Spotify Teardown: Inside the Black Box of Streaming Music | Hacker News Reader

Spotify Teardown: Inside the Black Box of Streaming Music

mitpress.mit.edu

86 pointsunicornporn15 comments

'Spotify Teardown combines interviews, participant observations, and other analyses of Spotify's “front end” with experimental, covert investigations of its “back end.”'

It would be helpful to see a list of its contents/preview. I can't quite tell what this is about - it seems like it promises some technical detail which doesn't square with the authors' backgrounds?

Yeah, I'm not a fan of this post. It's a link to a product with promises of discussion of technical details without any actual discussion of said details. HN has thousands and thousands of amazing free sites to link to, why do we need this marketing tease? Just because it has the words MIT and Spotify and Front/Backend on it? I wish I could downvote this post.

I think that they are referring to the music publisher interaction when they refer to the "back end" rather than the code itself.

It does look like there might be some technical detail, but I’m not sure how much there will be:

> The authors engaged in a series of interventions, which include establishing a record label for research purposes, intercepting network traffic with packet sniffers, and web-scraping corporate materials.

Look forward to reading this!

There's always been a really interesting relationship between technology and music in Sweden, from the founding of the Pirate Bay (which Rasmus Fleischer, one of this books authors co-founded) to the creation of Spotify. It's been at the forefront of that world in the 21st century and it seems to be carrying on, there are a ton of music tech start-ups in Stockholm[0][1].

0: https://medium.com/@josephstockholm/the-swedish-musictech-la...

1: https://musictechfest.net/

the abstract reads like it is from the grievance studies department (albeit a mild one).

"innovative" then goes on to use the most common tools and add a lot of fancy words. Is this MIT stuff?

Are they still using P2P technology, or is music streamed from central servers/CDN?

I think they moved to central servers when their mobile apps came out, as they couldn't use P2P there.

Unfortunately, the article is behind the paywall. I'd really like to know how many reads their backend makes, when you shuffe around your library. I have already 3 dead SD cards on my hands and they only had spotify library on them. duh...

https://www.reddit.com/r/spotify/comments/5c6p0t/spotify_exc...

Not that many really.

Since the mobile apps and desktop app share a lot of code, you could if you wanted profile the desktop app in similar scenarios to guesstimate usage patterns. As long as you remember that the desktop app will also load more data (bigger screen, so more data to handle), it should be fine.

Somehow, you could compare it to a browser usage pattern. There's some index, some various files that are cached... My guess in your case is bad quality SD card or bad SD drivers / reader.

Then it puzzles me why 3 cards are dead. I suppose they do not optimize their shuffle to pick a song from an already read flash page, so many shuffles -> many read disturbs -> page refresh (move to a fresh page, mark old as garbage) -> garbage collection -> repeat 1000x -> card death. Maybe, this is also amplified by bad drivers and bad cards, as you've said. Sadly, none of these are open source. sigh...

> Unfortunately, the article is behind the paywall.

It's a book that hasn't been released yet.

oh...ty 4 the info. somehow I got the impression that this was an article.

Reply on news.ycombinator.com