YPYajie Peng

Yajie Peng

*This is an algorithmically generated author record based on publicly available information.

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article


Yajie Peng has a total of 23 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Open Access

    RLive: Robust Delivery System for Scaling Live Streaming Services

    by Yu Tian, Gerui Lv, Qinghua Wu, Ruili Fang, Yajie Peng, Zhichen Xue, Rui Han, Chuanqing Lin, Xiaofei Pang, Ri Lu, et al.

  • Open Access

    M agpie : Improving the Efficiency of A/B Tests for Large Scale Video-on-Demand Systems

    by Hebin Yu, Haiping Wang, Chenfei Tian, Sandesh Dhawaskar Sathyanarayana, Shu Shi, Zhichen Xue, Shuaixin Yu, Haozhe Li, Yajie Peng, Xiaofei Pang, et al.

    With the exponential rise in video traffic, researchers and developers require more effective tools to validate the efficacy of designed algorithms for Video-on-Demand (VoD) system. However, traditional experimental platforms face two main challenges: a lack of realistic testing and the need for longer and significant effort. To overcome these limitations, we propose Magpie, an efficient experimental platform tailored for VoD systems. Magpie leverages a realistic operational setting, rapid testing, and high reproducibility to closely simulate online user environments without impacting production systems. Compared to conventional simulations, our evaluation demonstrates that Magpie reduces the disparity with online experiments by 85.6%. Deployed within our company-a leading video content provider in China-Magpie has efficiently validated over tens of algorithms, with 80% demonstrating enhanced performance in subsequent online tests.


  • Open Access

    Twist: A Multi-site Transmission Solution for On-demand Video Streaming

    by Haiping Wang, Ruixiao Zhang, Chaojun Li, Zhichen Xue, Yajie Peng, Xiaofei Pang, Yixuan Zhang, Shaorui Ren, Shu Shi

    Video traffic has witnessed exponential growth in recent years. As the cost optimization space of Content Delivery Network (CDN) has reached a plateau, content providers are expanding their network infrastructure to accommodate this surge. To address the cost issue, content providers have turned to cost-effective Peer-to-Peer Content Delivery Network (PCDN) solutions. However, the utilization of limited-capability PCDN nodes may lead to a decline in overall streaming performance compared to powerful CDN servers. To bridge this performance gap, we present Twist. Twist is a receiver-driven multi-site transport that leverages multiple cost-effective PCDN nodes to achieve performance equivalent to traditional CDN solutions. It incorporates a joint flow control to prevent throughput drops that can occur when using multiple content sources, and a proactive retransmission algorithm to handle frequent packet loss. Our evaluations demonstrate that Twist improves download speed on PCDN nodes by 1.71x-2.09x compared to traditional PCDN streaming solutions. Furthermore, extensive A/B testing verifies that Twist enables PCDN to achieve comparable download performance and Quality of Experience (QoE) as CDN. Over three years of commercial deployment, Twist has already served over 300 million users and has handled 35% of the video traffic, resulting in substantial cost savings for a world-leading content provider.


  • Multipath Smart Preloading Algorithms in Short Video Peer-to-Peer CDN Transmission Architecture

    by Dehui Wei, Jiao Zhang, Haozhe Li, Zhichen Xue, Yajie Peng, Rui Han

    The rapid growth of short video users has brought high traffic costs to content providers. Saving Content Distribution Network (CDN) costs while maintaining users’ playback Quality of Experience (QoE) is a significant problem for major short video platforms like Douyin (the Chinese counterpart of TikTok). Two commonly utilized cost-saving methods: PCDN (peer-to-peer CDN, fetching data from multiple low-cost but low-bandwidth edge devices with a multipath transmission protocol) and preloading control (minimizing unnecessary future playing data by downloading videos in segmented data ranges), each may hurt users’ experience respectively. Worse still, both methods combined have a synergistic negative effect over QoE. In this paper, we focus on experiences, algorithms, and prospects to solve this cost-QoE dilemma. We first introduce Douyin’s current PCDN multipath architecture and then review learning-based preloading techniques. Second, based on Reinforcement Learning (RL), we propose a Multipath-aware Smart Preloading algorithm, which consists of three schemes: one to decide the best size of the next range of preloaded data, another to design a water level valve algorithm that prioritizes preloading between currently playing video’s unfinished data and the next video’s beginning data, and the last one to determine the bitrate level of the next video. Douyin’s anonymous user feedback shows our Smart Preloading algorithm reduces traffic waste by ~26% while ensuring QoE. Third, we analyze and outlook the future of video systems, including trends in PCDN and other open issues.