Credit to @houlihan_rick; he's not entirely wrong. For *most* customers, data transfer pricing falls within a reasonable, expected band.

For a relatively small percentage of customers, this doesn't hold true.

The problem is that it's impenetrable and unpredictable.
It's an embarrassment to @awscloud's data transfer pricing both that I had to create this chart, and that I have to refer to it so frequently. Image
"Pricing reflects our costs" and other lies providers tell me.

Sending data into AWS from the internet is (generally) *FREE*.

Sending data out of AWS costs (generally) more than storing that data in S3 for 3 full months.
The way everyone finds out what data transfer is going to cost them is they run a workload and find out through experience.

This is a problem for discoverability, for predicting the costs of a workload, and for understanding how to structure for cost.
I'd rather understand that there's a cost consequence for things talking same AZ vs. cross-AZ and structure for that in my architecture when I build it, not have to retrofit that in later.
And then you squint at the corner cases and start asking hard questions. "Why is it more expensive to send data from one AZ to another in the same region than it is to send that same data from Virginia to Ohio?"
It's arcane, and bizarre, and it's a service you basically can't opt out of. You can get full line rate between basically any two points within AWS and that's damned near magic. But many workloads don't care about speed, they care about price.
In my datacenter I don't have to know the answer to things like "how many gigabytes do my application servers send to my database servers every month."

If my company values pricing predictability over lowering the price (many do!) then staying on-prem is the right move!
Let's be very clear here--Netflix didn't build their OpenConnect program for funsies (openconnect.netflix.com/en/). They (and anyone who wants to do streaming video at scale) are failed by AWS's data transfer pricing.
These are hard problems, and there aren't easy answers.

But I didn't build my business by spreading FUD about data transfer pricing. I discovered the horrors of data transfer pricing by looking at what was going on in my customers' accounts, and diving deeper.

• • •

Missing some Tweet in this thread? You can try to force a refresh
 

Keep Current with Corey Quinn

Corey Quinn Profile picture

Stay in touch and get notified when new unrolls are available from this author!

Read all threads

This Thread may be Removed Anytime!

PDF

Twitter may remove this content at anytime! Save it as PDF for later use!

Try unrolling a thread yourself!

how to unroll video
  1. Follow @ThreadReaderApp to mention us!

  2. From a Twitter thread mention us with a keyword "unroll"
@threadreaderapp unroll

Practice here first or read more on our help page!

More from @QuinnyPig

4 Mar
The @okta news is super great and all, but every time I hear their name I’m reminded of the time I had to sign an agreement not to attempt to hire their staff for a year in order to enter their office to speak at a meetup.
This was their “doorway NDA” and is almost certainly unenforceable, but it annoyed me something fierce.
Years ago I had to turn a meeting at @AppDynamics into a meeting at the nearby coffee shop for the same reason because I was at that time actively counseling one of their staff to leave.
Read 5 tweets
4 Mar
A third day, with my thoughts on consulting. This one is probably the most controversial, so let's get to it.

I don't think the best move is to do implementation work. A thread...
Hear me out. In the first thread I talked about finding a positioning that works as "the expert in An Expensive Thing."

In the second, I talk about value based pricing instead of hourly.

So you're now a very expensive expert here to solve a big expensive problem.
In the technology industry, I maintain that implementation opens up Pandora's Box of delivery risk, while simultaneously damaging the perception of your value.
Read 14 tweets
3 Mar
So in that NYTimes profile of me, @daiwaka wrote "Mr. Quinn said Amazon had never tried to rein in what he said."

That's not ENTIRELY true, and I do want to be fair to @awscloud.
While @awscloud has never once tried to stop me from publishing anything incendiary, or urged me not to deliver a Hot Take, they jump with a *QUICKNESS* when I say something that they perceive to be factually incorrect.
Their approach can best be summed up as "shitposters gonna shitpost, but the second it confuses a customer that is A Problem."

And what's more is, they're right. A few examples!
Read 7 tweets
2 Mar
As threatened, today I'm going to talk about pricing. This one will likely result in... feedback.
So, as you've probably heard if you've worked with, met, passed on a street, been in the same coffee shop as, or just had a dream about @jonathanstark, "hourly billing is nuts."

I used to disagree vehemently. I no longer do.
Unless you're an attorney, people are going to cap out in terms of what they pay you somewhere around $250-$300 an hour.

That's good money, right? $600K a year assuming 40 hours a week. Let's start there.
Read 30 tweets
2 Mar
"Join now to watch the Microsoft Ignite keynote!"
There we go. And as an apology for the trouble, @Microsoft has disabled the social profile nonsense.
What kind of crapass "cloud" conference is this?! They've got pre-keynote streams so you feel engaged and involved, they bother to *mention* the virtual sponsor expo hall so customers know it's there...

MS has so much to learn from AWS's approach to half-assing things.
Read 32 tweets
2 Mar
My personal guide for burned-out employees with chips on their shoulders. I recommend none of these. I am guilty of all of these. This is why I'm a terrible employee.
Put expenses on your own credit card and then submit them. If you experience pushback, stare them dead in the eye and say "okay, so don't pay it." See if they call your bluff.

(They almost never will.)
If someone asks you to work late tonight, you have plans. Maybe a date with your spouse. Maybe playing video games. Maybe you plan to cry yourself to sleep. Not their business; they're your plans.

Emergencies aren't "someone else fucked up the planning."
Read 18 tweets

Did Thread Reader help you today?

Support us! We are indie developers!


This site is made by just two indie developers on a laptop doing marketing, support and development! Read more about the story.

Become a Premium Member ($3/month or $30/year) and get exclusive features!

Become Premium

Too expensive? Make a small donation by buying us coffee ($5) or help with server cost ($10)

Donate via Paypal Become our Patreon

Thank you for your support!

Follow Us on Twitter!