6 ms·

> other than: it makes migrating to competitors cost-prohibitive in a subset of cases My theory: it forces third party services into the same cloud. Suppose y

by amluto 3y ago

> other than: it makes migrating to competitors cost-prohibitive in a subset of cases My theory: it forces third party services into the same cloud. Suppose you use AWS and you want to pay a third party SaaS provider for some service involving moderate-to-large amounts of data. Here’s one of many examples: https://www.snowflake.com/en/data-cloud/pricing-options/ https://www.snowflake.com/en/data-cloud/pricing-options/ And look at this remarkable choice: you get to pick AWS, Azure, or GCP! Snowflake is paying a lot of money to host on those clouds, and they’re passing those costs on to customers. Snowflake is big. They have lots of engineers. They are obviously cloud-agnostic: they already support three clouds. It would surely be much cheaper to operate a physical facility, and they could plausibly offer better performance (because NVMe is amazing), and they could split the cost savings with customers. But they don’t, and my theory is that egress from customers to Snowflake would negate any cost savings, and the variable nature of the costs would scare away customers. So my theory is that the ways that customers avoid egress fees makes the major clouds a lot of money. IMO regulators should take a very careful look at this, but it’s an excellent business decision on the parts of the clouds.

Eridrus 3y ago
Snowflake's margins are like 90%+ on top of the compute they sell, and they pass on all these costs including egress directly to customers.
- amluto 3y ago
  Their list price for storage capacity is only on the order of 2x what S3 charges, and Snowflake and S3 likely both offer discounts. Comparing compute costs is harder. If I were running a service like Snowflake, I would certainly appreciate the effortless scaling that the major clouds offer. But I also know what I pay for actual servers and roughly what I would pay in a major cloud, and I much prefer the former.
  - Eridrus 3y ago
    Yeah, storage is much more competitive, but there have been leaks on what hardware they run on, their margins are incredible, which makes sense since they are selling software, not hardware! Anyway, there is an explicit egress charge: https://docs.snowflake.com/en/user-guide/cost-understanding-data-transfer https://docs.snowflake.com/en/user-guide/cost-understanding-...
  - dalyons 3y ago
    tbf, snowflakes whole thing & what makes them "next gen" vs a traditional datawarehouse is that you dont may much for data at rest. You pay as little as possible for the "warehousing" part, and pay mostly/only for compute/querying.
    - pas 3y ago
      wait, you still pay your S3 storage bills, right? just with DataBricks. you don't pay when you don't use their software, cool. but then you pay for per core per second for the license in addition to running the VMs, right? sure S3-like storage is 'cheap', but if you want up-to-date data pipelines you'll end up running these things for many hours each day. and plus you pay for each bucket operation, and so on.
- api 3y ago
  “Let’s get rid of IT and stop using on prem software. Outsource it all to the cloud and SaaS. We’ll save so much money!” … said the entire world pretty much … This is hilarious. Now that nobody makes on prem software anymore and nobody knows how to do IT, it’s time to turn the screws. The prices are only going up. You can’t move your data because it’s all locked into proprietary silos and only readable by one app running in one place. I told you so. So did loads of others. Oh well.
- infogulch 3y ago
  That's not the point. The point is that if the customer and SAAS provider "happen to" be in the same cloud then nobody pays any egress costs, but if the provider operated out of their own datacenter then the customer would pay additional egress costs to AWS to export their data to the SAAS provider. Snowflake is a good example because by the nature of the business customers send them a lot of data. This not only makes the customer pay more, but it also makes their billing and cost accounting much more complicated and unpredictable.
nyc_data_geek 3y ago
It's pretty simple. Excessive egress costs = vendor lock in, and yes, forcing third party services into the same cloud (the walled garden), and limiting customer choice. Just another reason so many orgs are getting heartburn from going too deep too fast into the cloud.
- echelon 3y ago
  Amazing: Cloudflare -- Free for most services OVH Cloud -- Free and unlimited Scaleway -- Free for most services Great: Hetzner 20-60 TB / mo per instance $1.08 Not bad: Linode 1-20 TB / mo per instance $5.00 Oracle Cloud 10 TB / mo $8.50 A bit much: Backblaze 3x the amount of data stored $10.00 Bunny CDN -- $10.00 DigitalOcean 100 GB - 10 TB / mo per instance $10.00 UpCloud 500 GB - 24 TB / mo per instance $10.77 Vultr 2 TB / mo for most services $10.00 Uh... Fly.io 100 GB / mo $20.00 Are you actually serious? Microsoft Azure 100 GB / mo $78.30 Amazon Web Services 100 GB / mo $92.16 Railway -- $100.00 Zeabur 10-100 GB, depends on plan $100.00 Google Cloud Depends on service $111.60 Screw you guys: Render 100 GB - 1 TB, depends on plan $300.00 Vercel 100 GB - 1 TB, depends on plan $400.00 Netlify 100 GB - 1 TB, depends on plan $550.00 (We use Netlify and have well over 1TB of monthly traffic. They're insanely expensive for what they are. As soon as we have roadmap time to revisit it, we'll move away.) I'm starting to think of cloud as less of an asset and as more of a liability. We can leverage them for temporary scale, but in no way will we tie ourselves to a particular vendor.
  - tomcam 3y ago
    I’m guessing you can negotiate for sharply reduced rates. No one pays list if they can help it.
  - phonon 3y ago
    AWS Lightsail is actually pretty good, comparable to Linode (except it maxes out at 7 TB/month). https://aws.amazon.com/lightsail/pricing/ https://aws.amazon.com/lightsail/pricing/
    - 3y ago
jupp0r 3y ago
It's an interesting point, but I doubt it's the lions share of egress that's going to other data centers vs to customers. Fan out is where it gets expensive.
snapplebobapple 3y ago
Its this plus locking customers to one cloud, because egress kills intercloud syncing on any moderately large data set. Any smart customer would have a duplicated setup across clouds if egress cost what it actually cost instead of 100x plus what it actually costs
quickthrower2 3y ago
So an anticompetitive practice.
- zer00eyz 3y ago
  No. Everyone knew what the costs were going in. AWS, cloud was NEVER the cheaper solution. It was never faster, it was never better. It was easier, and easier came with a price. Like lambs to the slaughter lots of people embraced it. Amazon is a profitable company because of it.
  - DeathArrow 3y ago
    >It was easier, and easier came with a price. I kind of doubt it's easier. If you use your own servers you need some stuff to manage them. If you use cloud, you need some cloud engineers to manage your cloud infrastructure. On top of that, if your developers use cloud APIs and frameworks, they have to learn that.
    - karolist 3y ago
      It's easier but only at the beginning, when the needs are simple and you dream of unpredictable QPS spikes that cloud will magically handle for you, summed up pretty well in https://world.hey.com/dhh/we-have-left-the-cloud-251760fb https://world.hey.com/dhh/we-have-left-the-cloud-251760fb
      - whizzter 3y ago
        I think the QPS dream is something that was teased at to show feasibility (and scaling up services on the cloud IS simple when needed). I think the bread and butter for the big clouds however are customers like the shop I'm at (see sibling comment), we can do development work and put it on our clients (mostly non-IT that has no interesting in hiring their own department) and if we for some reason need to part ways it's easy to "hand-over-the-keys" since subscriptions can be handed over between organizations. But getting customers back is usually equally easy since there was never any fuzz when leaving so when they've become dissatisfied with the "cheap" options they don't remember leaving being a problem, regardless the winner is the cloud provider (but the costs are still small enough for each client that nobody is unhappy).
Spooky23 3y ago
That’s definitely a factor. You can even pay AWS for the infrastructure used by SaaS apps to aggregate more volume.
ctrw 3y ago
It's not a theory it's literally why two projects I was on used snowflake in aws.
rfoo 3y ago
While I agree with what you said in general, Snowflake is a poor example. Data warehouses like Snowflake really can use the Tbps+ aggregated bandwidth between S3 and EC2 in the same region. There is no way for this to work over the Internet.
ren_engineer 3y ago
Snowflake and most other cloud services offer on-prem for these use cases, so it really doesn't make sense for them to roll their own data centers because it would be pretty niche. Cloud works for startups because they are there already, on-prem for enterprise customers with their own hardware

by amluto 3y ago

Eridrus 3y ago

Snowflake's margins are like 90%+ on top of the compute they sell, and they pass on all these costs including egress directly to customers.

amluto 3y ago
Their list price for storage capacity is only on the order of 2x what S3 charges, and Snowflake and S3 likely both offer discounts. Comparing compute costs is harder. If I were running a service like Snowflake, I would certainly appreciate the effortless scaling that the major clouds offer. But I also know what I pay for actual servers and roughly what I would pay in a major cloud, and I much prefer the former.
- Eridrus 3y ago
  Yeah, storage is much more competitive, but there have been leaks on what hardware they run on, their margins are incredible, which makes sense since they are selling software, not hardware! Anyway, there is an explicit egress charge: https://docs.snowflake.com/en/user-guide/cost-understanding-data-transfer https://docs.snowflake.com/en/user-guide/cost-understanding-...
- dalyons 3y ago
  tbf, snowflakes whole thing & what makes them "next gen" vs a traditional datawarehouse is that you dont may much for data at rest. You pay as little as possible for the "warehousing" part, and pay mostly/only for compute/querying.
  - pas 3y ago
    wait, you still pay your S3 storage bills, right? just with DataBricks. you don't pay when you don't use their software, cool. but then you pay for per core per second for the license in addition to running the VMs, right? sure S3-like storage is 'cheap', but if you want up-to-date data pipelines you'll end up running these things for many hours each day. and plus you pay for each bucket operation, and so on.
api 3y ago
“Let’s get rid of IT and stop using on prem software. Outsource it all to the cloud and SaaS. We’ll save so much money!” … said the entire world pretty much … This is hilarious. Now that nobody makes on prem software anymore and nobody knows how to do IT, it’s time to turn the screws. The prices are only going up. You can’t move your data because it’s all locked into proprietary silos and only readable by one app running in one place. I told you so. So did loads of others. Oh well.
infogulch 3y ago
That's not the point. The point is that if the customer and SAAS provider "happen to" be in the same cloud then nobody pays any egress costs, but if the provider operated out of their own datacenter then the customer would pay additional egress costs to AWS to export their data to the SAAS provider. Snowflake is a good example because by the nature of the business customers send them a lot of data. This not only makes the customer pay more, but it also makes their billing and cost accounting much more complicated and unpredictable.

nyc_data_geek 3y ago

It's pretty simple. Excessive egress costs = vendor lock in, and yes, forcing third party services into the same cloud (the walled garden), and limiting customer choice. Just another reason so many orgs are getting heartburn from going too deep too fast into the cloud.

echelon 3y ago
Amazing: Cloudflare -- Free for most services OVH Cloud -- Free and unlimited Scaleway -- Free for most services Great: Hetzner 20-60 TB / mo per instance $1.08 Not bad: Linode 1-20 TB / mo per instance $5.00 Oracle Cloud 10 TB / mo $8.50 A bit much: Backblaze 3x the amount of data stored $10.00 Bunny CDN -- $10.00 DigitalOcean 100 GB - 10 TB / mo per instance $10.00 UpCloud 500 GB - 24 TB / mo per instance $10.77 Vultr 2 TB / mo for most services $10.00 Uh... Fly.io 100 GB / mo $20.00 Are you actually serious? Microsoft Azure 100 GB / mo $78.30 Amazon Web Services 100 GB / mo $92.16 Railway -- $100.00 Zeabur 10-100 GB, depends on plan $100.00 Google Cloud Depends on service $111.60 Screw you guys: Render 100 GB - 1 TB, depends on plan $300.00 Vercel 100 GB - 1 TB, depends on plan $400.00 Netlify 100 GB - 1 TB, depends on plan $550.00 (We use Netlify and have well over 1TB of monthly traffic. They're insanely expensive for what they are. As soon as we have roadmap time to revisit it, we'll move away.) I'm starting to think of cloud as less of an asset and as more of a liability. We can leverage them for temporary scale, but in no way will we tie ourselves to a particular vendor.
- tomcam 3y ago
  I’m guessing you can negotiate for sharply reduced rates. No one pays list if they can help it.
- phonon 3y ago
  AWS Lightsail is actually pretty good, comparable to Linode (except it maxes out at 7 TB/month). https://aws.amazon.com/lightsail/pricing/ https://aws.amazon.com/lightsail/pricing/
  - 3y ago

jupp0r 3y ago

It's an interesting point, but I doubt it's the lions share of egress that's going to other data centers vs to customers. Fan out is where it gets expensive.

snapplebobapple 3y ago

Its this plus locking customers to one cloud, because egress kills intercloud syncing on any moderately large data set. Any smart customer would have a duplicated setup across clouds if egress cost what it actually cost instead of 100x plus what it actually costs

quickthrower2 3y ago

So an anticompetitive practice.

zer00eyz 3y ago
No. Everyone knew what the costs were going in. AWS, cloud was NEVER the cheaper solution. It was never faster, it was never better. It was easier, and easier came with a price. Like lambs to the slaughter lots of people embraced it. Amazon is a profitable company because of it.
- DeathArrow 3y ago
  >It was easier, and easier came with a price. I kind of doubt it's easier. If you use your own servers you need some stuff to manage them. If you use cloud, you need some cloud engineers to manage your cloud infrastructure. On top of that, if your developers use cloud APIs and frameworks, they have to learn that.
  - karolist 3y ago
    It's easier but only at the beginning, when the needs are simple and you dream of unpredictable QPS spikes that cloud will magically handle for you, summed up pretty well in https://world.hey.com/dhh/we-have-left-the-cloud-251760fb https://world.hey.com/dhh/we-have-left-the-cloud-251760fb
    - whizzter 3y ago
      I think the QPS dream is something that was teased at to show feasibility (and scaling up services on the cloud IS simple when needed). I think the bread and butter for the big clouds however are customers like the shop I'm at (see sibling comment), we can do development work and put it on our clients (mostly non-IT that has no interesting in hiring their own department) and if we for some reason need to part ways it's easy to "hand-over-the-keys" since subscriptions can be handed over between organizations. But getting customers back is usually equally easy since there was never any fuzz when leaving so when they've become dissatisfied with the "cheap" options they don't remember leaving being a problem, regardless the winner is the cloud provider (but the costs are still small enough for each client that nobody is unhappy).

Spooky23 3y ago

That’s definitely a factor. You can even pay AWS for the infrastructure used by SaaS apps to aggregate more volume.

ctrw 3y ago

It's not a theory it's literally why two projects I was on used snowflake in aws.

rfoo 3y ago

While I agree with what you said in general, Snowflake is a poor example. Data warehouses like Snowflake really can use the Tbps+ aggregated bandwidth between S3 and EC2 in the same region. There is no way for this to work over the Internet.

ren_engineer 3y ago

Snowflake and most other cloud services offer on-prem for these use cases, so it really doesn't make sense for them to roll their own data centers because it would be pretty niche. Cloud works for startups because they are there already, on-prem for enterprise customers with their own hardware