Peer5/ShareFest:: Web based p2p file sharing using WebRTC Data Channels API

[tomByrer]

https://github.com/Peer5/ShareFest
"One-To-Many (mesh network) sharing application. Serverless. Eliminates the need to fully upload your file to services such as Dropbox or Google Drive. Put your file and start sharing immediately with anyone that enters the page. Pure javascript-based. No plugins needed thanks to HTML5 WebRTC Data Channel API - http://webrtc.org

TODO:
local storage
RESTful API
tests
+..."

I am not involved, thought many here would be interested.  Perhaps someone would like to help them with Local Storage?

[pir...@gmail.com]

I'm testing it and I'm impresed how fast it is... Compared with it,
ShareIt! (my own P2P WebRTC filesharing app) really sucks... :-( I'll
need to take a serious look on their code to take some ideas how to
improve it...

2013/6/5 tomByrer <tomb...@gmail.com>:

> --
>
> ---
> You received this message because you are subscribed to the Google Groups
> "Unhosted Web Apps" group.
> To unsubscribe from this group and stop receiving emails from it, send an
> email to unhosted+u...@googlegroups.com.
> For more options, visit https://groups.google.com/groups/opt_out.
>
>



--
"Si quieres viajar alrededor del mundo y ser invitado a hablar en un
monton de sitios diferentes, simplemente escribe un sistema operativo
Unix."
– Linus Tordvals, creador del sistema operativo Linux

[Fabio Barone]

this looks intriguing.

But there is a contradiction in the readme:

One-To-Many sharing application. Serverless. 

and

 Sharefest operates on a mesh network similar to Bittorrent network. The main difference is that currently the peers are coordinated using an intelligent server. 

??? 

2013/6/5 pir...@gmail.com <pir...@gmail.com>

[Tom Byrer]

On Wed, Jun 5, 2013 at 9:08 AM, Fabio Barone <holon...@gmail.com> wrote:
> But there is a contradiction in the readme:
>>
>> One-To-Many sharing application. Serverless.

>> Sharefest operates on a mesh network similar to Bittorrent network. The
>> main difference is that currently the peers are coordinated using an
>> intelligent server.

You are right; there is a STUN server to set up the handshake
with the public IP, or relaying via TURN if STUN did not work:
http://www.youtube.com/watch?feature=player_detailpage&v=p2HzZkd2A40#t=1075s
You can deploy the servers on Amazon, so you don't have to
self-host.

On Wed, Jun 5, 2013 at 4:48 AM, pir...@gmail.com <pir...@gmail.com> wrote:
> I'm testing it and I'm impresed how fast it is... Compared with it,
> ShareIt! (my own P2P WebRTC filesharing app) really sucks..

Perhaps you can use your wizard skills to help them out? They
have many bugs to squash. Then, when you know their system
well enough, you can fork if they won't take your patches anymore.


Thanks for the insights!

[pir...@gmail.com]

>> I'm testing it and I'm impresed how fast it is... Compared with it,
>> ShareIt! (my own P2P WebRTC filesharing app) really sucks..
>
> Perhaps you can use your wizard skills to help them out? They
> have many bugs to squash. Then, when you know their system
> well enough, you can fork if they won't take your patches anymore.
>

We have diferent focus perspectives, they seems to have developed a
simple framework where you drop your files and they are available to
others on the same shared link, some way similar to putfile (my
original idea), while I have been working on develop a full-fledged
P2P filesharing application and on the problem of removing the initial
handshake server ("the Holy Grail of Internet"), and now I'm working
towards to create a generic P2P framework and start working on content
publishing over it (i.e., browser-to-browser web surfing :-) ), but
don't doubt that we should interoperate. In fact, I think they don't
disagree to develop a common protocol and PeerJS likes the idea too,
maybe when reliable communications and interoperation comes to Chrome
we can start to work on it... :-) At this moment I'm trying to finish
and improve reliability on my WebP2P.io framework, that would be a
good start point then... :-)

[Shachar Zohar]

Hi Guys,

I just came across this thread,

First of all thanks for the complements, we hope that once google will introduce sctp, binary data channels the speed will further improve.

But there is a contradiction in the readme:

One-To-Many sharing application. Serverless. 

and

 Sharefest operates on a mesh network similar to Bittorrent network. The main difference is that currently the peers are coordinated using an intelligent server.

The notion of Serverless is refered to the data transfer, i.e the data it self doesn't touch any server.

We do have a server that handles the discovery of peers and to handle signaling between the peers and transfer metadata.

 I'll need to take a serious look on their code to take some ideas how to improve it... 

Pirahanna, you are more than welcome to check out the code absorb ideas, share ideas and don't hesitate to ask us stuff about architecture etc'.

 In fact, I think they don't disagree to develop a common protocol and PeerJS likes the idea too.

I think everyone are to gain from cooperation, this is one of the reasons we chose to open source sharefest. We believe the more people observe and give ideas the better the product will be, and the community will gain from taking ideas from it. 

We've been re factoring our code to make it more clear, and modular for developers to join the project. we'll merge this version in a week or so. And we're inviting developers to help us "squash those bugs".

Thank you!

Shachar, Peer5, Sharefest developer

[pir...@gmail.com]

> I just came across this thread,
> First of all thanks for the complements, we hope that once google will
> introduce sctp, binary data channels the speed will further improve.
>

More than currently it is??? ShareFest seemed to me blazzingly fast... :-P

>> I'll need to take a serious look on their code to take some ideas how to
>> improve it...
>
> Pirahanna, you are more than welcome to check out the code absorb ideas,
> share ideas and don't hesitate to ask us stuff about architecture etc'.
>

Thanks for the invitation :-) I'll first stabilice my WebP2P.io code
and later I'll start the discussions about how we all (WebP2P.io,
ShareFest, PeerJS and OpenPeer) we can move forward on the same
direction. With so much different perspectives and implementation we
could be able to fetch something great... :-D

P.D.: It's "piranna" ;-)

> We've been re factoring our code to make it more clear, and modular for
> developers to join the project. we'll merge this version in a week or so.

The same issue with WebP2P.io, that I'm splitting it from the core of
ShareIt! to became an independent project as a generic, application
agnostic P2P framework... :-P

> And we're inviting developers to help us "squash those bugs".
>

Cool :) How are you testing ShareFest? Do you have some unit test? I
would be interested on them, I don't know how I would create my own
ones... :-/

[Shachar Zohar]

Woops, sorry for the typo in the nickname.

We're using Jasmine to do unit tests, it's a great framework to do automatic javascript testing.

[pir...@gmail.com]

> Woops, sorry for the typo in the nickname.
>

Don't worry ;-)

> We're using Jasmine to do unit tests, it's a great framework to do automatic
> javascript testing.
>

I have heard good things both from Jasmine and Qunit, but my question
was more oriented about how do you test the network functionality
(online-offline, connection to servers...)