Tribler Deployment comments

[SOglf]

I will try to write a description what Tribler has and has not, what is needed for a fully working and USABLE anonymous torrent client.
Version 8.2.3

Let's start from the BAD side:

- Tribler lacks speed control mechanisms. To properly deploy any networking software which is autonomous when it comes to traffic, IT HAS TO HAVE a speed control mechanism. Examples include: tor relay node, qbittorent client, download managers, i2p, and similar software.
Without speed control, it's:

1. Confusing for new users
2. Problematic in bigger networks (why the hell this traffic is so big)
3. Even more confusing when setting speed limits in the interface and those DON'T WORK PROPERLY.
   Current workaround:

• Use TC or iptables packet limits. Unfortunately regular users are unable to do that.

- Tribler does not do the transfer of torrents inside it's overlay network. Overlay is there, it is working, it does relay large amount of traffic, but there is a one VERY MAJOR PROBLEM: ANY TORRENT CONNECTION for transferring actual data, ALWAYS USES EXIT NODES. If two tribler clients want to exchange data from the torrent, which one of the clients already has, DATA SHOULD BE TRANSFERRED INSIDE THE OVERLAY NETWORK, NOT EVEN TOUCHING IN ANY WAY exit nodes.

• Access to the WEB interface is problematic: at first you need to set a STATIC PORT. Then you need to copy the key from config file. Then you need to redirect the port because it only allows connections from 127.0.0.1, And then you have to KNOW HOW TO paste the key into the web interface. THIS IS NOT THE WAY TO BUILD AND ACCESS A WEB INTERFACE
  There should be a clear explanation, step by step how to access web interface on a headless install, or an option which will make it easy.

• Standard install with a debian running script CREATES A LOG FILE IN /tmp DIRECTORY and fills it with logs which can be GIGABYTES BIG. Tribler has to be started with: /usr/share/tribler/tribler --server, to avoid that.

• There is only a .deb package for linux systems, which is enough, but a whole VM or standalone binary, or an installation script would be nice. It's not a flaw, but this could be extended for easier deployments. Maybe less dependencies?

Now what tribler does FINE:

• Tribler is actually protecting your ip address. UDP packets are also not typical, and sometimes smaller. This is great.
• If a file already exists, adding magnet or torrent will start hashchecking. This is realy nice.
• Debian dependencies are not that big, the install is not problematic besides one hickup, but overall great.
• User interface is somewhat intuitive, it's okay, if the design was "data hiding" type, then it works as it should.
• Data transfers are immune to network outages, the same as with most torrent clients. This is important, because it's one of the strong sides of torrenting compared to http or other transfers.
• Tribler is autonomous. If set correctly, it can run itself, without any supervision.
• CPU usage is not the lowest, but I think it's very acceptable, the same with the memory.
• Install process for a regular user on a workstation is flawless

Overall, if it wasn't an experimental software, my score would be VERY POOR. But because it's a work in progress, and one of a kind solution, I'd say it's GOOD. Not bad, not great, but overall GOOD.

As we are facing unprecedented increase in attacks on privacy and freedom, Tribler has become one of the crucial elements of our future, and a critical part of torrenting in countries which are targeting torrent users.

[qstokkink]

Thanks for the feedback. Always good to hear the "first use" experience, even if it is not all positive. This gives us something to work toward.

I can offer some insights regarding what we are doing "behind the scenes" about some of your points:

Access to the WEB interface is problematic

Since the release of Tribler 8.0, we have started to transition into support for remote interface access. Safely exposing your machine to the Internet is difficult. So, for now, we're only personally guiding users to do so. Once we're confident that everything is safe and working, we will indeed make this easier.

CREATES A LOG FILE IN /tmp DIRECTORY and fills it with logs which can be GIGABYTES BIG.

We currently funnel the log output of tribler into a temporary file created with mktemp. You can find exactly how we do it here. If you want to absolutely minimize the log output, Tribler supports setting the --log-level as a command line argument: setting the logging to only CRITICAL will suppress just about all logging.

There is only a .deb package for linux systems, which is enough,

We also support Docker and Flatpak.