Notes on Reading Private Communication Through a Network of Trusted Connections: The Dark Freenet (click for pdf)

I learned that data encryption within a Freenet (not necessarily dark) is often restricted to two types: CHKs and SSKs (6). As a basic measure against profiling and traffic analysis, documents are relegated to set sizes — 32kb for CHKs (Content Hash Key, a type of symmetric cryptographic key generation in which the data includes the means of generating the key) and 1kb for SSKs (Signed Subspace Key, i.e. ain information publisher has an asymmetric key-pair used to sign documents within a free(/dark)net subspace. Any larger documents are split up into smaller ones to maintain this size imperative. Also, as anyone who has used any type of P2P net has experienced, search functions within the network are limited, but I just learned the reason why: directories aren’t plaintext, but are (usually) MD5 hashes of locations. So, unless you know the precise location of the document or information you’re trying to access, and thus can generate the matching MD5 hash, you can’t even discover that the directory is there. Kind of like how you can’t access an .onion site unless you already know the address. This was making my brain itch a little bit, but I’m good now.

Fuzzy Logic Operations within Computerized Social Networks applied to Darknet Operations:

“Instead [of Stanley Milgram’s small-world model], we use a method which draws off the small world models of Jon Kleinberg. The routing we perform is purely greedy: at each step, the desirability of the neighbors is ordered by the proximity of their identities to the route key K (seen as floating point number between 0 and 1 with periodic boundary). The question then becomes one of trying affect the randomized assignment of identities such that this becomes an efficient way of routing. For routing to be efficient Kleinberg’s results show that certain relation between the frequency of connections of different lengths (with respect to the identities) must be present, so our goal is to, to the degree that it is possible, assign identities so that this holds. The method we have chosen for doing this is a development of that described in [“Distributed routing in small-world networks” by O. Sandberg].” (7)

Basically, the propinquity of your friend Miss X to her friend Sir Y’s node which contains the data you’re attempting to access, this proximity is systematized via a fuzzy logic operator which outputs between a range of [0,1], 0 representing no connections to the desired node, and 1 representing the desired node. Thus, an informational subgraph is overlayed onto the subgraph of the world’s social network which contains the portion of your social network connected to the darknet.

caption id=”” align=”alignleft” width=”450” caption=”Fuzzy Logic Explained!”[/caption]

Sorry for the fuzzy logic geek out…I thought it was a cool application in informational network analysis.

“In our implementation nodes start upon joining the network with randomly selected identities seen as numbers between 0 and 1. These identities are then switched between the nodes using the simulated annealing like method first explored in [29], which causes nodes that are in some sense close in the network topology to also have nearby identities. (It is only when this property holds that greedy routing makes sense in anything but the final step.) The use of random positions, rather than fixed points in a grid, makes our model slightly different than the Kleinberg model that is the basis of previous analysis of this method, but similar continuum models have been previously studied in [15] and [14]. Nodes at a constant rate initiate random walks, which terminate after a fixed number of steps (current six, which simulation indicates is enough even in a large network). When the walk terminates, the node at which it was started and that at which it ended attempt to switch identities, which will happen with a probability specified by the algorithm. Ideally, one would hope to be able to assign the identities so that every step in a route for K brings us to a node whose identity is closer to K than the last, until we have found the best node on the network. In practice, this may not always be possible, but we still use this as heuristic to show us when to terminate a route. Currently the route continues until it has reached a maximal Hops-To-Live (HTL), which is motivated by attempting to balance a thorough search while limiting resource usage.” (7-8)

“Several enhancements to the basic algorithm described in [29] which could lead to better results have been suggested. Knowing the identity of ones neighbors’ neighbors is known to improve the performance of routing in Kleinberg type networks [24] by allowing nodes to route to the neighbor whose neighbors identities best match the query. Knowing the identities of neighbor’s neighbors reveals something about the surrounding network, but does not tell one who they are, so the basic principle of only revealing oneself to trusted peers remains. Another performance enhancement that we use is for nodes to be aware of which documents are present in their neighbor’s cache, by for instance neighbors passing Bloom filters [4] summarizing the contents of their cache to each other. A combination of both these techniques can greatly increase the number of successful searches and decrease the query length.” (8)

One enhancement to this system which I can conceive of would be the application of hash tags to identities. For instance, when searching on the darknet for a document, e.g. Kropotkin’s Conquest of Bread, your node would analyze the hash tags of your friends’ friends identities according to the likelihood of them having the document. So, for instance, you’re searching for Conquest of Bread, and you route on the darknet to your friend Miss X’s node. Miss X has two friends on the darknet, whose identities are invisible to you, but the hash tag descriptions of them are not. One of these friends has a tag identifying them as a Democrat, meaning the likelihood of their node containing the Conquest of Bread is low, say a 5% chance, and so they have an assigned value of .05 . Miss X’s other friend is an Anarchist, and so their likelihood is high, say 80%, and so they have a value of .80. Now I realise that the data stored on your node is not directly determined by you, but by what information is most commonly accessed on the darknet. However, there should still be some way to either 1) control what data is stored, or 2) apply hash tags for the type of data which has been pseudo-randomly stored. I’m not in a position to be creating a darknet right now, so anyone wanting to use this idea (if it’s even at all feasible) should feel welcome to do so.

Alright, I read further and realized I’m behind on the times. A very similar idea has already been implemented, called sinkstoring: “…each node keeps two seperate caches of data. One is a short term cache where all data that the node transfers is stored temporarily until it pushed out by other data. The other is meant to keep data longer by storing only inserted data that matches the nodes identity. The strategy used to populate this second cache, which we have dubbed sinkstore, is this: A node will attempt to place the document corresponding to a key K in its store if it is better located at the node, given its identity, than to that of any of its neighbors. For example, in the configuration shown in Figure 1, the node with identity 0.49 is a sink for K = 0.50: its current identity is closer to that value than any of its neighbors identities - if the query was an insert, the node with current identity 0.49 would store it.” (8)

One of the problems with darknets is the availabilty of trusted peers to new users. If each new user only knew one person already connected to the darknet, the resulting network would take on a tree formation, and the routing process would fail entirely. A possible solution to this could be integration of darknets with emerging decentralized social networking platforms, such as Appleseed and Diaspora, so that when users sign up for an account on the social network they are automatically integrated with a darknet system. If this were applied, new users shouldn’t have trouble connecting to multiple trusted peers. Again, someone please steal my ideas — that’s what they’re for.

Plus points to the authors of “Private Communication Through a Network of Trusted Connections: The Dark Freenet” for their excellent gender politics in using “she” wherever third person gender neutral pronouns would have been syntactically ambiguous. \<3