12. Experiments
• Performance metrics
– Average delivery probability (i.e., ratio
between the # of delivered messages
and the total number of messages that
should have been delivered)
– Average cost (i.e., number of replicas
per delivered message)
– Average latency (i.e., time elapsed
between message creation and delivery)
13. Experiments
• Communities help Bubble Rap • Bubble Rap relies on global centrality (bottlenecks)
• Contact sporadicity affects dLife/dLifeIC • dLife is affected by buffer limitation
• Focusing on content helps dLifeIC
• Buffer constraint
• Little interaction affects social-aware solutions
• Node delivery ability increases with # of
interests
14. Experiments
• Average cost
– Bubble Rap/dLife: increase
with load
– dLifeIC: decreases
• Average latency
– dLifeIC: less time to deliver
15. Conclusions
• Focusing on content
– Reach full delivery
– Very low associated cost
– Suitable delay
• Further investigation under very low
mobility scenarios
Content has a important role in
improving information delivery ☺