IPFS Scaling and Wishful Thinking

There is something to be said about the shooting star projects in the open source world. The kind of projects that somehow get better and better, and yet don't rely on crazy funding. Like 7zip, for example.

But they are the exception that cements the overall rule.

I think I first heard of IPFS back in the Steem days, maybe around the time DTube came about, maybe earlier, but not by much. I thought the idea was amazing back then, and although I did not grasp it completely, my participation in torrents made it somewhat understandable.

Seems obvious to me, even more so as I've gotten to play with IPFS for going on two years now, that although the promise is excellent, there are definite pitfalls we can't ignore forever.

The short of it is this:

IPFS has a scaling problem.

A small project, like storing NFTs for a game, seems to be the ideal use case. But I think all of us wanted, and I include myself here, for IPFS to rival industry standards. To stand toe to toe with S3, with AWS-type solutions.

Sadly, and I mean this sincerely, the biggest player in the game, the one company that figured out how to build a giant business model on top of IPFS, Pinata, is only able to do so because they are blending IPFS technology with centralized solutions.

To call Pinata IPFS is to describe half of it, and thus they prove my point yet again.

Now, it might be obvious to some, but it's worth mentioning that their solution to this scalability problem is proprietary. It's not like I can email them and get them to share their private repos with me.

But it's not too hard to imagine how they've solved it.

What is hard is building something like this from scratch.

Yes, even today with AI capabilities in play.

I'm sharing this because I think we are reaching a pivotal moment on 3speak.tv. Seems to me that some big changes have to happen, and somewhat soon too. More so if we want to move towards the big audiences.

Add to this little rant of mine the fact that Shipyard, the team who was behind the maintenance of IPFS' ecosystem, just called it quits.

I think the path forward is unavoidable.

After all, as I said in the beginning, there are only a handful of projects with tiny, decentralized, unfunded teams that actually deliver the goods.

And building on top of wishful thinking seems irresponsible.

MenO

Comments · 8

@pizzabot(60)· 22h
PIZZA!

$PIZZA slices delivered: @caparadiso_(4/5) tipped @meno

_{Please vote for pizza.witness!}
@caparadiso(60)· 6d
A very honest and well-argued post, @meno, thank you! 👏 I like how you start from the "shooting star" projects like 7zip and then show why IPFS is a harder case: the promise is excellent, but the gap with S3-type services is real. The Pinata point is striking: if even the biggest player needs centralized pieces behind the scenes, we should be careful with the word "decentralized". 🤔

It matters a lot for 3Speak. As a former IT guy (40 years) and now a 3Speak creator, I read your point about "big changes soon" as a call for realism. The news that Shipyard, the team behind IPFS ecosystem maintenance, has stopped makes the risk even clearer: building on "wishful thinking" isn't a plan. 🚧

Question: if you could choose, what would a realistic path look like for 3Speak, a hybrid model (IPFS plus centralized CDN) or a different storage layer altogether? And what can creators like me do in the meantime to help? 🚀

Matteo - Ca' Paradiso, Cagli 🌿 www.caparadiso.com
@mortezayousefi(68)· 6d
It is better to use your own dedicated servers
@harvestmoon299(57)· 6d
Please elaborate on the scaling problem! Maybe use IPNI to increase lookup speed? Of course, we need more participation with the pinning.
@epodcaster(67)· 6d
If you had a magic wand, what would the ideal decentralized solution be?
@alonicus(72)· 6d
Does 3Speak run on IPFS right now, and if it does, what are the implications for current levels of functionality ?

I must admit I'm something of a decentralisation purist, and have an especial dislike and distrust of Amazon/AWS even though I understand they can be a necessary evil.

It's a shame that we rely on nodes that are too complex for normal users to deploy and manage. It would be great if it was possible to create something that was more like a desktop app where you download it, say how much storage space you want to give to it and the maximum data usage (to avoid tripping over ISP "fair use" T&C's), and then everything else is handled in the background and it just delivers some rewards to your Hive/Hive engine wallet.
@steevc(81)· 7d
I've heard of this stuff and that it has costs. Video just creates huge files and storage can be expensive. You will never want it on the blockchain. Will you be mapping out the way forward soon?
@bozz(83)· 7d
I first heard about IPFS back when I was investing heavy in NFTs on WAX.