Why not WebDAV? about baikal

[netgusto]

We really want to stick with Calendar/Contacts (and maybe Todos) on Baïkal but not file storage.
This might be implemented on the side as a complementary project, though, if anyone is up to it.

[tmikaeld]

Please do, if you want to implement it, be sure to use streams, it's something that all other PHP implementations seem to have missed.

Reference: https://code.google.com/p/sabredav/wiki/WorkingWithLargeFiles

[flashtopia]

If you are going to implement it, please add S3 Support. Adding S3 would pretty much bake Baikal only working on linux, since the frameworks for intergrating with S3 S3cmd and windows S3 components vary. But having S3 support would be super sick!

[jeromeschneider]

@flashtopia it seems that s3cmd is a Amazon S3 client, not server; are your proposing that Baïkal should be able to connect to an Amazon S3 drive ? And maybe give access to Amazon S3 via WebDAV requests through Baïkal ?

[flashtopia]

There is the s3fs which i confused for s3cmd. Its a fuse file system backed with Amazon S3. It would be pretty neat to use something like that. I dont know how reliable it is.

Maybe S3Cmd could be used like this.
Dav client connects to baikal. If the file does not exist, then S3 is checked for it using s3cmd. The file is pulled and saved on the local server.

Files that are modified through dav, are pushed back to amazon s3 as backup.
Files which have not been modified or accessed for x days are deleted from the local server but maintained on s3.
Files that are deleted are deleted on S3 and local server.

This way the local server acts as a cache reducing s3 bandwith and requests, but you get the ability to use amazon s3 with dav. It would be super neat. Imagine you are working on a big web dev project, but dont touch it for 5 months, the files are safe on amazon s3 but not on your web server. Or a law office using WebDav can store documents. The case documents that are currently being worked on are available on the dav directory on the server & s3. Old case files are nice and safe on S3.

Its probably a lot of work. But if baikal is going to be the premier webdav server then having S3 fallback would be SUPER COOL.

• PS. my favorite feature request still remains a REST or JSON api.

[flashtopia]

http://code.google.com/p/s3fs/

[drzraf]

When you said "use webdav", do you speak about the storage backend for calendar entries ? for .ics files ?
In such case, indeed, that would be great. But note that Apache mod_dav already provides a workable file-storage solution.
The way I interpret this feature request is:

• install mod_dav
• setup webdav accounts access and directories for users
• setup bailkal to use <username>/calendar.ics as the storage for <username>'s calendar
  is that right ?

[tmikaeld]

No, storage of any file - and in this case use Baikal for user management.
Currently we use a separate self-made solution using SabreDAV, since it's the most feature rich WebDAV library.

But sure, you could implement the same using the apache mod_dav module, but SabreDAV don't require it.

[m0yellow]

@flashtopia why do you think a law office likes the idea of their case files being abroad on another server where nobody can call support if the service is unavailable. It even breaks their obligation to keep their files save. As there is a third party involved, which didn't sign a NDA.
I did understand s3 always as something where you could park a lot of files (a lot of binary files, not text or documents), to have them accessible around the world. All the lawyers I worked for, can have their years work of documents (including scans) on USB sticks (or even one). Which they can deposit in their bank save with the important contracts. I would suggest if you think you need it, why not go for a synchronizing software, that runs on changes or even via cron every 15 min. You have s3 support, and it is not tied to this project, you can even sync even more files from the local server to s3.

[sint]

not sure if this is really needed. i would rather have a perfectly working caldav/carddav server than another overfeatured owncloud clone.

[tmikaeld]

I'd rather have a complete *DAV set - make it a lightweight/unbloated owncloud clone then.

It makes no sense to have one system for each dav type.

[blinkenlight]

+1 for webDAV support - I'm using Baikal to break free of Google / other cloud services using my Android phone. So far it works great for contacts and calendars, with tasks coming along nicely. Last thing missing would be memos / notes, and webDAV sounds like just the ticket for that. I was aware that Apache can help with that, but of course I'd prefer an integrated solution... :)

Any news about a small and simple webdav support for files storage ? Nothing fancy. Just a storage quota per user should be enough for many needs.