Gossip-based dissemination

Middleware for Network Eccentric and Mobile Applications

Volumn , Issue , 2009, Pages 169-190

  Friedman, Roy ^a   Kermarrec, Anne Marie ^b   Miranda, Hugo ^c   Rodrigues, Luís ^d  

^a TECHNION ISRAEL INSTITUTE OF TECHNOLOGY   (Israel)

^b INRIA   (France)

^c FCUL   (Portugal)

^d INESC ID   (Portugal)

Author keywords

[No Author keywords available]

Indexed keywords

EID: 84859778345     PISSN: None     EISSN: None     Source Type: Book    
DOI: 10.1007/978-3-540-89707-1_8     Document Type: Chapter

References (4)

1. The probability of retransmission is a parameter in GOSSIP3. In the experiments reported in [353], for uniformly distributed networks whose density is sufficient to ensure connectivity [344] and the number of nodes is up to a few hundreds, the value of 0.65 was deemed the best.
2. The reliability of a message dissemination in a publish/subscribe systems also refers as the completeness of the dissemination: a dissemination is complete when all the nodes subscribing to an event receive it.
3. The way interests are broadcast and messages disseminated is extremely similar to the so-called gossip-based approaches presented above.
4. The standard deviation is the relevant metric because the space for storing data items is constant.