Abstract:In this paper, we address the problem of content placement in peer-to-peer systems, with the objective of maximizing the utilization of peers' uplink bandwidth resources. We consider system performance under a many-user asymptotic. We distinguish two scenarios, namely "Distributed Server Networks" (DSN) for which requests are exogenous to the system, and "Pure P2P Networks" (PP2PN) for which requests emanate from the peers themselves. For both scenarios, we consider a loss network model of performance, and determine asymptotically optimal content placement strategies in the case of a limited content catalogue. We then turn to an alternative "large catalogue" scaling where the catalogue size scales with the peer population. Under this scaling, we establish that storage space per peer must necessarily grow unboundedly if bandwidth utilization is to be maximized. Relating the system performance to properties of a specific random graph model, we then identify a content placement strategy and a request acceptance policy which jointly maximize bandwidth utilization, provided storage space per peer grows unboundedly, although arbitrarily slowly, with system size.
Networking and Internet Architecture,Distributed, Parallel, and Cluster Computing,Performance,Probability
What problem does this paper attempt to address?
### Problems the paper attempts to solve
This paper aims to solve the content placement problem in peer - to - peer (P2P) video - on - demand (VoD) systems, with the goal of maximizing the utilization rate of the upstream bandwidth resources of peers. Specifically, the paper considers two scenarios:
1. **Distributed Server Network (DSN)**: Requests come from external users and can be regarded as exogenous requests to the system.
2. **Pure P2P Network (PP2PN)**: Requests come from the peers themselves and can be regarded as endogenous requests to the system.
For these two scenarios, the paper uses a loss network model to evaluate system performance and determines the asymptotically optimal content placement strategy under a limited content catalog. In addition, the paper also explores the situation where the size of the content catalog increases with the number of peers, and finds that in order to maximize bandwidth utilization, the storage space of each peer must grow without bound, although the growth rate can be arbitrarily slow.
### Main contributions
1. **Content placement strategies**:
- For the DSN scenario, the "proportional - to - product" content placement strategy is proposed and proven to be optimal in the limit of large - scale systems.
- For the PP2PN scenario, a different content placement strategy is proposed and its optimality is proven.
2. **Theoretical analysis**:
- Using tools such as the loss network model and Hall's theorem, the mathematical expressions of system performance are derived.
- Through asymptotic analysis, it is proven that in the limit of large - scale systems, the proposed placement strategies can minimize the proportion of rejected requests.
3. **Simulation verification**:
- Through extensive simulation experiments, the effectiveness of the proposed content placement strategies is verified.
- The performance of the sampling - based pre - allocation strategy and the demand - driven cache update strategy are compared, and the results show that both can match the theoretical optimal values well.
### Key formulas
1. **Probability distribution of content placement strategies**:
\[
p(j)=\frac{1}{Z}\prod_{c\in j}\hat{\nu}_c, \quad j\in J
\]
where \(\hat{\nu}_c\) is the normalized popularity of content \(c\), and \(Z\) is the normalization constant.
2. **Definition of system load**:
\[
\rho\triangleq\frac{\sum_{c\in C}\nu_c}{BU}
\]
where \(\nu_c\) is the request rate of content \(c\), \(B\) is the number of peers, and \(U\) is the upper limit of concurrent requests per peer.
3. **Probability of accepting requests**:
\[
A^{(B)}_c = \exp\left(-\sum_{S:c\in S}\bar{y}^{(B)}_S\right)+O\left(B^{-\frac{1}{2}}\right), \quad \forall c\in C
\]
where \(\bar{y}^{(B)}_S\) is the Lagrange multiplier of the optimization problem.
### Conclusion
Through theoretical analysis and simulation experiments, the paper proves that the "proportional - to - product" content placement strategy is optimal in the DSN scenario and can maximize bandwidth utilization in the limit of large - scale systems. For the PP2PN scenario, an effective content placement strategy is also proposed. These results provide an important theoretical basis and practical guidance for designing efficient P2P VoD systems.