Ask HN: Why do AWS, Azure, and GCP charge so much for outbound data transfer?

[Kareem71]

Some cloud providers seem to be able to offer a much better deal i.e https://www.digitalocean.com/pricing/

whats the logic behind their data pricing structure?

[gostsamo]

Keeping your data hostage. The more you have with them, the more you will have to pay in order to go to the competition. And wherever is your data, there must be your highly-paid services that utilize it.

[maltalex]

This is probably the right answer.

Outbound bandwidth fees are essentially a form of "exit tax" that makes it more costly to use resources outside the cloud (like other clouds). It's there to lock you into your current cloud's ecosystem even if the grass is greener elsewhere.

Moving data around is already a lot harder than moving other resources (like compute). The bandwidth fees also add a monetary cost to such moves.

[kohanz]

Some providers provide great pricing and service, but please, unless your project is a hobby project, do not use DO's storage and CDN solution (called Spaces) in production. Made that mistake once and caused a lot of headache and regret. In fact, if you're considering DO at all for production, subscribe to their status page for a steady stream of incidents to your inbox. And if it's not an incident, it a 2-day long "maintenance" that gets announced literally one minute before it starts. I've stayed subscribed partially for the comedy of it at this point (migrated away from DO a while ago). DO engineering and devops must be a lively place.

As for the pricing reason, part of it is reliability, but part is that all of the other product offerings (which are available at the big players and not elsewhere) are subsidized by these costs that everyone needs. For example, Amazon Cognito is dirt cheap, why do you think that is? Come for the X, Y, Z and make us money back through storage/egress.

[tiernano]

I dont completely agree with the pricing, dont get me wrong, but i can see why they do it... You are not just paying for bandwidth. You are paying for the rest of the infrastructure. So, for example, in Azure's case (which is what we use in $dayjob) we get VNETs, NSGs (basic firewalls), Routes, DNS, IP Addressing, etc, plus then bandwidth. Yes, i would like the bandwidth bill to be cheaper, but you also need to look, again, in the case of Azure, of the rest of what they are offering in a global network and number of POPs around the world...

[idunno246]

This is the argument aws uses. In a dc you might need a bunch of high price Cisco equipment, and that’s part of what is getting priced into the bandwidth.

the cynical part of me agrees though that making moving out difficult/expensive is a large part of why too. Probably some combination of both

[lrossi]

It’s a matter of supply and demand.

Most of the cloud customers are using it to serve data more than receiving data. Meanwhile, the cloud providers are connected to the Internet via mostly symmetrical links. This means that they are underutilized in the upload direction, and loaded in the download direction. So they charge more in the direction where there is demand.

Download traffic is not the only one that is costly. Traffic between their datacenters from different regions may be billed as well.

[aaccount]

Cloud providers are full of hidden costs

[speedgoose]

Because they can.

Check Hetzner, OVH, or Scaleway too.

[atian]

Im not sure AWS’s peering can compare to Hetzner’s or Scaleway’s.