Freenet Frequently Asked Questions

Or even (find your IP address from ipconfig/ifconfig/winipcfg and substitute it for 192.168.1.1):

And if you want to grant full access (i.e. change config settings, restart, etc) to the node (WARNING: Be very careful who you give full fproxy access to!):

Problems installing with OpenJDK

Some versions of OpenJDK, particularly the one included with Ubuntu 8.04, have some problems with SSL which cause the installer to fail. Please install the Sun JRE, at least version 6. On Ubuntu or Debian, open a root terminal and type:

	apt-get install sun-java6-bin
	update-java-alternatives -s java-6-sun
	

Why are there so many messages in my logfile with a backtrace attached?
Fred (and freenet in general) are still very much in development, and if something goes wrong it's worthwhile to know exactly what went wrong.

How can I change from the main network to the test network?
The test network is a separate network which allows the developers to see exactly what is going on. There is no anonymity on the test network. There is a separate installer (for Linux, OS/X, for Windows). This can break quite often, so you should probably have some idea what you are doing or at least be prepared to reinstall regularly!

Freenet won't start and the launcher says it can't find freenet.ini
This is due to an unfortunate bug, fixed in 1249. You can fix it by renaming freenet.ini.tmp to freenet.ini.

Freenet won't start, it says "Service did not respond to signal"
This should be fixed now, let us know if you see it. If you have a very old install, you might be interested in the steps here.

I have Kaspersky anti-virus, and Freenet doesn't install, or shows "Download/upload queue database corrupted!"
Kaspersky can be a problem with Freenet. See here. We recommend you turn off Kaspersky during install and during node startup, and exclude the directory you installed Freenet in (most likely C:\Program Files\Freenet or C:\Program Files (x86)\Freenet).

What's new? Is there a changelog?
On every new build, a brief summary of all the main changes is posted to the support and devl lists and the eng.freenet board on Freetalk. This is usually relayed to FMS and Frost too. Alternatively, for a much more detailed view, check out the git repositories. Also, you should check the developer blogs (from the default bookmarks, or over the web, e.g. toad), but be warned they are often not regularly updated and frequently go off on rants on unrelated topics!

Publisher answers

If I publish something in Freenet, how will people find it? Don't they have to know the key I used?
Yes, people will have to know what key you used to publish your information. This means you will have to announce your key in some way.

The most common way to do this is to send a message, containing your key and brief description of your information, to the author of one of the existing Freenet sites. Most of the "portal" sites which are linked from the Freenet web interface (fproxy) read the Freetalk or FMS forums, and there are boards specifically for announcing sites (usually the boards are called "sites"!). You could also send your key to people by using the Freenet mailing lists, in the IRC channel (irc.freenode.net #freenet), by private e-mail, or by advertising your Freenet site on your World Wide Web site. If you're feeling extravagant, you could even try skywriting it. (Graffiti is not recommended, for legal reasons.)

How do I publish a Content Hash Key (CHK)?
A Content Hash Key is based on the actual content contained within it - and as such, the key will only be known after it has been inserted into Freenet. To insert a CHK, simply insert it as "CHK@", Freenet will tell you what the actual CHK is once the insertion completes.

Can Freenet documents be updated / deleted?
Currently, a document posted to Freenet with the same name as one already present may actually serve to propagate the existing document. There is also currently no means of deleting a document from Freenet. Documents that are never requested are eventually removed through disuse.

However, you can use an Updatable Subspace Key (USK) to provide a form of updatable freesite: your node will automatically look for later editions of the site (after you visit it, or always if you bookmark it), and show you the latest version. You can force it to search for the latest version by changing the number at the end of the key to negative.

Contribution answers

I have this great idea....
Good! First step: read the mailing list archives. Odds are good that someone else had the same idea and discussed it with the group. Either a flaw was found in the idea, or perhaps it was decided to postpone implementing the idea until later. Some examples of ideas already discussed are storing information by content hash, key redirection, signed keys/data, use of UDP, server discovery, URLs, document versioning, and others. If you don't see the idea discussed in the archives, by all means bring it up in the appropriate mailing list.

Can I contribute to the Freenet Project?
Absolutely. Even if you don't have the time or skills to become a co-developer of the project, you can contribute in other ways:

• Help test Freenet by installing and configuring the server software on your machine.
• Install the client software on your machine to test retrieving information and publishing your own.
• Work on the Freenet web site (including the FAQ).
• Contribute your ideas to the discussion lists.
• Translate the user interface into your local language.

If you are a developer, you can also help by designing and developing other applications to run on Freenet. Or maybe even mention us on your website and promote us just like this web hosting guide did.

People that want to contribute to Freenet in any way, by contributing artwork, new ideas, or even correcting spelling/grammar mistakes or ideas for adding new pages/paragraphs on this site, are also welcome to email.

How can I access the code and website?
See the developer page for details of our git repositories etc.

What tools do I need to help develop?
To build and deploy the Freenet server, you will need Java tools compatible with Sun's JDK 1.6 or later. You can download the source tarballs on the download page for a specific build, or use git to get an up to date copy of the source, see here for details. Further instructions for building and deploying the server are included with the code itself. Generally speaking, joining our IRC channel is a good idea: #freenet on irc.freenode.net

Is there a Help Site that goes deeper into the questions newbies may have about Freenet, and where people can contribute too?
Have a look at our wiki. An older wiki, which is now read-only, but has a fair amount of content so is sometimes helpful is here. There are also several implementations of wiki's over Freenet. The most recent one is called Jfniki. There is a link in the default bookmarks on the Browse Freenet page after you install Freenet.

Where can I report bugs?
You can use our bug tracking system or send a mail to our support mailing list.

I'm a theoretical computer scientist/mathematician, how can I help? (research challenges)
See here.

Security answers

Can I use my regular browser to browse Freenet?
Freenet has a web interface: much of the content on Freenet is in the form of "freesites", and downloads, configuration and friend connections can be managed from the web interface. However, because of weaknesses in current browsers, we strongly recommend that you use a separate browser for Freenet. Specifically, browser history stealing, in all its forms, is a major threat if you share a browser between Freenet and the WWW at large: malicious web pages will be able to probe which freesites you have visited, and report this information to their owners.

With recent browsers, privacy/incognito mode may be sufficient, and the rabbit applet on the system tray on Windows will try to start a browser running in this mode. However, this is not 100% reliable in our experience, so be careful.

Won't attack X break Freenet's anonymity?

Short answer: Yes, on opennet. Maybe, on darknet.

Long answer:

Freenet has a different threat model to Tor and the Mixmaster remailers. Freenet is designed to resist censorship: The network must therefore be robust, and content must be distributed without requiring a central server, whether anonymous or not. Anonymity is important for requesters and especially for those who upload content in the first place. The typical example is a corporate or government whistleblower. Generally to find the originator of some content, the attacker must be able to predict the data in advance, must be able to move across the network relatively quickly, and must be able to perform the attack while the data is being inserted; after that, it is distributed across the network and is much harder to trace, and the originator may have left the network. However, if by chance or by overwhelming force the attacker is connected to the whistleblower (or e.g. seizes the computers of everyone on the network), he may be able to identify this much more quickly. All of this is vastly more difficult on a darknet, where everyone connects only to their friends, where it is very hard for an attacker to find nodes, and where to connect to a given node he must social engineer its operator! Freenet does support opennet mode (plug and play), but darknet is far more secure, and far more difficult to block on a national firewall.

Tor on the other hand is designed to anonymise real-time data streams, on the assumption that the list of nodes can be public, that there is a free world where nodes can be operated safely, that the authors of controversial content will be able to either host (hidden) web servers themselves or upload it to other (hidden, but usually centralised) storage systems, and so on. And Tor has a concept of a "client", which is somebody who uses the service without providing any value to it; on Freenet, every node relays data for its neighbours. Hence the attacks on Freenet are completely different to the attacks on Tor. Both compromise to some degree to enable more or less real-time performance.

If you can use the darknet, trust your friends, don't reinsert files, always use the "Insert a random, safe key" option, and change your anonymous identity after some volume of inserts, you should be relatively safe using Freenet. However this has not yet been quantified. If you can connect, build up some trust in your anonymous persona, insert your controversial content, and then disappear, again, you are better off with Freenet, especially if the content is a website (but if you are connecting on opennet, beware of seednode compromises). In some other cases, Tor is better.

We are still working on Freenet's security and there are major security enhancements which have not yet been implemented, most of which will go in before 1.0. Cryptographic tunnels similar to Tor's onion routing are one possibility, which would greatly reduce the impact of many of the below attacks, but there are several other enhancements planned, both to anonymity and to network robustness/undetectability.

• Harvesting: Simply by running some powerful Freenet nodes, an attacker can identify most of the opennet (Strangers network) relatively easily. These nodes can then be attacked one by one (subject to resources), their traffic analysed, or simply be blocked on a national firewall. Connecting only to friends (darknet) largely solves this problem. ISPs may be able to identify Freenet nodes with some effort, although we make this fairly difficult: Freenet's current protocol is designed to be hard to detect, and steganography will be introduced at some point. However, traffic flow analysis, or brute-force blocking of all peer to peer traffic (e.g. traffic between IP addresses marked as "consumer" rather than "business"), both of which would hit a lot of things other than Freenet, would likely be effective for quite some time.
• Bootstrapping attacks: Unless a node only connects to friends, it will have to connect to the opennet "seednodes" to announce itself and get initial peers to connect to. At the moment there are relatively few seednodes and the list is maintained manually. The seednodes could be blocked easily by a national firewall etc, but also, there is little to prevent attackers from setting up their own seednodes and submitting them, and then "capturing" any new Freenet users who connect to their nodes, in order to observe their traffic etc. Freenet will try to announce to multiple seednodes, but see the below section on "correlation attacks", which generally are feasible with only a single connection to the target. So this is a question of resources - if the attacker has the resources to surveil all new Freenet nodes, he has a good chance of pulling it off. In future we may have more seednodes, and only reveal a small proportion of them to each node, as Tor does with its hidden bridges, but that will not prevent attackers from creating lots of malicious seednodes and getting them into the official lists, and it will likely still be possible to block all the seednodes with some effort (something similar has already happened to Tor hidden bridges in China). Combined with harvesting and adaptive search attacks, this attack explains why opennet is regarded by many core developers as hopelessly insecure. If you want good security you need to connect only to friends. Hit and run inserts are possible, and can be relatively safe in terms of many of the other attacks, but you are taking the risk that the opennet seednode you connect to may be malicious.
• Datastore attacks: This is largely solved as of build 1224, we don't cache our local requests or inserts, and neither do the nodes immediately connected to us, to a depth of at least 2 hops (3 on inserts). However, if your node is older than that, seizing the store might give a bad guy some interesting information. Also note that the client-cache caches local requests (but not inserts), so it should be encrypted and passworded by setting the physical security level to HIGH, or turned off. You should also encrypt the swapfile in particular and the whole system if possible to prevent information leaks from the web browser, media players etc. Note that some incriminating data (e.g. the list of bookmarks) is still stored in plaintext; we're working on it, but did I mention you should encrypt your whole system?
• Correlation attacks: If you are connected to a node, and can recognise the keys being requested (probably because it was posted publicly), you can show statistically that the node in question probably requested it, based on the proportion of the keys requested from that node, the locations of nearby nodes, the HTL on the requests and so on. This will be largely eliminated by tunnels (but these will be quite expensive so may need to be turned off by default except for predictable blocks), but in any case it requires a rather powerful attacker compared to the next attack... Note also that if you only connect to your friends, a remote attacker will have to either co-opt your friends or social engineer you into giving them a connection; either way, connecting to the entire network this way is rather expensive: If they already suspect you personally they'll probably bug your keyboard rather than trying to connect to your Freenet node!
• Adaptive search: If you want to find the author of some content, and you can predict the exact keys which will be inserted, and you are able to connect to new nodes at will, you may be able to listen out for the keys, guess where they must have come from, connect to nodes near there, and if your guess is correct, get more keys which gives you a more accurate fix on the originator, so the attack gets faster and faster and eventually converges on the originator. This attack is most powerful with inserts of big, predictable files, but the "Insert a random, safe key" option will make the keys unpredictable even if the content is guessable, by using random encryption keys. The downside is it produces a different key each time for the same file, and you can never safely reinsert the same file to the same key. Given that Freenet's data persistence is currently relatively poor, this is a problem. Anyway, if you can use the random keys option, the attacker is unable to move towards you until after you announce the file: Most of his samples will come not from the actual content inserts but from chat posts. There are far fewer of these, and changing your pseudonymous identity periodically will help, provided the attacker cannot easily connect the new identity to the old one. Using a dedicated identity for posting sensitive content, which doesn't chat too much, again will help. Another thing which makes a huge difference is connecting only to your friends (i.e. using darknet): This makes it extremely difficult for an attacker to get new connections closer to where he thinks you must be, just as it helps with correlation attacks. So the biggest problems with this attack are 1) Files which are not very popular fall off Freenet relatively quickly, so need to be reinserted, but it is not safe to reinsert to the same key (this is why we have the "Insert a canonical key" option, for those who don't care about attacks), and 2) Chat can still be attacked. Tunnels will help to deal with both problems, and by default will only be used for predictable keys so can be relatively slow without this causing problems in practice. Also there is work going on on various techniques to allow users to do reinserts safely via for example preventing the attacker from seeing requests started before he connected. Another important point is this only works if the source is uploading new content, or chatting, regularly; creating and bootstrapping a new pseudonymous identity over a short period, doing a single insert (of any size) with the safe random key option, and announcing it, should be relatively safe from this attack, even on opennet - but see the section above on bootstrapping attacks.
• Traffic analysis: Freenet provides minimal protection against global traffic analysis (basic message padding etc); if the attacker also has nodes on the network, the extra data will likely be helpful. We certainly do not guarantee that it is impossible to trace data transfers from one node to the next with detailed traffic data, however it is hoped that this will fall down on the busier nodes. One day we will implement steganographic transports and/or constant bitrate links as an option for more paranoid users. Note that on Tor-style networks, global traffic analysis will defeat the network completely: all that is needed is to observe both the entry and exit points.
• Swapping attacks: It is possible to attack the location swapping algorithm, and thereby disrupt routing on friend-to-friend networks. This has been demonstrated by the authors of the Pitch Black paper. We are working on a solution, but sadly at the moment most users use opennet.

More information on the current practical state of Freenet security is available here.

Is Freenet vulnerable to flooding attacks?
Short answer: no.

Long answer:
We don't think so. Aside from protecting freedom of speech, Freenet is also designed to be an efficient dynamic caching system. If information is requested a lot from a limited number of nodes, the nodes that the requests pass through will cache the information, lowering the load on the network. If information is inserted on a limited set of nodes and then subsequently requested a lot from a separate set of nodes, with repetition, the sets will close in on one another in the network topology until they are "neighbors" and only the originally targeted nodes are suffering from the attack.

In other words, in order to harm Freenet with a flood you need to consistently change your point of entry into the network and continually insert and request new data, and you will still only increase the workload for the network that is linear to your own. Given an immense will and capacity greater than the total of the entire network, it is possible to cripple any public network (including the Internet itself) with floods, but it is our intention to always keep Freenet as resistant to this as theoretically possible.

Curiously enough, the above analysis only applies to Opennet. On Darknet, you might have a little more success, although it would be much harder to change your entry point in any significant way. Nonetheless, you have a reasonably low bandwidth multiplier (the total number of nodes visited, around 20 on average), and you are severely limited by the number of nodes you can connect to, which will be low on a darknet.

Why hash keys and encrypt data when a node operator could identify them (the data) anyway if he tried?
Hashing the key and encrypting the data is not meant a method to keep Freenet Node operators from being able to figure out what type of information is in their nodes if they really want to (after all, they can just find the key in the same way as someone who requests the information would) but rather to keep operators from having to know what information is in their nodes if they don't want to. This distinction is more a legal one than a technical one. It is not realistic to expect a node operator to try to continually collect and/ or guess possible keys and then check them against the information in his node (even if such an attack is viable from a security perspective), so a sane society is less likely to hold an operator liable for such information on the network.

What about hostile "cancer" nodes within the network?
The existence of malicious nodes within the network is the most difficult problem that a distributed network must face, and has been the bane of many previous ideas. Many systems (such as multiplayer gaming networks) try to avoid malicious nodes by keeping the protocol and code closed, but we have yet to see an example of that working in the long run. And anyway it is opposed to Freenet's philosophy.

Freenet is based on a balance of positive and negative feedback loops that bring requests for information to a node when it is functioning well, and keep requests away from it when it is not. The key to avoiding "cancers" is (as in the body) to make sure these loops can correctly identify even the most carefully designed malicious node and not keep sending requests to it. This issue is not fully dealt with by the current test code, but you can rest assured that a number of possible solutions have been on the table and discussed for some time now. Several have been implemented (enforcing hashes or signatures on content, per node failure tables, backing off from a node that causes timeouts ...)

What about attack Y?
Freenet is still in testing and there are bound to be attacks found that we have not dealt with yet. So if you do manage to figure out a truly new kind of attack, we are interested in hearing about it. Please keep in mind what Freenet is and what it is not, however. No single network can offer everybody everything, and there are security issues that Freenet, by it's nature, may not deal with to extent you might wish. If this upsets you, all of our code is freely available, so you are free to take as much of it as you like and write your own distributed network that suits your desires.

What private data does Freenet store? How do I get rid of it? How can I secure my computer so I am safe when running Freenet?
First of all, we strongly suggest that you install Freenet inside an encrypted drive using, for example, Truecrypt. It is not possible for Freenet to prevent all leaks of private data, especially if you download media files etc. Even if you only browse freesites and use the chat plugins, there will still be potentially incriminating data in your swapfile, which needs to be encrypted (on recent versions of Windows you could try the command "fsutil behavior set encryptpagingfile 1", but really the solution is to encrypt your whole system including swap). It is also essential that you use your web browser in privacy mode, or with cache and history turned off; we try to do this if you launch Freenet via the rabbit icon, but there are no guarantees as unfortunately this functionality seems buggy in current web browsers. Browser plugins could also be a problem, and you should use a separate browser for Freenet if in any doubt. Be careful with the files you download from Freenet - not only could anyone seizing your computer see you have them (media files are likely to be written to disk even if you open them directly in your web browser and never save them), but also they could contain threats to anonymity themselves, such as calling back to a malicious website etc; this is possible in for instance PDFs and some video formats. Freenet tries to warn you about this when it can't filter out such malicious content: Currently it can only filter HTML pages, GIF/PNG/JPEG images and CSS, and MP3s, but we will add support for Ogg soon and other formats later. And of course there are many other threats - you should take standard security precautions, such as not running operating systems that are no longer updated, not running software not from a trustworthy source, using appropriate security software etc (if you have a firewall make sure it allows the two UDP ports Freenet needs through).

Because not all users will have installed encrypted drives at the time when they first install Freenet, Freenet itself attempts to encrypt all the potentially incriminating data that it stores on disk. Details are below but as explained, leaks are inevitable: you really should encrypt your disks!

The main datastore does not store data you request or insert (or that is requested or inserted by nearby nodes), because it can be probed by other nodes: This was introduced to fix this attack publicised by The Register. Freenet has a separate client-cache, which stores data which you have recently requested to avoid having to go back to the network every time (which would not only reduce speed but also security, by giving attackers more opportunities to see your requests). Also, Freenet stores the list of your downloads and uploads (which you can see on the Filesharing menu), their current progress, and various other data, in the file node.db4o (or node.db4o.crypt). The actual data is kept in the persistent-temp-* directory. Unless you set the physical security level to LOW, this data is encrypted. At MAXIMUM, the encryption keys are never written to disk, so the data is effectively wiped on restarting the node; otherwise the encryption keys are stored in a file called master.keys (on HIGH this is passworded). You can wipe the data by either using the panic button on the downloads/uploads page or by securely deleting master.keys. Freenet also creates temporary files for other requests, which are also encrypted unless physical seclevel is LOW, which are in temp-*. Also, some plugins may create their own data files, which may contain for instance messages you have posted or downloaded from chat forums, and currently bookmarks and recently completed files are stored in plain text. It is our intention to move these into node.db4o or store them in separate encrypted databases, as soon as we have automatic backups for node.db4o. See here for details on some of the files.

I set a password and now I forgot it, what can I do?
The password protects your downloads and uploads and the client-cache (cache of what you've recently browsed on Freenet). It is stored in the file master.keys. There is no way to recover the password, but if you forget it you can wipe your downloads and uploads and the client cache by securely deleting the file master.keys. See the question on private data and local security for more information.

Freenet keeps complaining about clock skew
Freenet will have problems if your clock is constantly being rewound. Usually this happens when something is resetting your clock regularly in big jumps. On linux, you should run ntpd to make sure your clock isn't too far off (this isn't vital but it's helpful), but if you see clock skew errors, try adding the -x option to it to avoid big backwards jumps. Also, running ntpdate on startup so there is one big jump before freenet starts is a good idea. This can also happen on Windows sometimes, let us know how you managed to fix it ... generally it's not all that serious though, especially if big jumps in the clock are only once a day.

Where can I get a high quality copy of the rabbit icon?
Here.