RUBY/ConsectedClientApps/transfer2webdav
========================================

Homepage:: http://www.consected.com/transfer2webdav
Copyright:: Copyright 2010, Consected LLC, Massachusetts, USA.
Source:: http://github.com/philayres/transfer2webdav

ConsectedClientApps is a small collection of client-side applications to
support common document and workflow functions performed in an office
environment. This program, transfer2webdav allows a user to simply make
transfers to a WebDAV server without requiring knowledge of WebDAV client
software, or even any specific client to be installed. The design attempts to
allow transfer2webdav to be used on any major platform without change.

There are three modes of operation:

1) Send a single file
2) Transfer all the files in a directory
3) Monitor a directory for transfer when new files are added

What makes transfer2webdav a little different is the approach it takes to
managing credentials, while remaining (hopefully) platform independent.
Rather than prompting the user for WebDAV credentials for each file or when
starting monitoring, the transfer2webdav configuration stores the credentials so
that they can be used without further interaction by the end user. The password
is not just stored in plain-text, and the configuration can not be used by
another user. The configuration ties a specific user's configured credentials to
the user logged on to the operating system, so even if another user 'borrows'
the configuration file, the configured credentials will be useless. A user can 
set up multiple configurations to run transfers simultaneously to different
servers or remote directories if required.


Examples
--------

CONFIG

Setup a configuration for the user, running the CONFIG command, in the form:

transferdocuments.rb CONFIG 'configuration_name' 'url' 'username' 'password'

e.g.

ruby transferdocuments.rb CONFIG 'my_config'\
  'https://aserver.net/webdav_server/folder/' \
  '[email protected]' 'plain-text pa$$word'

After attempting to connect to the WebDAV server and testing the existence of
the supplied subfolder path, the configuration is saved to:
~/.consected_client_apps/transfer2webdav/{configuration_name}_config

Note: the configuration directory and user_config file are chmod'ed to prevent 
reading by unauthorized users.


MONITOR

Monitor a directory for new files, deleting them after they are transferred:

transferdocuments.rb MONITOR 'configuration_name' 'localdirectory' \
  {'webdavsubfolder'}

e.g.
ruby transferdocuments.rb MONITOR 'my_config'\
'/home/bob/Desktop/Consected Transfer' 'webdav-subfolder' &

The monitor will transfer all files already existing the in the directory when
started up, in case anything was left from a bad shutdown previously. Obviously,
you will need to have a valid configuration in place for this to start up at
all.


STORE

Transfer a single file with:
transferdocuments.rb STORE 'configuration_name' 'filename' {'subfolder'}

e.g.
ruby transferdocuments.rb STORE 'my_config'\
'/home/bob/Desktop/Consected Transfer/a-document.pdf' 'webdav-subfolder'

This will return with success or failure. This works nicely for applications
that can call a program after creating a file (some scanner apps do this).


STOREDIR

Transfer a directory full of files with:
transferdocuments.rb STOREDIR 'configuration_name' 'localdirectory' \
  'webdavsubfolder' {DELETE}

e.g.
ruby transferdocuments.rb STOREDIR 'my_config'\
'/home/bob/Desktop/Consected Transfer' 'webdav-subfolder' DELETE

This will return with success or failure. You can leave DELETE off the end if
you don't want the local files to be deleted after transfer. Only a single
directory will be stored, it does not recurse. It also ignores any hidden files
(i.e. those starting with a '.' character).

Monitoring a Directory as Files Change
--------------------------------------

The reason for the MONITOR option is the need to transfer files that can appear
at any time from a range of different applications. This gives us no control
over how the application stores its data. Use cases that transfer2webdav tries
to address are:

1) Files moved from another directory
2) Files saved directly to disk by an application
3) Files copied from another location(for example Explorer, Nautilus)

The first item, 'move', is easy to address, since it is an atomic action that
triggers the notification of the new file. The file is complete the moment it
appears in the monitored directory.

A test document scanning program revealed that just looking for a new file was
insufficient. The program (gscan2pdf) writes out all the scanned pages of a
document to a PDF, and is relatively slow at doing so. The file it writes to is
opened on the disk, and appears to MONITOR as a new file of zero size.
Fortunately, when gscan2pdf is finally complete, the data gets written to disk
pretty rapidly. This appears to MONITOR as another file notification. This makes
it easy to just ignore zero sized files, in the expectation that they will
eventually grow into real files that force a new notification.

Surprisingly, #3 is the hardest, and probably applies to many application.
When copying from local disk, the action is often fast enough to complete within
the directory monitoring grace period . But if you copy a large file from the
network, or from another WebDAV server, you will see the file gradually growing.
MONITOR receives a notification for every chunk of data that hits the disk,
making things tricky, since you can't move the file just because it suddenly
exists and is larger than zero bytes. MONITOR takes the approach that, as soon
as a file exists with some size, it needs to wait for the file to stop growing.
It basically goes into a loop for the file, polling it every few seconds to see
if it has grown since the last check. If it has stayed a stable size, the
transfer is allowed to proceed. If the file size has changed, the transfer is
delayed for another waiting cycle, to see if next time it stops changing.

MONITOR does attempt to play nicely, therefore every file is delayed by a few
seconds to check it is a consistent size, but in doing so, the who MONITORing
cycle is not held up. Rather than using sleep, MONITOR implements the
eventmachine periodic time to manage the polling of a single file. This allows
monitoring to continue, and multiple newly arrived files to get queued up for
simultaneous checking and subsequent transfer. If we used 'sleep', throughput
would be limited to the poll time for checking file sizes, since every action
would queue up synchronously behind the poll. At least this allows multiple
polls to be waiting simultaneously.


Discussion About Stored Credentials
-----------------------------------

The design is intended to allow configuration of stored credentials to be made 
by standard users without IT assistance, especially in environments where users 
move between PCs regularly, or have morethan one PC that they work on. The 
server does not need to be made aware of changes to the client PC that is being
used to connect to it, since when processed correctly on the server the 
credentials result in a username / password-hash that can be compared against a 
standard user database. 

For our needs, the design avoids the need for the server to manage an additional 
public key for authentication, in addition to standard user credentials. 

Taking a look at the user configuration file, we see: 

{:url=>"https://aserver.net/webdav_server/subfolder/", 
 :username=>"[email protected]",
 :key=>"53e9876f148fa9d3ea2be7f616cdb5d5a1b8ae8d"}

(I know it should try harder with the config file...)

A clientname ties this configuration to a user by doing `whoami` (or the
Windows equivalent) and appending the hostname of the client machine.

The key is a hash of the username, clientname and plain-text password. The
clientname is not stored in the configuration file, since this should be
generated at the time of authentication. This prevents carelessly backed up
files from identifying themselves quite so easily if stolen.

At a minimum, this means that for the credential configuration to be stolen, the
attacker will need to know the login id, client hostname and get the config file
to use the credentials. All are possible in an insecure environment, but in that
case an RSA private key would be equally vulnerable.

To handle authentication, the ConsectedDav class prepares credentials for the 
login to the WebDAV server that follow this scheme:

userid = 'storedcredentials//' + current_clientname + '/' + config.username + '/'
password = config.key

We could therefore create a user in our system with this userid and password,
but that's not very secure, since the key becomes the equivalent of a plain-text 
password stored on the client. Our WebDAV server authentication goes a step 
further and does _something_ like this:

if userid[0..18] == 'storedcredentials//'
  username = userid[19..-1].split('/')[1]
  user = UserProfile.find_by_username(username)
  return user if Digest::SHA1.hexdigest(userid + user.userPassword) == key       
  return false
else
    # Do a standard username and password login
end

This allows us to maintain a standard authentication model.

Installation
------------

Installation varies by platform (due to the way different operating systems
handle directory change notification).

Windows:
Install Ruby and Gems using the installation script on the installation media
gem install win32-changenotify win32-event em-dir-watcher net_dav eventmachine

Linux:
sudo apt-get install ruby ruby-gems
sudo gem install rb-inotify em-dir-watcher net_dav eventmachine

Mac:
Install Ruby and Gems
sudo gem install em-dir-watcher net_dav eventmachine


    
License
-------

    This file is part of transfer2webdav.

    transfer2webdav is free software: you can redistribute it and/or modify
    it under the terms of the GNU Lesser General Public License as published by
    the Free Software Foundation, either version 3 of the License, or
    (at your option) any later version.

    transfer2webdav is distributed in the hope that it will be useful,
    but WITHOUT ANY WARRANTY; without even the implied warranty of
    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
    GNU Lesser General Public License for more details.

    You should have received a copy of the GNU Lesser General Public License
    along with transfer2webdav.  If not, see <http://www.gnu.org/licenses/>.