Gossip can be considered as a genre of communication, since it has its own descriptive and prescriptive conventions for production and interpretation, which communicators tacitly understand and which are used to evaluate the success of a particular exchange. Gossip is a simple but very impactful form of communication. As a communication mechanism with very old roots in human evolution, we use it to create social ties and interpret the world around us (in particular, risks). A gossip protocol or epidemic protocol is a computer-based peer-to-peer communication procedure or process based on the way epidemics spread.
Some distributed systems use peer-to-peer gossip to ensure that data is disseminated to all members of a group. Some ad hoc networks do not have a central registry and the only way to disseminate common data is to trust each member to transmit it to their neighbors. Listening to positive stories and sharing information about the skills and abilities of others can strengthen an organization. However, some leaders are actively engaged in gossiping in the office (even virtually) with few negative consequences.
While negative gossip may be more exciting and entertaining, it has few potential benefits (it may help identify those who are more or less trustworthy). Gossip protocols have also been used to achieve and maintain the coherence of distributed databases or with other types of data in consistent states, count the number of nodes in a network of unknown size, disseminate news in a solid way, organize nodes according to some structuring policy, creating so-called overlapping networks, calculating aggregates, classifying the nodes of a network, choosing leaders, etc. There are probably hundreds of variants of specific protocols gossip because every usage scenario is likely to adapt to the specific needs of the organization. The term epidemic algorithm is sometimes used to describe a software system that employs this type of gossip-based information propagation.
The term “convergent consistency” is sometimes used to describe protocols that achieve an exponentially rapid dissemination of information. To do this, a protocol must propagate any new information to all nodes that are affected by the information within the logarithmic time in the size of the system (the mixing time must be logarithmic in terms of the size of the system). Gossip protocols can be used to spread information in a way quite similar to the way a viral infection spreads in a biological population. After every appointment with a water cooler, the number of people who have heard the rumor almost doubles (although this doesn't explain why he gossiped twice with the same person; maybe Dave tries to tell Frank the story, only to discover that Frank has already heard it from Alicia).
Most commonly, gossip protocols are those that are specifically executed in a regular, periodic, relatively lazy, symmetric and decentralized manner; the high degree of symmetry between nodes is particularly characteristic. If bandwidth allows, this means that a gossip system can support any classic protocol or implement any classic distributed service. People gossip and no rule, regulation or requirement will stop people from engaging in one of the most primitive forms of communication.
Leave Reply