Search

Your search keyword '"Geng, Jinkun"' showing total 44 results
Start Over Author "Geng, Jinkun"
44 results on '"Geng, Jinkun"'

1. Design and Implementation of a Scalable Financial Exchange in the Public Cloud

Author

Haseeb, Muhammad, Geng, Jinkun, Butler, Ulysses, Hao, Xiyu, Duclos-Cavalcanti, Daniel, Sivaraman, Anirudh, and Narayana, Srinivas

Subjects
Computer Science - Networking and Internet Architecture
Abstract

Financial exchanges are migrating to the cloud, but the best-effort nature of the public cloud is at odds with the stringent latency requirements of exchanges. We present Jasper, a system for meeting the networking requirements of financial exchanges on the public cloud. Jasper uses an overlay tree to scalably multicast market data from an exchange to ~1000 participants with low latency (250 microseconds) and a 1-microsecond difference in data reception time between any two participants. Jasper reuses the same tree for scalable inbound communication (participants to exchange), augmenting it with order pacing and a new priority queue, Limit Order Queue (LOQ), to efficiently handle bursts of market orders. Jasper achieves better scalability and 50% lower latency than the AWS multicast service. During bursty market activity, LOQ nearly doubles the order processing rate.

Published
2024

2. Nezha: Deployable and High-Performance Consensus Using Synchronized Clocks

Author

Geng, Jinkun, Sivaraman, Anirudh, Prabhakar, Balaji, and Rosenblum, Mendel

Subjects
Computer Science - Distributed, Parallel, and Cluster Computing, Computer Science - Databases, Computer Science - Networking and Internet Architecture, C.2.1, C.2.4, C.4
Abstract

This paper presents a high-performance consensus protocol, Nezha, which can be deployed by cloud tenants without any support from their cloud provider. Nezha bridges the gap between protocols such as Multi-Paxos and Raft, which can be readily deployed and protocols such as NOPaxos and Speculative Paxos, that provide better performance, but require access to technologies such as programmable switches and in-network prioritization, which cloud tenants do not have. Nezha uses a new multicast primitive called deadline-ordered multicast (DOM). DOM uses high-accuracy software clock synchronization to synchronize sender and receiver clocks. Senders tag messages with deadlines in synchronized time; receivers process messages in deadline order, on or after their deadline. We compare Nezha with Multi-Paxos, Fast Paxos, Raft, a NOPaxos version we optimized for the cloud, and 2 recent protocols, Domino and TOQ-EPaxos, that use synchronized clocks. In throughput, Nezha outperforms all baselines by a median of 5.4x (range: 1.9-20.9x). In latency, Nezha outperforms five baselines by a median of 2.3x (range: 1.3-4.0x), with one exception: it sacrifices 33% latency compared with our optimized NOPaxos in one test. We also prototype two applications, a key-value store and a fair-access stock exchange, on top of Nezha to show that Nezha only modestly reduces their performance relative to an unreplicated system. Nezha is available at https://github.com/Steamgjk/Nezha., Comment: Accepted by 49th International Conference on Very Large Data Bases (VLDB 2023)

Published
2022
Full Text
View/download PDF

4. Diffusion Policies as Multi-Agent Reinforcement Learning Strategies

Author

Geng, Jinkun, Liang, Xiubo, Wang, Hongzhi, Zhao, Yu, Goos, Gerhard, Founding Editor, Hartmanis, Juris, Founding Editor, Bertino, Elisa, Editorial Board Member, Gao, Wen, Editorial Board Member, Steffen, Bernhard, Editorial Board Member, Yung, Moti, Editorial Board Member, Iliadis, Lazaros, editor, Papaleonidas, Antonios, editor, Angelov, Plamen, editor, and Jayne, Chrisina, editor

Published
2023
Full Text
View/download PDF

5. Xenos : Dataflow-Centric Optimization to Accelerate Model Inference on Edge Devices

Author

Zhang, Runhua, Jiang, Hongxu, Tian, Fangzheng, Geng, Jinkun, Li, Xiaobin, Ma, Yuhang, Zhu, Chenhui, Dong, Dong, Li, Xin, Wang, Haojie, Goos, Gerhard, Founding Editor, Hartmanis, Juris, Founding Editor, Bertino, Elisa, Editorial Board Member, Gao, Wen, Editorial Board Member, Steffen, Bernhard, Editorial Board Member, Yung, Moti, Editorial Board Member, Wang, Xin, editor, Sapino, Maria Luisa, editor, Han, Wook-Shin, editor, El Abbadi, Amr, editor, Dobbie, Gill, editor, Feng, Zhiyong, editor, Shao, Yingxiao, editor, and Yin, Hongzhi, editor

Published
2023
Full Text
View/download PDF

9. Jasper: Scalable and Fair Multicast for Financial Exchanges in the Cloud

Author

Haseeb, Muhammad, Geng, Jinkun, Butler, Ulysses, Hao, Xiyu, Duclos-Cavalcanti, Daniel, Sivaraman, Anirudh, Haseeb, Muhammad, Geng, Jinkun, Butler, Ulysses, Hao, Xiyu, Duclos-Cavalcanti, Daniel, and Sivaraman, Anirudh

Abstract

Financial exchanges have recently shown an interest in migrating to the public cloud for scalability, elasticity, and cost savings. However, financial exchanges often have strict network requirements that can be difficult to meet on the cloud. Notably, market participants (MPs) trade based on market data about different activities in the market. Exchanges often use switch multicast to disseminate market data to MPs. However, if one MP receives market data earlier than another, that MP would have an unfair advantage. To prevent this, financial exchanges often equalize exchange-to-MP cable lengths to provide near-simultaneous reception of market data at MPs. As a cloud tenant, however, building a fair multicast service is challenging because of the lack of switch support for multicast, high latency variance, and the lack of native mechanisms for simultaneous data delivery in the cloud. Jasper introduces a solution that creates an overlay multicast tree within a cloud region that minimizes latency and latency variations through hedging, leverages recent advancements in clock synchronization to achieve simultaneous delivery, and addresses various sources of latency through an optimized DPDK/eBPF implementation -- while scaling to a thousand receivers. Jasper outperforms a prior system, CloudEx, and a commercial multicast solution provided by Amazon Web Services. We present different deployment models and their performance impact. A deployment model where MPs and the exchange do not have to trust each other is realized using confidential computing.

Published
2024

10. HiPower: A High-Performance RDMA Acceleration Solution for Distributed Transaction Processing

Author

Zhang, Runhua, Cheng, Yang, Geng, Jinkun, Wang, Shuai, Gao, Kaihui, Shen, Guowei, Goos, Gerhard, Founding Editor, Hartmanis, Juris, Founding Editor, Bertino, Elisa, Editorial Board Member, Gao, Wen, Editorial Board Member, Steffen, Bernhard, Editorial Board Member, Woeginger, Gerhard, Editorial Board Member, Yung, Moti, Editorial Board Member, Tang, Xiaoxin, editor, Chen, Quan, editor, Bose, Pradip, editor, Zheng, Weiming, editor, and Gaudiot, Jean-Luc, editor

Published
2019
Full Text
View/download PDF

14. A Novel Clustering Algorithm for Database Anomaly Detection

Author

Geng, Jinkun, Ye, Daren, Luo, Ping, Lv, Pin, Akan, Ozgur, Series editor, Cao, Jiannong, Series editor, Coulson, Geoffrey, Series editor, Dressler, Falko, Series editor, Ferrari, Domenico, Series editor, Gerla, Mario, Series editor, Kobayashi, Hisashi, Series editor, Palazzo, Sergio, Series editor, Sahni, Sartaj, Series editor, Shen, Xuemin (Sherman), Series editor, Stan, Mircea, Series editor, Xiaohua, Jia, Series editor, Zomaya, Albert, Series editor, Bellavista, Paolo, Series editor, Thuraisingham, Bhavani, editor, Wang, XiaoFeng, editor, and Yegneswaran, Vinod, editor

Published
2015
Full Text
View/download PDF

15. Predicting Severity of Software Vulnerability Based on Grey System Theory

Author

Geng, Jinkun, Ye, Daren, Luo, Ping, Hutchison, David, Series editor, Kanade, Takeo, Series editor, Kittler, Josef, Series editor, Kleinberg, Jon M., Series editor, Mattern, Friedemann, Series editor, Mitchell, John C., Series editor, Naor, Moni, Series editor, Pandu Rangan, C., Series editor, Steffen, Bernhard, Series editor, Terzopoulos, Demetri, Series editor, Tygar, Doug, Series editor, Weikum, Gerhard, Series editor, Wang, Guojin, editor, Zomaya, Albert, editor, Martinez, Gregorio, editor, and Li, Kenli, editor

Published
2015
Full Text
View/download PDF

16. Nezha

Author

Geng, Jinkun, primary, Sivaraman, Anirudh, additional, Prabhakar, Balaji, additional, and Rosenblum, Mendel, additional

Published
2022
Full Text
View/download PDF

20. CloudEx

Author

Ghalayini, Ahmad, primary, Geng, Jinkun, additional, Sachidananda, Vighnesh, additional, Sriram, Vinay, additional, Geng, Yilong, additional, Prabhakar, Balaji, additional, Rosenblum, Mendel, additional, and Sivaraman, Anirudh, additional

Published
2021
Full Text
View/download PDF

30. Accelerating End-to-End Deep Learning Workflow With Codesign of Data Preprocessing and Scheduling.

Author

Cheng, Yang, Li, Dan, Guo, Zhiyuan, Jiang, Binyao, Geng, Jinkun, Bai, Wei, Wu, Jianping, and Xiong, Yongqiang

Subjects
DEEP learning, FIELD programmable gate arrays, WORKFLOW, BOTTLENECKS (Manufacturing), ARTIFICIAL neural networks, IMAGE processing
Abstract

In this article, we investigate the performance bottleneck of existing deep learning (DL) systems and propose DLBooster to improve the running efficiency of deploying DL applications on GPU clusters. At its core, DLBooster leverages two-level optimizations to boost the end-to-end DL workflow. On the one hand, DLBooster selectively offloads some key decoding workloads to FPGAs to provide high-performance online data preprocessing services to the computing engine. On the other hand, DLBooster reorganizes the computational workloads of training neural networks with the backpropagation algorithm and schedules them according to their dependencies to improve the utilization of GPUs at runtime. Based on our experiments, we demonstrate that compared with baselines, DLBooster can improve the image processing throughput by 1.4× – 2.5× and reduce the processing latency by 1/3 in several real-world DL applications and datasets. Moreover, DLBooster consumes less than 1 CPU core to manage FPGA devices at runtime, which is at least 90 percent less than the baselines in some cases. DLBooster shows its potential to accelerate DL workflows in the cloud. [ABSTRACT FROM AUTHOR]

Published
2021
Full Text
View/download PDF

33. DLBooster

Author

Cheng, Yang, primary, Li, Dan, additional, Guo, Zhiyuan, additional, Jiang, Binyao, additional, Lin, Jiaxin, additional, Fan, Xi, additional, Geng, Jinkun, additional, Yu, Xinyi, additional, Bai, Wei, additional, Qu, Lei, additional, Shu, Ran, additional, Cheng, Peng, additional, Xiong, Yongqiang, additional, and Wu, Jianping, additional

Published
2019
Full Text
View/download PDF

39. Dante

Author

Li, Zhetao, primary, Gui, Fei, additional, Geng, Jinkun, additional, Li, Dan, additional, Wang, Zhibo, additional, Li, Junfeng, additional, Cheng, Yang, additional, and Zafar, Usama, additional

Published
2018
Full Text
View/download PDF

40. HiPS

Author

Geng, Jinkun, primary, Li, Dan, additional, Cheng, Yang, additional, Wang, Shuai, additional, and Li, Junfeng, additional

Published
2018
Full Text
View/download PDF

42. LOS

Author

Huang, Yukai, primary, Geng, Jinkun, additional, Lin, Du, additional, Wang, Bin, additional, Li, Junfeng, additional, Ling, Ruilin, additional, and Li, Dan, additional

Published
2017
Full Text
View/download PDF

Catalog

Books, media, physical & digital resources
See catalog results