Abstract
This paper investigates modified random timers based on uniform and exponentially distributed timers for feedback scalability for large groups. We observe the widely-used probability distribution functions and propose new ones that are aware of network delays. The awareness of network delays of our proposed modified p.d.fs proves to be able to achieve lower expected number of messages compared to the original ones given that the parameters are optimized for the network variables: the number of receivers, and the network delay. In our analysis we derive an equation to estimate the optimized parameter based on these network variables. We also simulate the p.d.fs for heterogenous network delays and find that each receiver only needs to be aware of its network delay.
Original language | English |
---|---|
Pages (from-to) | 437-444 |
Number of pages | 8 |
Journal | IEICE Transactions on Communications |
Volume | E87-B |
Issue number | 3 |
Publication status | Published - 2004 Mar |
Keywords
- Feedback suppression
- Large group multicast
- Scalability
ASJC Scopus subject areas
- Software
- Computer Networks and Communications
- Electrical and Electronic Engineering