Positing that Facebook purchased Instagram for its architecture strikes me as odd. They (FB) have solved all of the problems that Instragram faces and more for a userbase roughly twenty times as large, and while the infrastructure is interesting for those at companies aspiring to Instagram's success, for your average systems engineer at Facebook or Google I would think it seems pedestrian in comparison

I read that title as kind of tongue in cheek. I'm pretty sure Facebook could have built the exact same functionality for less money and in less time as it took to initiate and finalize the acquisition. Winning their users over may have been harder to do. I guess they just figured it would be easier to just buy the users. shrug

I can't see this as a user acquisition play. The number of Instagram users that are already on Facebook must be, what, 90%? Maybe more?

So, it's obviously not for talent ($100mm per head? no way...), users, or tech. That leaves it as a competitive play. They took Instagram off the market to keep it out of the hands of Apple and Google.

But to keep what out of the hands of Apple and Google? Apple and Google don't need the talent or tech either. Instagram's only real value was users. Competitive play or not... it had to be for the users. What else is there? Break room snacks?

Whether it's acquisition or retaining, every user owns an expensive phone, the demographic advertisers lust for.

Agreed. Also a pretty topical rebuttal to the other flood of articles claiming they were bought purely for the users - widely reported as being valued at $37 each, a number arrived at by assuming there's absolutely nothing else of value in the transaction.

Given the "prices" Google are rumored to be paying to acquire great engineering staff, a team of three engineers who've got a proven track record of building and supporting 100million users on their own - surely they'd be worth some not-down-in-the-noise-floor percentage of even the seemingly insane valuation here?

And a founding team who've "created" $1billion in value in ~550 days? Even if you only credit them with last weeks $500mil rumors - surely that founding team would be worth several tens of million dollars even if you didn't get the company they founded?

Paying $1B for a company partly because they've "created $1billion in value in ~550 days" is a bit of a tautological line of reasoning.

They also had ~50 million in the bank after a funding round that was so recent they could not have spent much of the money. Which means FB got a 5% discount on that billion dollar price tag and helps to make all those estimates just that much more reasonable.

Didn't Sequoia make a big YouTube investment suspiciously immediately before acquisition?

This is one user that will be deleting this off of his phone.

Darn-it!

Facebook definitely had other motives for doing this. One such speculation is here http://news.ycombinator.com/item?id=3819245

Others have been for their user base while some say it's for a talent acquisition (which I highly doubt). The title doesn't seem to imply that's why they bought Instagram.

I've got other suspicions too…

I wonder what Instagram's privacy policy says about things like the exif data on all your photos? Almost certainly all geo-tagged. A bunch of data like "how often and how quickly individual users upgrade their iPhones".

Cynical-me wonders how much value FB placed on the non-FB social graph that Instagram built when it hoovered up everybodies contact list?

are companies only allowed to have one reason for their acquisitions? i assume the complexity of instagram's architecture would reflect the technical prowess of the team and this would be very important to facebook

The article mentions they have a team of 3 engineers. You don't pay that much cash for 3 dudes who managed to boot some instances on EC2 and install an off-the-shelf open source stack.

You are depreciating their work in a way that is not that of a real gentleman. Maybe you are also working on a startup and would like to make it to where Instagram guys are? Me too, you know.

And for their tech stack what its impressing is that they chose very wisely IMO: if I had to build from scratch any Web or mobile app I wouldn't do differently (Python n PostgreSQL, etc)

He's being a bit flippant but his point is correct. You don't spend a billion dollars just to get three engineers, no matter how clever they are.

I don't know... These ones were clever enough to get someone to pay a billion dollars for them.

That's pretty clever.

Did the pay for the three engineers, or the thirty million users?

> You are depreciating their work in a way that is not that of a real gentleman.

I do the same exact kind of work. So am I depreciating myself? I may be brilliant and handsome to boot, but I don't think I'm worth 333 million dollars. Which was my point.

Oh sure, I might be worth several mil. I'm that good. But that's still a tiny fraction of what Facebook paid. So no, it wasn't about the brains.

There are stories around of Google spending 5mil or more to acquire (or in some cases, keep) great engineers.

I wonder what the real "market value" for a proven team of 3 great engineers who were the technical base that allowed a company to grow from nothing to $1billion in under 2 years?

It wouldn't surprise me to find those three guys as a team offered 20 or more million if the right company thought they were on the market.

Lars Rasmussen made out like a bandit when he went to FB a few years back, and that was after the "failure" (at least relative to Instagram) of Wave, and before the current "bubble" as well.

IMO, getting bought for $1B is not really the same as "growing to $1B". They may have grown to a ton of users and may have raised some VC money. But that is not the same. Did they even have a revenue stream? People with insane amounts of money overpay for shiny new toys all the time. Have you ever seen MTV Cribs? lol

They have 13 employees of which 3 were heading web ops.

3 web ops engineers for 30+ (before android release) million users & 100+ EC2 instances = WOW!

Amazon Web Services just got a huge shout-out today with this news. I self manage a large Web property that has 25+ AWS instances, plus some metal with ServerBeach and 1000TB.com. I also personally do all the development work on the site, as well as managing day-to-day operations and most support.

Yes, it is possible to run a multi-million dollar business by yourself with lots of servers, thanks to the cloud - in this case AWS.

Holy Cow! www.radioreference.com is that big already? Props! 570k users and 3k feeds = that infra? Streaming must be a bitch.

I love the instagram app and concept and execution, but I wonder why they even need 100+ EC2 instances (my guess is that almost all of them just serve photos). It appears instagram has little web traffic and it's not so heavily loaded with users 24/7 like facebook is.

http://siteanalytics.compete.com/instagram.com/

What you are seeing there is only the visitor traffic metrics for the landing page and the URLs shared out. You aren't seeing the back-end API centric traffic from the app to the core servers.

Twitter is in a similar situation, and according to compete data (inaccurate i know but vaguely indicative), twitter has more than 100 times more traffic, and they had ~70 servers in 2009 (http://www.quora.com/How-many-servers-does-Twitter-have). Just curious here.

I completely agree with your take on it; but pls note the entirely different problem Instagram solved vs. Twitter. If you build a real-time streaming application for pure text based objects, your stack will look very different. But collecting, storing, replicating and serving images as your primary social object != same thing.

Hence, even when Twitter 'added' photos, they went with Photobucket so as to not change the nature of their stack. It had taken them 3 years to really get ahead of their adoption curve for 'just' text.

You think API requests from all those mobile clients are some how "free"? Plus, I'd guess that they are resizing photos on the server.

Behold the power of the cloud.

So you're saying this is the most valuable Django application ever? :)

No. It's just the most expensive ;-)

I'd never trade my URL shortener or my podcast+connected-TV-backend for a photo sharing app.

This is a real feather in the Django community's cap. Eat that Rails

How long do I have to be on HN to get a downvote button? This isn't about Django vs. Rails or any other framework, it's about the stack that Instagram is built on, nothing more. There are no comparative metrics.

For some reason I was thinking Drupal the first time I read through the article... Drupal.

I'll be interested to see what happens with Pinterest (also Django).

How much does it cost to run it?

Facebook didn't buy them for anything like a billion dollars, $1000 plus the rest in shares is more like it.

Instagram's gamble on Instagram has become a gamble on Facebook (who are risking nothing of value).

I'm intrigued they went with XFS for the filesystem. Is there any more info on the thinking behind that decision?

They pointed out the thinking behind it. If you check out Eric Hammond's blog at Alestic, it expands on what was mentioned here:

http://alestic.com/2009/09/ec2-consistent-snapshot

Excellent, thanks for that. Looks mostly like a legacy decision which wouldn't necessarily hold today.

Except for the part where you want integrity; AFAIK only xfs guarantees truly atomic snapshotting thanks to xfs_freeze.

Furthermore, it's also (by an extremely large margin) your best choice if you need a filesystem bigger than 16 TB.

Allegedly other filesystems have had a similar freeze added since then.

Actually it was added to the vfs layer IIRC, around 2.6.30.

Why wouldn't? is there a better alternative to freeze a filesystem and take a AWS snapshot from it?

http://git.kernel.org/?p=linux/kernel/git/torvalds/linux-2.6... would indicate that the ability to freeze has been added to other filesystems since ec2-consistent-snapshot was first written. I've got no data on whether they're equivalently robust. If freezing was the only reason to go for xfs, then that decision would seem to be open to more alternatives now.

Is this 3 backend engineers? Was there another team that maintained the app and django app?

A not so interesting infrastructure that FB will probably overhaul. I think the author does not realize that facebook runs on custom server hardware, php/Hiphop, a bunch of open source stuff they contribute to like cassandra, memcached, varnish etc. The size of facebook and its platform is simply staggering. I 'm pretty sure their engineers are going to be hard to be impressed by these.

This architecture went far enough to get them acquired for a lot of money, so while it may not be a miracle of ingenuity, it's worth studying as an example of something that worked.

Or, we've learned that even the biggest web companies can now be built with fairly vanilla technology and that obsessing over clever bespoke scaling schemes is largely a waste of time.

I think it'd be more accurate to say "Even /fairly large/ web companies can now be built [et cetera]". Good luck putting Google on AWS. Or (relevant to the name of this website) AWS.

I think now you could even build a fledgling Google on AWS. Once you start getting real traction then you may outgrow it but there's really no point in even talking about those steps until you have a few million users beating down your door.

... in the sense that the typeface of famous books is worth studying as an example of how to write a good book?

It does raise the question of whether you even need any real ingenuity on the back end for this sort of startup...

Thinking that you need ingenuity is often the root of problems.

you probably need some to create it though

That's a long list of open source projects. I wonder did Instagram contribute to any of them, or help out in any way?

I am going to guess "no". For three people to scale a site to that size, I imagine it was a daily battle just to keep the thing flying. Not to mention, if they ran into a problem that could potentially be solved with some deep magic done on a forked OSS project... or another ten extra large instances, given their valuation I can imagine which direction they would lean.

three engineers and a cloud? sure beats three yards and a cloud of dust...