Re: [tlug] hosting in Japan

[Curt Sampson <cjs@example.com>]

On 2017-02-13 09:59 +0800 (Mon), Raymond Wan wrote:

> Quite frankly, I chose AWS because they were giving some talk /
> training in Hong Kong for free.  Then there was the "first year free"
> deal and, as expected, once that year ended, I was reluctant to look
> at an alternative.  I'm sure that's their marketing ploy...
> 
> On a few occasions, nothing was running on my instance, but it became
> very slow.....

That "first year free," as far as compute goes, is a single t2.micro
instance which gives you 144 minutes per day of CPU time (overall, 10%
of a single core). So it could well be the case that at some point you
did something like a large build maxing out a core for 15 minutes or
so and ran out of CPU credits.

That's not the only thing you can run out of; you also get just a
certain amount of disk I/O (IOPS or I/O Operations per second) that
you need to manage as well.

There are a huge number of different types of VMs and disk storage on
AWS EC2, and the pricing becomes even more complex if you try to save
money by reserving instances, such that you get a discount but you
have to pay for them whether you use them or not.

Google Cloud Compute is a lot simpler; though they do have different
machine "types" these are just suggested configurations picked out of
the full spectrum of configurations; you can bring up a couple of
sliders at any time to decide how much CPU and memory you want on any
instance and the rate is linear. Attached storage is also somewhat
simpler, though the IOPS are tied the size of the disk you choose. And
they just automatically give you a discount on the rate if you use
something for more than a certain number of hours per month, so you
don't need to think about that.

cjs
-- 
Curt Sampson         <cjs@example.com>         +81 90 7737 2974

To iterate is human, to recurse divine.
    - L Peter Deutsch