AWS lowers data transfer prices for PrivateLink, Transit Gateway, Client VPN | Hacker News Reader

AWS lowers data transfer prices for PrivateLink, Transit Gateway, Client VPN

aws.amazon.com

47 pointsbdb16 comments

How about lowering the data transfer prices going out to the internet? $0.09/GB is a joke (and in poor taste).

For what it's worth, we run "edge compute" with two separate mid-sized CDNs and pay about $0.005 per GB. This required a bit of negotiation, but has no minimum commitment and is on a monthly term. Yes, even lower fees can be had if you build your own infrastructure and pay for bandwidth on a CDR, but this is a fairly apples-for-apples case where we are not doing that.

So yes, AWS, GCP, Azure bandwidth pricing is very high. Oracle is, amazingly, the outlier with lower bandwidth pricing.

Oracle has to compete somehow (they're a distant fourth in public cloud market share) and bandwidth pricing is an obvious place for them to try. It's compelling, but I don't know if it's compelling enough for a lot of people to be willing to give Oracle money...

I wouldn't be surprised if data transfer fees as a percentage of their total revenue account for >10%. As long as GCP (standard tier 0.085) and Azure (0.08 non-MS-backhaul standard routing) don't charge extremely lower rates, it's not beneficial to lower their rate.

How about the price of EBS volumes as well? I believe the price for gp2 has been $0.10/GB (USE1) since 2014 and gp3 only reduced the price slightly.

I always assumed that when they stop dropping the price of older instance types, storage types and other such things, that’s their subtle way of encouraging both new and existing users to chose or migrate to the newer ones in order to not just get the new features but to obtain the cost savings.

How about lowering the price of an essential service NAT gateway? Each NAT gateway is $33 per month (without bandwidth charges) and generally if you want to be HA you need one NAT gateway per AZ.

I agree. Why is this an hourly charge? Same for endpoints in a private subnet.

Are there actual compute resources necessary to make endpoints and nat gateways work?

I assume the hourly charge is because NAT gateways are just EC2 instances, but they should be able to deploy using t4g instances and the monthly costs should be lowered significantly.

NAT Gateway and a bunch of our other fancy networking gadgets are based on a thing publicly called Hyperplane, and at the end of the day it's just EC2 instances. I'm skeptical that we'd ever use T series instances for this kind of thing for multiple reasons, but there are efforts to move internal systems to Graviton.

The reason for the hourly cost is a bit subtler than that, but we are working on the cost of this stuff.

(Source: Work at AWS.)

Thanks for chiming in, I had no idea that was backed by actual EC2 instances and the charges make more sense now.

So that means all external traffic in a private subnet with a NAT Gateway is routed through a single instance? I’ll have to read up more on Hyperplane.

Can you say what instance types are currently used for NAT gateways then? m5.large?

I'm a little removed from where I'd need to be to answer that very precisely, and probably shouldn't anyway.

They're not small- these are large, multitenant fleets handling huge numbers of NAT gateways all at once. The system has several layers, that scale on different dimensions and have different requirements. And those "requirements" can be very weird.

The part that actually moves packets around is probably mostly full-size c5n's, for the bandwidth.

You are singing my song here...

Yeah but were the prices sky high to begin with and still sky high now? That's the usual AWS hustle.

Does this mean running a VPC in each AZ connected via Transit Gateway is 100% cheaper than a single VPC w/subnets in each AZ?!

Regularly see EC2 interzone account for 1/3 of EC2 cost this would be a huge win if so (albeit totally ridiculous).

Reply on news.ycombinator.com