Comments on: ETech 2007 SmugMug Amazon Slides are Up!

By: Brief Introduction To Cloud Computing — Keenpath

Brief Introduction To Cloud Computing — Keenpath — Wed, 05 Mar 2008 13:39:42 +0000

[...] a successful photography community, saved an estimated $2 $1 Million dollars (edit per this link) in 2007 by using Amazon S3 cloud storage. Organizations and individuals can benefit from mass [...]

By: Lessons on How to Use Amazon Web Services « HighContrast

Lessons on How to Use Amazon Web Services « HighContrast — Thu, 21 Feb 2008 02:50:29 +0000

[...] AWS outage comes a link to a great presentation on how to successfully use AWS from SmugMug founder Don MacAskill given at ETech last [...]

By: Melvin Ram

Melvin Ram — Sun, 03 Feb 2008 00:04:58 +0000

Your slides won’t open for me. It says it’s unsupported or damaged. I tried downloading it a few times to make sure it wasn’t just a corruption during download. Could you email it to me. Thanks.

~ mel

By: Can Amazon Make Startupping A Science? « GigaOM

Can Amazon Make Startupping A Science? « GigaOM — Thu, 13 Sep 2007 15:58:25 +0000

[...] But the money part is important too. SmugMug CEO Don MacAskill reiterated he’s saved $692,000 using AWS for his photo-sharing startup. He said that figure was as of April of this year and would be higher if he recalculated. [...]

By: EwanSilver.com » June 21 2007

EwanSilver.com » June 21 2007 — Thu, 21 Jun 2007 21:55:29 +0000

[...] Amazon Web Services and the lack of a SLA Interesting comments from Greg Linden and Doug Kaye where they take differing views and I think they are both right. Those SLAs suck if you are trying to run a business that is succesful - +1 to Greg. The economics of EC2 make it a very cheap way of boot strapping an idea. The probability is that the idea will fail any way. +1 to Doug. SmugMug pdf from Greg [...]

By: SmugBlog: Don MacAskill » Blog Archive » Amazon S3: Show me the money

SmugBlog: Don MacAskill » Blog Archive » Amazon S3: Show me the money — Mon, 30 Apr 2007 18:18:09 +0000

[...] UPDATE 4/30/07: I’ve updated these numbers since this post was written in November 2006. It’s now been 12 months and we’ve saved almost exactly $1M. You can see the numbers in my ETech slides. [...]

By: Jacques Marneweck

Jacques Marneweck — Mon, 02 Apr 2007 20:19:06 +0000

Don, interesting presentation. How are you abusing libcurl to get added speed?

By: PanMan

PanMan — Mon, 02 Apr 2007 07:32:44 +0000

Thanks for the explanation: that makes sense.

By: onethumb

onethumb — Sun, 01 Apr 2007 23:06:14 +0000

@PanMan:

Yes, that’s a big drawback to proxying - if you’re actually proxying lots of your data.

But our “hot” data, the data that’s accessed most of the time, is stored locally at SmugMug, so we’re only doing proxy reads in case we don’t have the data locally or we have an internal failure somewhere.

So for the vast majority of data, we write it once to S3 and then we’re done.

Make sense?

So the answer to your other question is yes, 90% of our data lives at S3, but not in our own datacenter now. We do trust them that much.

Finally, we use libcurl.

By: PanMan

PanMan — Sun, 01 Apr 2007 22:37:02 +0000

Hi, like the presentation!
But one point: the big drawback of proxying, instead of redirects, is that you pay for the data transfer 3 times, instead of 1, right? (out of S3, into your datacenter, and out of it). Isn’t that a strong reason to use redirects? Why still use the proxying?
Another question: Do you still have all your own data as well, or is there data that’s only in S3? (do you trust them that much?).
3th question: what cURL library do you use? The PHP internal one?
Thnx!