Web hosting in Asia is quite a different experience from what many people have when buying services from providers in Europe and North America. Aside from the obvious language and interface barriers that may be in place, it seems the Asian services offer more processing power for the money, and bandwidth is completely free1. Earlier this year, when the majority of my web services were running in Japan, I rarely had to monitor a server's load or its throughput as I knew the machines could handle just about anything (reasonable) that was thrown at them. More than this, if there was a rush of traffic to a website or if a particular podcast file was downloaded a few thousand times, I never had to worry about big charges coming back to me at the end of the month. This sort of peace of mind was great, as it meant I could expand my services and organically grow the platforms while intentionally spending the money required to make it happen.
Since moving my 10Centuries platform over to Amazon's Web Services, though, I've been consistently nervous about the popularity of the tool. Will I wake up one day to discover that a machine was set up to copy the same mp3 files again and again and again, racking up hundreds of dollars in fees? Will I receive an email letting me know that the auto-scaling has reached its maximum and the servers are being inundated with traffic that I need to pay for? Will some other unexpected occurrence take place that could bankrupt the project at any given moment?
I really hope not.
The reasons for moving 10Centuries from my Japanese provider to Amazon came down to factors such as auto-scaling, easy storage solutions, response times for the primary audience, and fewer geographic blockades2. Was it the right decision, though? When the platform was running on Japanese servers, I needed roughly 18 people to sign up for subscription packages in order to break even and have a little left over for things like new theme designs or domain names. On Amazon I need 68 people to sign up for subscription packages in order to break even … and that's not taking into account any costs that will be incurred should the service receive interest from a larger group of the population. With every measure of success, the Amazon Tax will need to be considered.
This makes me nervous3
In January I had set this year's 10Centuries budget at 90,000円, which is about $900 USD. This would cover the cost of two servers, a bunch of domain names, a handful of theme templates that I would convert, plus some unforeseen expenditures. That amount was gone by May as various costs quickly added up and drained the bank account. The existing subscriptions are helping a lot, but I can't shake the feeling I've made a terrible mistake moving the service to Amazon. Unless I can get the costs way down in the next 45 days, the service will need to return to Asia where services are cheaper … even though this will come with a number of very clear disadvantages to development going forward.
Using Amazon's infrastructure comes with a lot of benefits. There's no denying this. The amount of risk people leave themselves open to, though, can be quite scary.
Bandwidth is often free or so cheap it may as well be free.
The Great Firewall of China will occasionally refuse to allow traffic from Japanese servers through. Traffic from American servers, however, appears to be largely unaffected.
Very nervous … especially considering how some people have no problem wasting other people's money "for the lulz".