Implement browser compatibility

[tribou]

Hello!

I'm coming from the direction of testing y.js for a web-based shared document editor. However, I started looking into your architecture for secure-scuttlebutt, and I really like what I've found so far.

I was curious, is there a browser implementation available or planned for this version of ssb? I ran across dominictarr/localstorage-scuttlebutt along with dominictarr/r-edit, but I'd hate to forego the latest end-to-end security architecture you've been working on lately. Is it possible to run this version of the database in the browser?

[dominictarr]

first, there is "secure-scuttlebutt" and "scuttlebutt" they are based on the same basic idea of a protocol, but otherwise, completely different. those other ones work with scuttlebutt, which isn't secure per se. I am intending to get this working in the browser at some point though, and some sort of mutable document representation on top of it, but not really planning on updating "scuttlebutt"

you might also be interested in http://github.com/dominictarr/epidemic-broadcast-trees it's also about the part essential protocol part to scuttlebutt, but makes it more efficient and scalable!

[tribou]

Thank you for the clarification! I'll check out epidemic-broadcast-trees 👍

Regarding your intentions to get secure-scuttlebutt working in the browser, perhaps we could use this issue to track the progress? I'd be glad to help out where I can. Could you provide a high-level list of steps still needed for that to happen?

The idea that I'm chasing is to create a better opportunity for code-sharing between React, iOS/Android React Native, Electron, and an event-sourcing/AWS Lambda backend. I'm thinking the backend may need some custom logic to provide an initial authentication handshake and persist a snapshot of the state. Anyway, right now I'm just looking for next steps to get this package closer to the browser. Thanks for your help!

[dominictarr]

Well, the main hurdle is getting https://github.com/flumedb in the browser. I am working on this at the moment, actually. the main problem is getting the indexeddb wrapper fast enough to be reasonable. I need to be able to load 100k records (100mb or so) into indexeddb in under 5 minutes.

[christianbundy]

This issue is my main priority -- my daily driver is a Chromebook, and being able to use it in a mobile browser (or app) would be a total game-changer. If there's anything I can do to help this along, I'd be more than happy to throw a few hours at it. Thanks!

[dominictarr]

oh awesome! so the biggest thing single obstical is to get the flume (database) stuff working well in the browser. I want to be able to load say, 150mb of data in at least 10 minutes. that really isn't too much data, but indexeddb (our only storage option, basically) is really quite slow.

From what I've looked into it currently, the best thing is build this api: https://github.com/flumedb/aligned-block-file on top of indexeddb. I also tried going into indexeddb directly, https://github.com/flumedb/flumelog-idb but it's slow with many small writes. It's fairly fast with large writes, which is why I think the best option might be to write a block store on top of indexeddb.

[mlegore]

Is there a fork somewhere of this WIP?

[Rob-pw]

@mlegore, I believe at: https://github.com/ssbc/secure-scuttlebutt/tree/bump-flumedb, I've not yet tested it.

[Atavic]

Decentralization Protocols on Firefox 59: pyllyukko/user.js#366

[stale]

This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.

[tribou]

I still would love to see a browser implementation, but it sounds like we may need a different way of architecting it to make it performant?

[christianbundy]

Me too! Unfortunately I don't think this is something that we can resolve without replacing flumedb or making significant changes upstream, so I wonder whether this would be better tracked in the flumedb repo. What do you think?

[stale]

Is this still relevant? If so, what is blocking it? Is there anything you can do to help move it forward?

[regular]

ssb-browser-core got this covered, I think?

[tribou]

Nice, thanks for sharing. Would love to see it added to the ssbc org. Regardless, looks like a huge step forward in available browser APIs!

[regular]

It is! Very exciting! cc @arj03

[stale]

Is this still relevant? If so, what is blocking it? Is there anything you can do to help move it forward?