1: \begin{abstract}
2: This paper presents an asynchronous distributed algorithm to manage multiple trees for peer-to-peer streaming in a flow level
3: model. It is assumed that videos are cut into
4: substreams, with or without source coding, to be distributed to all nodes.
5: The algorithm guarantees that each node receives sufficiently many
6: substreams within delay logarithmic in the number of peers. The algorithm works by constantly updating the topology so
7: that each substream is distributed through trees to as many nodes
8: as possible without interference. Competition among trees for limited
9: upload capacity is managed so that both coverage and balance are
10: achieved. The algorithm is robust in that it efficiently eliminates
11: cycles and maintains tree structures in a distributed way. The algorithm
12: favors nodes with higher degree, so it not only works for live streaming
13: and video on demand, but also in the case a few nodes with large degree
14: act as servers and other nodes act as clients.
15:
16: A proof of convergence of the algorithm is given assuming instantaneous
17: update of depth information, and for the case
18: of a single tree it is shown that the convergence time is
19: stochastically tightly bounded by a small constant times
20: the log of the number of nodes. These theoretical results are
21: complemented by simulations showing that the algorithm works well even when
22: most assumptions for the theoretical tractability do not hold.
23:
24:
25: %We assume videos are cut into substreams, which are distributed to users by directed trees. We model the delay using tree depth, and require each user to be covered by enough trees distributing substreams.
26:
27: %Our algorithm is simple to implement because it is composed of random sampling and step updates. Our algorithm is composed of two parts, one for the coverage and the other for the balance.
28:
29: % Ji -- note how "homogeneous" and "heterogeneous" are spelled. -BH
30:
31:
32: \end{abstract}
33: