Btw, Google already has a BlobStore API which is in beta. Maybe they're releasing it at the conf. If so, not big news. It's the exact same price as Amazon and has a clunkier API (IMHO).
Competition is sorely needed in this space. I hope Google undercuts Amazon pricing. IIRC, Amazon has maintained the 15c per GB-Month pricing for 4+ years. I would have expected them to pass along at least some of the price reductions of bare HDDs.
I'm hoping for something like the $5 for 20 gig they offer for photos. I'll keep dreaming. :)
It's an App Engine API, so unless you're running your application on App Engine you're out of luck. Of course, you could write a wrapper around it to make it accessible over a RESTful API, but what's the point now that Google Storage has been announced?
This can only be good. More competition will probably mean better prices, more features, higher reliability and more innovation. And if someone can compete with Amazon on this, it's certainly Google.
My first guess would be that Google has a lot more cool things to do with its datacenters than Amazon, so up to now any extra capacity was used for a new product/project. Only now did they have enough spare to start leasing it to outsiders (which either says something about Moore's Law, or about a decline in the rate of innovation at Google).
I hope the bit about easy migration from S3 means a compatible API, just change the base URL. Would that make the S3 API the de facto standard for cloud storage APIs?
Eucalyptus also uses this interface. It would seem that it is becoming the de facto standard. Being able to immediately use all of the existing tools would go a long way for adoption.
"Google To Buy Dropbox For An Undisclosed Amount" does not sound far-fetched to me given this announcement. From a strategic viewpoint it makes perfect sense.
It would provide their new storage service with the necessary intial momentum at the expense of S3.
They've taken investment. They are by definition interested in either a) selling or b) going public. While I would happily buy shares in a DropBox IPO, I'm afraid option a is much more likely.
for the right price everyone is willing to sell out. And by looking at the Google acquisition history they don't mind paying big bucks for a great service.
App Engine and EC2 don't really compete in my mind for this reason. App Engine seems to compete more with things like Heroku, although the interface complaint certainly applies there too.
I find App Engine and Heroku to be in stark contrast in this respect.
App Engine requires you to code your app in a particular way to work on their platform, and that way will (to a large extent) only work on App Engine.
Heroku, on the other hand, requires you to code your app in what is pretty much the recommended way to do Rails apps anyway. While they do impose a few limitations that aren't present on other platforms (e.g. no local writable storage), accommodating those limitations wouldn't prevent you from moving to another platform. There's no proprietary Heroku Database API that you'd have to reimplement in order to move to a physical-server-colo setup, for example.
App Engine requires you to code your app in a particular way to work on their platform, and that way will (to a large extent) only work on App Engine.
This is true, but it does make for a very nice deployment story; and Ian Bicking's silver lining (http://cloudsilverlining.org/) is (I think) a very interesting attempt to tell that story in a more generic way.
Google is the industry leader at monetizing search, but nothing else they do comes close.
I am truly curious - is the reason that Google keeps investing time and money in "line-extension" businesses to distract competitors, keep the antitrust police away, what?
Google is the industry leader at monetizing search
Google's stated mission is to "organize the world's information".
There is tremendous power in information. Not so much in your own, but on the correlation of your data to everyone else's; the collective statistics about anything is where the power lies.
In gmail, google reading my email tells them to sell me product x and we think that's the end of it. But there is so much more available in terms of how the aggregate masses behave with regards to product x and its correlated points.
This extends right into file storage. Yes I am saying that Google will "read" your data. Even better, I'm saying that Amazon more than likely is already doing this.
They're doing it not because they give a damn about you, but because they want the data points that come from the analysis of many separate instances of unrelated bits and bytes. That analysis is worth a fortune. It's a modern day gold rush.
They've already got a ton of storage for end users - I imagine Gmail attachments alone are a pretty massive corpus.
What this adds is information about site traffic - who's using what from where. Between Google Analytics, search trends, and toolbar, they've already got a lot of that. But I imagine that's part of the rationale for this.
"In gmail, google reading my email tells them to sell me product x and we think that's the end of it."
I don't want to put on a tin-foil beret, but how much revenue do they actually make from ad-words on the side of GMail? I can't imagine it's much, but then I'm probably biased in that I don't even notice they're there.
The pragmatic part of me realizes GMail is a service that improves 'lock-in' to the 'Google Ecosystem'... the paranoid part of me thinks about how valuable a "search everyone's inbox" rather than a "search my inbox" feature would be.
Yes I am saying that Google will "read" your data. Even
better, I'm saying that Amazon more than likely is already
doing this.
It would be an unforgivable breach of trust for Amazon to steal customer data that is stored in AWS. Amazon is possibly the world's most customer-oriented company. I'm sure that such behavior would not be tolerated inside the company.
well actually I think this is the right direction for them, they lost a lot of time trying to catch up in things were they can't win, like buzz or alike.
Apart from the business perspective, at least this is one thing where they can be strong, and able to match the expectation of users.
Also I can't see how this is different from Amazon, business side: they got an infrastructure in order to run their other business. But there is no point in using this just internally, if this can be provided as an high quality service. Looks sane to me.
Why was this downvoted to zero? Asking to improve in the future.
My point was that line extension has been shown time and time again to be toxic for businesses and Google's bevy of marginal non-search projects seem only to reinforce that this is accurate.
Al Ries has dedicated an influential book to the subject called "Focus" and I've seen the damage line-extension can cause first hand in businesses I've worked in and invested in.
Considering the variety of points of view this comment spurred, I was suprised to see it = 0 which tells me that HN doesn't want this type of comment in the future?
There will be a point in time where they either don't have the same dominance in search or that it will become increasingly irrelevant. Given the money that they are making they can afford to stick an arm into everything just to see if there is a big business there, all the while helping the core business with more data.
Much better strategy I thing rather than the exploiting cash cows until they dry up, which a lot of business does.
I think AppEngine and storage are developer bait. What's notable in the TechCrunch article is bit about authentication via GoogleID. Let's say Google entices 100K developers to provide applications of some sort which rely on a Google ID for auth. A huge segment of internet users would acquire a Google ID _AND_ recognize that they have one. So, Google becomes a de facto record of the world's identity (much like Facebook?). What sorts of services could be built on top of this? How much is that worth?
The first is that it commoditizes the Internet further, which means more people will use it to look at Google ads.
The second is, as other people have said, that they can potentially get valuable data out of it. It's actually rather scary how few web sites Google is not able to keep track of, if you consider both ads, analytics and search traffic.
That is very true. Google's interests are very simple to fathom: They want more people on the internet. Anything that helps (in this case making it easier to develop internet oriented products) towards that is a worthy product to develop.
Because I guess they see 'cloud computing' in their future and generating money from that. I guess they need some kind of beefier storage service as they start making more ground with their Google App product line.
They've already developed the vast majority of the infrastructure needed for this product in the course of working on their own projects. Why not start selling it? It's pure gravy for them, and it eventually means even greater economies of scale for innovations in power consumption, horizontal scalability, etc.
Google has gotten so big on the web, anything they can do that will increase the usability, adoption and speed of the web may help their bottom line. In this sense, it may be the same strategic initiative as getting into the ISP space and offering Chrome.
Interesting! It had just occurred to me that I could write a competitor to s3 using app engine and blobstore. I guess someone at Google realized that too.
This is what I had thought about. If you don't care that the API is published/sorta-standard, it's not a big deal to deploy a lightweight pass-thru app to AppEngine.
The pricing model is what will be most interesting about the announcement.
Could not agree more. Video encoding is still damn expensive. We're using encoding.com and although they have great customer support, we're still paying roughly $1/video to encode. 5% of these videos will account for 95% of the views.
Why pay $1 to encode an unpopular video then $0.001 for the bandwidth?
Encoding is the only piece of our video model that currently doesn't scale directly.
A hacker I know (who has I team I believe - the debuggable.com guys - both successful bloggers and core teamembers of the cakephp community) has been working on transload.it for a while, it's a pretty good service despite still being in beta, but I haven't put it through its paces like I'm sure you would. (I was working on a wordpress plugin.) It sits on ec2 and will save right into your s3.
We actually offload right to Akamai so the s3 integration isn't a big deal for us, but anything to drop the price of bandwidth or encoding is a win these days :(
Tho I must say the encoding guys are awesome so it will take a damn good product to make me leave them.
I am very confused. You can rent a 12GB RAM 8-core Nehalem system for under $400 per month - surely such a system can encode much more than 400 videos in 30 days?
Correct. 400 videos in 30 days is no problem. However, video encoding is purely a 'peak' architecture issue. If your videos take longer than 10 minutes (give or take a few) to encode, they leave or complain.
This is fine at 4am but during peak times, we see 100-200 videos uploaded in the course of 5 minutes... and we're a small video 'network.'
During off-peak times, we see maybe 5-10 videos per hour. So it doesn't really help us to have a full time computer or four.. we've gone down this route. The 'cloud' tech is really the best way to do it.
Competition is sorely needed in this space. I hope Google undercuts Amazon pricing. IIRC, Amazon has maintained the 15c per GB-Month pricing for 4+ years. I would have expected them to pass along at least some of the price reductions of bare HDDs.
I'm hoping for something like the $5 for 20 gig they offer for photos. I'll keep dreaming. :)