Amazon + Two Guys + $0 = Next YouTube | Hacker News Reader

Amazon + Two Guys + $0 = Next YouTube

blogs.smugmug.com

15 pointsdood9 comments

The configuration for each virtual instance that Amazon Cloud gives you is a 1.86Ghz with 160GB non peristant storage. Which means that even if not more computing power is required you will be required to pay for another instance if you need more than 160GB. If we suppose you run one balancer, and two instance for database for replication, you need to pay at least 3 instances. Which is $72x3=$216/month=$2512/year. For the traffic of 1.5mbit/s which is 474GB, means you need at least $94/month=$1128/year. So, for me I don't see really the advantage. If we suppose that with $2000 you buy a pc with 1TB and two dual core's at 1.8Ghz. So, 512+1128=$1640/year for the bandwidth. You can lease a dedicated T1 line (1.5mbits x 1.5bmits) for $5200/year. So, there is not really advantage going to Amazon. The only thing you save is maintaining hardware...

I would use Amazon only as backup like smugmug does...other than that for me, there is not avdantage...

O'r you could use Amazon S3 instead of EC2 for storage.

* Pay only for what you use. There is no minimum fee, and no start-up cost. * $0.15 per GB-Month of storage used. * $0.20 per GB of data transferred.

EC2 is designed for computation, S3 for storage.

I would like to know your answer to this then:

http://blogs.law.harvard.edu/philg/2007/01/05/best-web-server-program-for-a-lot-of-static-files/#comment-19962

If this is the question you are talking about:

"Leaving the files in S3 with global read perms would open up the possibility of a botnet attack whereby an attacker could simply request all your files continuously and thereby drain your bank account by using up bandwidth as fast as Amazon can deliver it. Is there any way to address this kind of abuse? Otherwise, agreed, serving directly from S3 would be a fine idea." (how do you do blockquotes, btw?)

then it doesn't really apply. If you have an EC2 instance running as your server, you aren't allowing global public access to your S3 data. You control bandwidth usage through your EC2 server the same way you would with any other server.

And I think that even with raw S3, there are ways, such as "use once URLs", that allow you to get around this problem.

Exactly. You have control over the read-access permissions on any content on S3. Anything set to 'private' cannot be accessed by the outside world without a URL with an encrypted key salted with your private key and an expiration date. I am already doing this with my startup and it is a wonderful feature, with no fears of 'botnet attacks'.

I would like to know your answer to this then:

http://blogs.law.harvard.edu/philg/2007/01/05/best-web-server-program-for-a-lot-of-static-files/#comment-19962

isn't part of the whole point that the thing scales up or down as you need it to? so, yeah, if you know up front what kind of constant bandwidth/storage/cpu load you are going to need then you're in a really good position to shop around for a better deal.

I finally got around to trying EC2 this weekend, and it's pretty slick. I just wanted a *nix box to try out SBCL threading. I love the convenience.

Anyone know off-hand what extended instruction set these machine instances offer? I.e. MMX? SSE? SSE2?

Reply on news.ycombinator.com