When we were discussing different importance of links and how we can connect two nodes together through decentralized search, I thought of how we can be matched to other people depending on certain factors, and a big one that came to mind was matchmaking in video games. There are two major factors that go into this matching; how good the network connection is between two players (which is also often affected by location, i.e. physical distance apart), and skill disparity between the two players.
To do this, usually the skill level of each player is targeted first, and the system will look for the player in closest proximity. In a paper on matchmaking strategies, they took the example of League of Legends, a popular online multiplayer game with millions of players worldwide, and analysed the way the grouped players together to be in a match. In this case, the main factor was the ranking of each player, and they would match to the closest players with similar ranking, broadening the scope of the search and the allowed difference in ranking as time spent waiting for the queue increased.
One of problems brought up in the article is that this matchmaking does not take into account latency, which would be part the network connections strength, and affected by physical distance between players. It suggests that to handle a multiple criteria search on such a large scale, decentralized architecture is required.
This is where something like decentralized search from our lectures could be used. Something like this already exists for player matchmaking in games, and is called peer-to-peer matchmaking. Peer-to-peer to put is simply, is a decentralized communication model, and in this context would be two players (nodes) communicating between each other to play a game, rather than a centralized system dealing with hosting servers and networks.
This network is the same as a Kleinberg model, and through the game, each player is given the matchmaking strategy to connect with other players in the P2P network, and can navigate through this searchable network for another player. Although this system still has it’s flaws, it allows players to quickly connect to others, and is an interesting example of the applications of a decentralized search and the analysis of large networks in the context of a popular video game.
Articles: https://www.researchgate.net/publication/260086948_Matchmaking_in_multiplayer_online_games_Studying_user_traces_to_improve_the_user_experience
https://link.springer.com/article/10.1007/s12083-019-00725-3#Sec3