Hacker News Reader

Comment by skilled | Hacker News Reader

I still can’t believe this even happened. They could have archived a static version of the site and served it from AWS and put a banner on top of the site saying it is hosted on AWS.

I guarantee you they would have recouped the costs of keeping the site alive in no time.

What a stupid decision.

View on HN

But that exists on archive.org right?

Archive.org doesn't have the best usability or speed.

But they have a static version of the site which seems to be what a lot of people on this thread want Amazon to create. The contents are already not lost. (Though I can't speak for the forums.)

What's the engineering cost to move the dynamic site over to a static site and fix any issues there within when you have no engineers working on staff anyway. What's the ongoing price of dealing with things like legal threats take down notices etc. Especially when you have no engineer and no good way to just go remove individual pages or deal with those types of issues. What's the engineering cost to go deal with changing regulations like gdpr and right to be forgotten we need to go remove comments or other user posted information from someone. What is that giant mess going to look like and how do you get an engineer to spin up on it fix each of those things on a one to two year timer probably in between any given issue. How excited is that engineer going to be a 3 to 6 months break in something that's actually useful to their career to go patch someone else's tech debt. What level of quality of work does that lead you to get. Where do you even put the headcount for an engineer or who do you steal them from that is less important than a dead archived website that's been sitting around for years. All in I think a base level engineer at Amazon probably makes around 250 to 300 when you include all the support in for a and all of the benefits and everything else. That's for an entry-level engineer who probably will have trouble understanding the weird ass bespoke infrastructure that has been created for this one-time thing in a dead corner of the company. Well you may not agree with any of these things these are all considerations that would go into a decision like this and all the complications this is not some shit just running on your digital ocean droplets and it's all things that companies have to take into consideration when they are the size of Amazon and have a Target on their backs for regulators and lawsuits.

No I'm not saying either way if I support or don't support this decision all I'm saying is all these things would go into consideration into a decision like this from just a pure engineering level. I've sadly had some projects that I worked on in the past were they looked at it and they didn't want to do archival because there was just not worth the legal risk to deal with changing laws like gdpr so they just shut the feature down. This is often the calculus that happens how much revenue is it making what is the legal liability risk of keeping it up what is the ongoing engineering cost to maintain it or spin up an engineer and fix it in between. Given that 300K price and just assuming you got to do one year of fixes over the next 10 years that site better be bringing in multiples of $300,000 a year in revenue to even make it worth it to not just shut it off. Esp if you can't quantify community good will.

Personally one of the lessons I learned is if I build a system like I did for gaming again I will make it so that we can shut it into a top and list archival mode that can be run minimally from a CDN but sadly I didn't think of that when we built it and it was a little expensive to do it the right way after the fact since I was the only one that actually understood the tech.

> What's the engineering cost to move the dynamic site over to a static site and fix any issues there within when you have no engineers working on staff anyway. What's the ongoing price of dealing with things like legal threats take down notices etc. Especially when you have no engineer and no good way to just go remove individual pages or deal with those types of issues. What's the engineering cost to go deal with changing regulations like gdpr and right to be forgotten we need to go remove comments or other user posted information from someone.

archive.org seem to manage it on donations alone, for millions of websites. So presumably it can be done very cheaply, if you have a lean organisation.

It's possible that Amazon is not that lean organisation, of course.

`wget -r` is free.

Huh. Yeah. In fact anybody could do this and then re-host it statically (another thread has the costs for it and it's not bad). That would put Amazon in the interesting position of defending IP that it had previously deleted, but large corporations are pretty comfortable throwing lawyers at silly things like that.

Only let the GET and HEAD requests through, probably. I'd say about a day fiddling with the load balancer, at most.

>I guarantee you they would have recouped the costs

Do you know the costs. Please share

For a static version? I am willing to bet a single server with 10 Gb connection can do it

The very basic math on this is:

1,560,000 pages x 150 KB/page = 234 GB

234 GB x $0.023 per GB = $5.38

this is just to host all the pages themselves.

Data transfer per month: (Assuming 10 million pageviews)

10,000,000 x 150 KB = 1,500 GB

1,500 GB x $0.09 per GB = $135

GET and all other Requests:

$0.0004 per 1,000 requests

30,000,000 / 1,000 x $0.0004 = $12

---

Storage cost: $5.38

Data transfer cost: $135

Requests cost: $12

Total estimated cost per month: $5.38 + $135 + $12 = $152.38

---

This is NOT considering images themselves, though you get a general idea of what I am implying.

Enjoy the math. I was happy to share it with you.

One of the more useful comments I've read in weeks. Thanks.

Or $3.36 if your put it on Cloudflare R2 (s3 compatible) and host it there

Okay, well, in that case I will put down $500 to pay for that for ten years or whatever and one of you guys can wget -r it and put it on Cloudflare R2.

Reply to this post once you have set up the Cloudflare account and posted the content on it and I'll send you an email you can give me admin credentials on so I can put billing details in.

It's rate limited. A distributed effort to create a complete backup will be starting soon https://wiki.archiveteam.org/index.php/DPReview

Very affordable to keep online as an archive or infrequently updated blog.

If the founder really wanted it back, the correct response would have been asking Amazon in a public forum if they would like to pass on ownership. If done in a gracious and respectful tone, I have no doubt that would have engendered all around good feels. It could have even compelled Amazon to put one half quarter of engineering headcount into wrapping the site up in an AWS account to hand over.

The founder should have said they'd be happy to take on costs and future burden, host it on AWS, and continue to link to Amazon products and storefront. Should have said that they were grateful for Amazon's stewardship and that they'd like to continue carrying the torch. An easy all-around win.

The actual response from the founder was too negative and blame ridden if their intention was to resume operation.

Reply on news.ycombinator.com