Article Preview
Top1 Introduction
This paper is an extended and redeveloped version of authors’ presentation for the 27th Conference of Open Innovations Association FRUCT (Meskanen, Niemi, & Kuusijärvi, 2020).
Privacy concerns and collection of data by the centralized social media services are driving the need and development towards decentralized and distributed solutions. These concerns are strengthened as people gain more knowledge about what information of theirs have/can be used without them being totally aware of, as not all people read the terms of service texts in detail. Numerous cases of data breaches by centralized platforms have been reported thus far (Lynskey, 2019) (Isaac & Frenkel, 2018). In the premise of these requirements, a decentralized solution is needed, in order to fulfil the end-users needs and wants, the end-user requirements. In order to achieve a decentralized social media platform, a p2p solution is required. A number of decentralized solutions have been proposed, such as (Cutillo, Molva, & Strufe, 2009) (Graffi & Masinde, 2020), that do address the base for achieving the goals towards decentralized solutions, and in some cases taking into account privacy aspects at some level, but they do not achieve untrackability in the networks.
Transforming from a centralized solution to a decentralized solution does not come without a price either. The p2p networks are open in terms of connections and data exchange between peers, including that all the communicating nodes see the end-point addresses of the other connected nodes. This cannot be fully avoided with any generic p2p solution as such (at least without onion routing), as peers communicating with each other will have to create connections with each other in order to exchange data and function as a true p2p network. This is the case when the communicating nodes cannot directly connect to each other, e.g., they may be behind distinct Network Address Translation (NAT). The problem and concern addressed in this paper is that the identification data (the user ID/IP address) should not be shared with unauthorized other person(s) involved in the p2p network.
In order to achieve this, the authors propose a set of protocols to share identification data in the environment to a set of users who should have access to a certain user’s data. To exchange data with other users, a user needs to know the actual end-point addresses of their friends, i.e., IP addresses and the ephemeral IDs. The premise of HELIOS platform1 is that users can communicate with each other in a p2p network, while taking into account the overall security and privacy of users. The restrictions and technical problems explained above need to be solved, in order to make the user untraceable. One solution to achieve some level of untraceability is to change the used end-point address and the ID of the user at certain intervals.
The libp2p2 networking library enables connecting to other nodes using a Circuit Relay (or relay in networking terms) in case the nodes are behind a NAT network and cannot connect to each other directly. Therefore, the relay feature can also be used to connect to the p2p network via a relay and exchange data with the other nodes, without connecting directly to the other nodes and, in particular, without exposing IP addresses. Another way to achieve this kind of privacy-enhancing feature would be to use another gateway outside of user’s normal networks to connect to the p2p network. Of course, that gateway, e.g., a VPN server, would see the IP address of the user in this arrangement.
Peer-to-peer networks have many advantages compared to networks with centralized control. One of the advantages is privacy. Nobody in the network has a full view about what is going on. Sensitive data can also remain local when the p2p paradigm is in use.
On the other hand, one of the advantages of a centralized system over a p2p system is easier management. For example, nodes in the network can easily find each other with the help of a centralized database that contains up-to-date information about the network endpoints of all nodes.