Hacker News Reader

Comment by Zak | Hacker News Reader

> Even without staff, DPReview‘s server costs are likely quite high given the huge amount of image data that has been accumulated over the years.

That's a matter of perspective. If I had to pay them out of pocket at retail prices, I imagine I'd consider them high. If I was a multinational corporation that owned the world's largest cloud computing service, I probably wouldn't.

View on HN

Retail costs for cloud storage have been in a race to the bottom for years, and the margins are not large.

For storage, maybe, but I would think the bandwidth is the actual cost not the storage, and cloud is definitely not in a race to the bottom on bandwidth pricing.

Bandwidth went the same way, AWS/Azure/GCP don't translate it to the customers purely for commercial reasons[1]. Also, Amazon.com (or whatever is the division owning DPReview) doesn't pay the retail price for traffic-out, for sure. The discounts for very big customers are very important.

[1} https://blog.cloudflare.com/aws-egregious-egress/

It really depends on the type of storage. Long term storage has great margins over time because capex is one time and the opex is relatively low.

Wasabi gets $6/tb/mo... indefinitely.

That doesn’t really change anything. The opportunity cost would be the same, which is what matters for companies. Those resources could be sold for market value rates to a customer if they weren’t being allocated to DPReview.

Opportunity cost implies that there is an alternative that AWS forgoes if they allocate resources to DPReview. Can’t they just provision more hardware?

Hardware cost money? The rate at which AWS buys more hardware is consistent with their income from their cloud rental business. If they're not going to buy the hardware otherwise, then evidently at cloud market rates the marginal income isn't worth it.

Hardware, rackspace are all resources. If something is using up hardware and rackspace, there is opportunity cost.

The richer we are the less we can afford.

If they cared about saving money they would try to sell it. Something which doesn't bring many millions is to small for Amazon to care even if it cost rounding error on a balance sheet to run it.

The problem is that at a big/huge company, every little project imposes some share of costs and distraction all the way up the line and on supporting groups. And, yes, that means something could be reasonable as at least lifestyle business for someone. But unless it more than incidentally supports more central parts of Amazon's business, a million dollar side business isn't really very interesting to a company like Amazon. And have more than a few of those and people start to reasonably ask what Amazon is actually focused on.

That doesn’t stop them from selling it.

I think you are describing cost in the traditional sense, and not opportunity cost.

Why is it an either/or? Is Amazon bandwidth-constrained?

Reply on news.ycombinator.com