Scaling Instragram

http://speakerdeck.com/u/mikeyk/p/scaling-instagram

by tferris • 14 years ago

112 35 14 years ago

tferris• 14 years ago

Instragram is the modern fairy tale showing that any "2 product guys" "with no real backend experience" are able to rise from rags to riches.

DavidAbrams• 14 years ago

And also that the bubble is back, big time.

ma2rten2• 14 years ago

I am starting to get a little bit tired of people screaming "bubble" whenever someone mentions Intragram. Could you make a good case for it?

If one particular person (Zuckerberg) values one particular company higher then most other people, I don't see that as a sign for a bubble.

jaredsohn3• 14 years ago

Instragram -> Instagram (the title was wrong)

apu• 14 years ago

On slides 61-69, it seems to suggest that they were storing their photos directly in the database. Is this correct? It seems rather inefficient to do this, rather than storing the images in the filesystem.

seanmccann• 14 years ago

No, photos are stored on S3. Only photo meta data is stored in the db.

rplnt• 14 years ago

Is there a video/audio for this talk?

jaredsohn3• 14 years ago

Submission for a TechCrunch writeup of the talk, which includes some quotes and also embeds the same slides (via scribd):

http://news.ycombinator.com/item?id=3831865

jaredsohn3• 14 years ago

This talk was given at AirBnB in SF. Talks there are fairly new and typically occur every other Wednesday. There is a talk tonight by Yehuda Katz and upcoming talks are posted here: http://www.airbnb.com/tech_talks

zdgman• 14 years ago

Wasn't this same article up on HN the same day they got acquired? Lot's of interesting stuff in here in general, especially when they get to talking about when their app actually uploads a photo.

marcusf2• 14 years ago

Slide 54 is insanely important. At my current place we've invested tons in instrumentation and metrics. Every IO operation is guarded with a semaphore and access is monitored and exposed. Every cache in the system logs everything from hit and miss rate to volume etc. Every async operation logs operation trails. We also ship a ton of tools for everything from bread and butter stuff like stack dumping, pinpointing long running threads etc. to more sophisticated stuff like being able to break in to a page rendering and profiling various element hierarchies, either over time or following specific users.

As I said, we've invested so much in getting the entire stack monitored very granularly, but every time our customers run in to issues it pays dividends.

For reference, we run on the JVM which is very pleasant. A ton of tools interop with it and basic things like MBeans being integrated in to JVM makes monitoring a lot easier.

SanderMak• 14 years ago

How much overhead does this setup incur?

marcusf2• 14 years ago

Run time or development time? Runtime is mostly negligible except for places with heavy congestion (caches + rendering) where you can turn the heaviest operations on and off as needed -- in our case the call to `System.getCurrentTimeMillis()` is usually too expensive to do unless we really need it.

In terms of development, it's not a lot of work continuously, but it took some while to write up, test and benchmark standard libraries for eg monitorable semaphores. In general though, what it does is raise the entry threshold for new developers a bit as you need to have a feel for when something requires monitoring and when it doesn't.

jvehent• 14 years ago

I wish each and every single one of the open source developer who, every day, build amazing tools to make things like instagram possible would receive a tiny portion of that 1 billion dollars.

But that's not how the world works....

The real revolution of this century is not how many sheeps I can throw per day, but how huge a success the open source movement is, and what incredible companies we can build with it.

famousactress• 14 years ago

That'd actually be a really cool pledge to make as a company. Sorta like the 1% for the planet initiative, but something similar to pledge a slice of profits or revenue, or exit-event-cash to the OSS projects you depended on most.

d503• 14 years ago

A few companies do actually, including DuckDuckGo and Tarsnap.

http://fosstithe.org/

http://www.gabrielweinberg.com/blog/2010/11/help-me-start-a-...

unwind• 14 years ago

Won't somebody please think of the nit-pickers, and fix the typo in the title? At least it's only in the submission to HN.

rabidsnail• 14 years ago

Audio: http://soundcloud.com/bob-poekert/tech-talk-mike-krieger-of

Scaling Instragram

Related Stories