Gossip-based aggregation in large dynamic networks
read more
Citations
Internet of things: Vision, applications and research challenges
PeerSim: A scalable P2P simulator
State Estimation and Sliding-Mode Control of Markovian Jump Singular Systems
Gossip-based peer sampling
Design Patterns from Biology for Distributed Computing
References
Collective dynamics of small-world networks
Linked: The New Science of Networks
Reaching Agreement in the Presence of Faults
Small Worlds: The Dynamics of Networks between Order and Randomness
Epidemic algorithms for replicated database maintenance
Related Papers (5)
Frequently Asked Questions (11)
Q2. What are some examples of aggregation functions?
Examples of aggregation functions include network size, total free storage, maximum load, average uptime, location and intensity of hotspots, etc.
Q3. What are the two classes of realistic small-world topologies?
The Watts-Strogatz and scale-free topologies represent two classes of realistic small-world topologies that are often used to model different natural and artificial phenomena [1, 28].
Q4. How do the authors calculate the nth central moment?
To calculate the nth central moment, given by (w − w)n, the authors can calculate all the raw moments in parallel up to the nth and combine them appropriately, or the authors can proceed in two sequential steps first calculating the average and then the appropriate central moment.
Q5. What is the mechanism used to terminate a protocol?
To implement termination, the authors adopt a very simple mechanism: each node executes the protocol for a predefined number of cycles, denoted as γ, depending on the required accuracy of the output and the convergence factor that can be achieved in the particular overlay topology adopted (see the convergence factor given in Section 3).
Q6. What is the structure of the graph between the two extremes?
For intermediate values of β, the structure of the graph lies between these two extreme cases: complete order and complete disorder.
Q7. What is the exact implementation of dynamic queries?
The exact details of the implementation of dynamic queries (if necessary) will depend on the specific environment, taking into account efficiency and performance constraints and possible sources of new queries.
Q8. What is the reason why node crashes are important?
This represents another important source of error, although the authors note that from their point of view node crashes are more important because the authors model leaves as crashes, so in the presence of churn crash events dominate all other types of failure.
Q9. Why are static topologies considered unrealistic in the presence of churn?
While static topologies are unrealistic in the presence of churn, the authors still consider them due to their theoretical importance and the fact that their protocol can in fact be applied in static networks as well, although they are not the primary focus of the present discussion.
Q10. What is the probability of a variance reduction step?
In Section 3.2 it was proven that ρ = 1/e (where ρ is the convergence factor) if the authors assume that during a cycle for each particular variance reduction step, each pair of nodes has an equal probability to perform that particular variance reduction step.
Q11. How can the authors describe the waiting time between two consecutive selections of a given node?
When iterating AVG, the waiting time between two consecutive selections of a given node can be described by the exponential distribution.