The increasing demand for computational power in big data and machine
learning has driven the development of distributed training methodologies.
Among these, peer-to-peer (P2P) networks provide advantages such as enhanced
scalability and fault tolerance. However, they also encounter challenges
related to resource consumption, costs, and communication overhead as the
number of participating peers grows. In this paper, we introduce a novel
architecture that combines serverless computing with P2P networks for
distributed training and present a method for efficient parallel gradient
computation under resource constraints.
Our findings show a significant enhancement in gradient computation time,
with up to a 97.34\% improvement compared to conventional P2P distributed
training methods. As for costs, our examination confirmed that the serverless
architecture could incur higher expenses, reaching up to 5.4 times more than
instance-based architectures. It is essential to consider that these higher
costs are associated with marked improvements in computation time, particularly
under resource-constrained scenarios. Despite the cost-time trade-off, the
serverless approach still holds promise due to its pay-as-you-go model.
Utilizing dynamic resource allocation, it enables faster training times and
optimized resource utilization, making it a promising candidate for a wide
range of machine learning applications.

通过将无服务器计算与点对点网络相结合，提出了一种在资源限制下实现高效并行梯度计算的新型架构，相比传统的点对点分布式训练方法，在梯度计算时间上取得了显著提升，同时也引入了更高的成本。

探索无服务器计算对点对点训练机器学习的影响

Exploring the Impact of Serverless Computing on Peer To Peer Training  Machine Learning

Decentralized, distributed storage offers a way to reduce the impact of data
silos as often fostered by centralized cloud storage. While the intentions of
this trend are not new, the topic gained traction due to technological
advancements, most notably blockchain networks. As a consequence, we observe
that a new generation of peer-to-peer data networks emerges. In this survey
paper, we therefore provide a technical overview of the next generation data
networks. We use select data networks to introduce general concepts and to
emphasize new developments. Specifically, we provide a deeper outline of the
Interplanetary File System and a general overview of Swarm, the Hypercore
Protocol, SAFE, Storj, and Arweave. We identify common building blocks and
provide a qualitative comparison. From the overview, we derive future
challenges and research goals concerning data networks.

本论文调查了下一代数据网络技术，重点介绍了 Interplanetary File System，Swarm，Hypercore Protocol，SAFE，Storj 和 Arweave 等技术，通过对这些技术的概述和比较，展示了 P2P 数据网络的基本概念和未来发展方向。

IPFS 与相关技术：下一代对等网络数据网络的质性比较

IPFS and Friends: A Qualitative Comparison of Next Generation  Peer-to-Peer Data Networks

Designing distributed and scalable algorithms to improve network connectivity
is a central topic in peer-to-peer networks. In this paper we focus on the
following well-known problem: given an $n$-node $d$-regular network for
$d=\Omega(\log n)$, we want to design a decentralized, local algorithm that
transforms the graph into one that has good connectivity properties (low
diameter, expansion, etc.) without affecting the sparsity of the graph. To this
end, Mahlmann and Schindelhauer introduced the random "flip" transformation,
where in each time step, a random pair of vertices that have an edge decide to
`swap a neighbor'. They conjectured that performing $O(n d)$ such flips at
random would convert any connected $d$-regular graph into a $d$-regular
expander graph, with high probability. However, the best known upper bound for
the number of steps is roughly $O(n^{17} d^{23})$, obtained via a delicate
Markov chain comparison argument.
Our main result is to prove that a natural instantiation of the random flip
produces an expander in at most $O(n^2 d^2 \sqrt{\log n})$ steps, with high
probability. Our argument uses a potential-function analysis based on the
matrix exponential, together with the recent beautiful results on the
higher-order Cheeger inequality of graphs. We also show that our technique can
be used to analyze another well-studied random process known as the `random
switch', and show that it produces an expander in $O(n d)$ steps with high
probability.

本文针对给定 n 个节点和 d 个邻居的无向图，提出了一种分布式的算法，使用随机翻转来改进图的连通性以获得新的拓展图，且经研究表明可以在 O（n ^ 2 d ^ 2 log n）的步骤内实现高概率有效。

局部边翻转法扩展算子

Expanders via Local Edge Flips

We introduce a model for decentralized networks with collaborating peers. The
model is based on the stable matching theory which is applied to systems with a
global ranking utility function. We consider the dynamics of peers searching
for efficient collaborators and we prove that a unique stable solution exists.
We prove that the system converges towards the stable solution and analyze its
speed of convergence. We also study the stratification properties of the model,
both when all collaborations are possible and for random possible
collaborations. We present the corresponding fluid limit on the choice of
collaborators in the random case. As a practical example, we study the
BitTorrent Tit-for-Tat policy. For this system, our model provides an
interesting insight on peer download rates and a possible way to optimize peer
strategy.

应用稳定匹配理论，我们提出了一种基于协作节点的去中心化网络模型，证明系统会收敛于唯一的稳定解，同时研究了模型的分层性质及其收敛速度，并在 BitTorrent Tit-for-Tat 策略上进行实证研究，为优化节点策略提供了新思路。