess you are paying 20 cents on the c1.xlarge based on the spot instances(since the ongoing price is 58 cents per hour)
So with the spot instance price you are getting around 66% off which you are paying 1/3 of what the actual going rate is.
On the c1.medium instances I was paying as low as 1.8 cents each and as high as 2.8, very few times I was paying up to 4.5 cents per hour and almost never 5 cents or more.
The western regions always start around I think 2.5 cents per c1.medium instance, on the eastern region they always start around 1.8 cents. However, the bitch here is that the price flunctuates a lot more in the eastern region instances that the price of 1.8 cents is not really as long as the price of 2.5 cents.
Now, it is important to notice the limits here on the instances you can fire on the spot instances.If you check their FAQ, you can launch 100 spot instances per region on the lower specs instances, and 20 on the bigger ones.
So 100 instances times 8 regions, that is a maximum of 800 instances before you have to call in and get permission to run more instances.Unfortunately the c1.xlarge is part of their 20 spot instances limit . Thus, allowing you to run a maximum of 20 instances per limit which is equal around 80 instances of c1.medium, while the c1.medium instance you can run a maximum of 100 spot instances so you get 20% more power if you go with the smaller instances.
The price also flunctuates a lot more on the very small instances like the micro and the bigger instances. Also, don't go for the medium, small, or large instances those are the most in demand so you hardly ever get a discount on the spot instances( you probably only get 40-66% off at most).
With the c1.medium instances I was paying as low as 1/8 of the actual price of the going on-demand price.
For your needs, what I would recommend is that you take advantage of amazon spot instances the smart way. Since those fuckers have some gay shit where the price could spiked up to 100000000% percent, basically paying like up to 1000 dollars per hour on some instances , you have to remember that you need something that queries the current going price of the spot going prices so you can reduce your costs.
I was using Iron.io IronWorker to run a task every 5 mins when I was running a lot of c1.medium instances so I can get the price of the going current prices and decide which region was the cheapest at the current moment to launch my instances. However, I never really used more than 30 instnaces at the time, I had plan to use like 300 but never got the change yet.
The High CPU instances are perfect to run browsers, I was running a bunch of docker containers with firefox/chrome installed with a virtual display like xvfb.