惯性聚合 高效追踪和阅读你感兴趣的博客、新闻、科技资讯
阅读原文 在惯性聚合中打开

推荐订阅源

T
The Blog of Author Tim Ferriss
I
InfoQ
H
Hackread – Cybersecurity News, Data Breaches, AI and More
aimingoo的专栏
aimingoo的专栏
小众软件
小众软件
有赞技术团队
有赞技术团队
J
Java Code Geeks
Apple Machine Learning Research
Apple Machine Learning Research
大猫的无限游戏
大猫的无限游戏
Engineering at Meta
Engineering at Meta
B
Blog RSS Feed
博客园_首页
Y
Y Combinator Blog
V
Visual Studio Blog
Google DeepMind News
Google DeepMind News
M
MIT News - Artificial intelligence
雷峰网
雷峰网
博客园 - 司徒正美
OSCHINA 社区最新新闻
OSCHINA 社区最新新闻
H
Help Net Security
P
Proofpoint News Feed
B
Blog
云风的 BLOG
云风的 BLOG
奇客Solidot–传递最新科技情报
奇客Solidot–传递最新科技情报

cs.DC updates on arXiv.org

DUAL-BLADE: Dual-Path NVMe-Direct KV-Cache Offloading for Edge LLM Inference Progressive Semantic Communication for Efficient Edge-Cloud Vision-Language Models Efficient, VRAM-Constrained xLM Inference on Clients Folding Tensor and Sequence Parallelism for Memory-Efficient Transformer Training & Inference DORA: A Scalable Asynchronous Reinforcement Learning System for Language Model Training AMMA: A Multi-Chiplet Memory-Centric Architecture for Low-Latency 1M Context Attention Serving RaMP: Runtime-Aware Megakernel Polymorphism for Mixture-of-Experts Spark Policy Toolkit: Semantic Contracts and Scalable Execution for Policy Learning in Spark Internet of Everything in the 6G Era: Paradigms, Enablers, Potentials and Future Directions PolyKV: A Shared Asymmetrically-Compressed KV Cache Pool for Multi-Agent LLM Inference A Survey on Split Learning for LLM Fine-Tuning: Models, Systems, and Privacy Optimizations ITAS: A Multi-Agent Architecture for LLM-Based Intelligent Tutoring Latency and Cost of Multi-Agent Intelligent Tutoring at Scale TACO: Efficient Communication Compression of Intermediate Tensors for Scalable Tensor-Parallel LLM Training FreeScale: Distributed Training for Sequence Recommendation Models with Minimal Scaling Cost CommFuse: Hiding Tail Latency via Communication Decomposition and Fusion for Distributed LLM Training A Taxonomy and Resolution Strategy for Client-Level Disagreements in Federated Learning Usable Agent Discovery for Decentralized AI Systems Cloud to Edge: Benchmarking LLM Inference On Hardware-Accelerated Single-Board Computers Data-Free Contribution Estimation in Federated Learning using Gradient von Neumann Entropy Shard the Gradient, Scale the Model: Serverless Federated Aggregation via Gradient Partitioning Promoting Simple Agents: Ensemble Methods for Event-Log Prediction GraphLeap: Decoupling Graph Construction and Convolution for Vision GNN Acceleration on FPGA AGNT2: Autonomous Agent Economies on Interaction-Optimized Layer 2 Infrastructure FedSIR: Spectral Client Identification and Relabeling for Federated Learning with Noisy Labels Stream-CQSA: Avoiding Out-of-Memory in Attention Computation via Flexible Workload Scheduling A Delta-Aware Orchestration Framework for Scalable Multi-Agent Edge Computing Federated Learning over Blockchain-Enabled Cloud Infrastructure Optimal Routing for Federated Learning over Dynamic Satellite Networks: Tractable or Not? Sherpa.ai Privacy-Preserving Multi-Party Entity Alignment without Intersection Disclosure for Noisy Identifiers
EBFT: Simplifying BFT Consensus Through Egalitarianism
Jianyu Niu, Runchao Han, Shengqi Liu, Fangyu Gai, Ivan Beschastn · 2020-12-03 · via cs.DC updates on arXiv.org

We present Egalitarian BFT (EBFT), a simple and high-performance framework of BFT consensus protocols for decentralized systems like blockchains. The key innovation in EBFT is egalitarian block generation: nodes randomly and non-interactively propose blocks containing client transactions, rather than relying on a leader to do so. Apart from deterministic safety and liveness guarantees standard in BFT protocols, the egalitarian design provides two novel features: (i) EBFT is resilient to attacks targeting the leader, such as bribery and targeted DoS attacks, and (ii) EBFT does not require any fail-over protocol to detect and replace the faulty leader. EBFT consists of three protocols: EBFT-Syn for synchronous networks, EBFT-PSyn for partially synchronous networks, and EBFT-Turbo that builds on EBFT for high performance. We implement EBFT and evaluate its performance on AWS. To compare EBFT with state-of-the-art BFT protocols, we build EBFT-PSyn based on Bamboo, an open-source platform for prototyping partially synchronous BFT protocols. We evaluate EBFT-PSyn and HotStuff on EC2 with up to 16 nodes. The evaluation shows that EBFT-PSyn achieves better throughput and latency than HotStuff. To demonstrate its simplicity and practicality, we build EBFT on the Go version of Bitcoin, btcd. We implemented EBFT-Syn, EBFT-PSyn and EBFT-Turbo in about 920 LoCs in total. This indicates that EBFT can be built on top of existing blockchains with relatively little effort. We evaluate these protocols on EC2 instances with up to 256 nodes. Our evaluation shows that EBFT-Syn (resp. EBFT-PSyn) achieves a latency of 6 (resp. 1) seconds, and an optimized version of EBFT-PSyn processes up to 3.6k transactions per second and has a latency of 8 seconds.