Hacker News Reader

Comment by eslaught | Hacker News Reader

What I find interesting about this is that the Google Drive is now significantly less expensive than Amazon S3, and even comparable to Amazon Glacier.

    Google Drive: $10/1000GB = $0.01/GB
    Amazon S3: $0.85/GB for the first 1 TB
    Amazon Glacier: $0.01/GB
    (all prices per month)

This means that any service built on top of Amazon S3 will never be less expensive than Google Drive. Period.

Sure, there are caveats. Google Drive is not programmatically accessible like S3. The terms of service probably disallow the same use cases as S3. The pricing of Google Drive almost certainly depends on users not using all their storage. Google may rate-limit heavy users or do something else to discourage people from fully utilizing the product. etc. But the point still stands: 3rd party products built on top of Amazon's infrastructure will not be able to beat this price without significant competition from Amazon.

View on HN

> What I find interesting about this is that the Google Drive is now significantly less expensive than Amazon S3

Which shouldn't be too surprising, since the (more expensive than Drive) Google product that is more of a direct competitor to S3 (Google Cloud Storage) is still an order of magnitude less expensive than S3 (Cloud Storage is $0.085/GB for the first 1TB.)

EDIT: Well, actually, that's not true; parent's cited prices for S3 were wrong. GCS and S3 are, at least for the first TB, exactly the same price at $0.085/GB.

> This means that any service built on top of Amazon S3 will never be less expensive than Google Drive. Period.

That depends. With drive you pay for allocated capacity, with S3 don't you pay for actual usage? Depending on the level of overcommit, its possible for less expensive per GB "capacity" priced service to be profitable on top of a more expensive per GB "usage" priced service.

Plus, its at least theoretically possible, especially if a lot of the stored content is not efficiently compressed on its own, that the consumer capacity-priced service could also use compression and/or deduplication to further reduce usage on the usage-priced backend.

> Google Drive is not programmatically accessible like S3.

Google Drive most assuredly is programmatically accessible via an API.

> > This means that any service built on top of Amazon S3 will never be less expensive than Google Drive. Period.

> That depends. With drive you pay for allocated capacity, with S3 don't you pay for actual usage? Depending on the level of overcommit, its possible for less expensive per GB "capacity" priced service to be profitable on top of a more expensive per GB "usage" priced service.

Normally I would agree, but the price gap is so large (close to 10x), you could easily downgrade to the next smallest capacity level in Google Drive. You would only actually lose money if you wanted to store more than 100 GB but less than 10/0.085 = 118 GB. That means for most overcommit levels, you're still winning by a reasonable margin.

> Plus, its at least theoretically possible, especially if a lot of the stored content is not efficiently compressed on its own, that the consumer capacity-priced service could also use compression and/or deduplication to further reduce usage on the usage-priced backend.

Compression is unlikely to yield enough, but perhaps deduplication would save you. Still, 10x is big gap to make up for.

> > Google Drive is not programmatically accessible like S3.

> Google Drive most assuredly is programmatically accessible via an API.

Admittedly, I'm not too familiar with the Google Drive API, but my impression is that it is geared toward entirely different use cases (compared to S3-like storage services). Could you build a business on top of it?

Amazon S3/Glacier don't just provide some random cloud storage, but availability from specific locations, guaranteed bandwidth/rate, detailed logging, API, etc.

While Google Drive might be an interesting place to just store "some stuff", it's not really comparable for any larger scaled application.

I'm not sure what bandwidth Google Drive provides, but looking at Dropbox, I'd never want to serve any customer with that "crappy" < 1 MiB/s speed. It's okay for all the small changes Dropbox is designed for, but for serving as public storage space, it's not really acceptable.

You are assuming that you are using your entire allowance in Google drive.

Mmmm... kind of not.

You forget that the first 15 GB of Drive is free.

You don't get the first 15 GB of S3 (for each of your users) for free.

Isn't Amazon S3 $0.085 / GB for the first TB/month (1/10 of what you've said)?

That's still 8.5x Google Drive. Being competitive with glacier jives with what I've heard the underlying cost of storage hardware is (still with room for profit). However, from what I hear running the request serving on top of that is fairly expensive (which is why glacier has big restrictions).

AWS says they're a cost plus org which means they drop prices as their costs drop. But that plus can be quite large.

Being #1, first to market, etc. has a big advantage. Despite knowing (or believing) these things, my org still happily spends plenty at AWS.

Google Drive has an API but it's certainly not meant for S3 use-cases serving content on websites :)

https://developers.google.com/drive/v2/reference/

When you talk about Glacier, you must remember there's no instant access to files and there's bandwidth costs associated with downloading your files. The price for storage may be the same but the overall costs of Glacier are much higher.

Deduplication.

This is a big reason why service like Dropbox and Google Drive work so well. If two or more people have the same exact file in their drive, it doesn't take up more physical storage than the original file.

Reply on news.ycombinator.com