ABR algorithms dynamically select the bitrate of chunks based on network capacity. To estimate capacity, most use throughput prediction while recent works start to leverage delivery time prediction. We in this paper examine all components predictor for algorithms, i.e., input features, mapping function and output target. build an automated video streaming measurement platform, collect extensive dataset under various environments, containing 2500+ sessions. Through analysis, we find that...

Low-latency live streaming has imposed stringent latency requirements on video transport networks. In this paper we report the design and operation of Alibaba low-latency network, LiveNet. LiveNet builds a flat CDN overlay with centralized controller for global optimization. As part this, present our routing computation path assignment, as well fast data transmission architecture fine-grained control frames. The performance results obtained from three years demonstrate effectiveness in...

Content Centric Networking (CCN) performance by definition depends on the in-network caching efficiency. We propose CodingCache which utilizes network coding and random forwarding to improve efficiency under multipath forwarding. Its advantage is that existing strategies can be easily incorporated with it for better performance. evaluate extensive simulation experiments China Telecom topology a unique dataset consisting of video access logs from PPTV system. The results demonstrate compared...

Large cloud service providers have built an increasing number of geo-distributed data centers (DCs) connected by WAN to host their diverse services. While we seen a large body work on traffic engineering WAN, the characteristics production DC networks remain not well understood. In this paper, report network observed in Baidu's (DCN) that consists tens DCs. Baidu hosts both traditional services like Web and Computing, as emerging services, such Analytics, AI, Map. We analyze DCN from...

Increasing bandwidth demands of mobile video streaming pose a challenge in optimizing the Quality Experience (QoE) for better user engagement. Multipath transmission promises to extend network capacity by utilizing multiple wireless links simultaneously. Previous studies mainly tune packet scheduler multipath transmission, expecting higher QoE accelerating transmission. However, since Adaptive BitRate (ABR) algorithms overlook impact scheduling on throughput prediction, adaptive can even...

Named Data Networking (NDN) shifts the communication paradigm from quest of where content is to what be consumed. In such a new Internet architecture, transmission control mechanisms are particular importance and have carefully designed enable efficient data transmission. Existing work advocates use TCP-like reactive for NDN control. this paper, we show that statefull adaptive forwarding properties makes proactive possible. We achieve by using Explicit Congestion Notifications (ECN), which...

This article presents SOFIA, a service-oriented information-centric networking architecture. SOFIA is designed by exploring the design space between host abstraction and content abstraction. It can facilitate ICN to easily support various applications beyond retrieval. decouples flexible service processing efficient data transmission into two stack layers: network layers. Service requesting driven name processed intermediate routers according rules at layer. Applications build on layer...

Video delivery performance is the main factor that affects Internet video quality. Characterizing performance, especially throughput, can help content providers as well service (ISPs) in system optimization and network planning. Based on a unique dataset consisting of 20 million download speed measurements , this paper comprehensively studies throughput large-scale commercial video-on- demand (VoD) system. We observe user exhibits large variation over time day across provincial locations. In...

In real-time communication (RTC) systems, accurate bandwidth prediction is crucial for encoding and transmission strategies to optimize users' quality of experience (QoE) in various network environments. this paper, we propose an offline reinforcement learning (RL) method predict RTC video streaming. We use a representative algorithm, named Implicit Q-Learning (IQL), train the model. To improve performance, carefully preprocess given dataset redesign neural structure reward function....

Video providers heavily rely on geographically distributed content distribution networks (CDNs) to place video as close users possible, with an aim of improving quality and avoiding single point failure at the server side. The effectiveness CDNs is mostly dependent consumption patterns. Currently, are offering access from different platforms (e.g., mobile devices PC clients), which might result in distinct patterns finally affect efficiency CDN caching. Nevertheless, type effect Internet...

The prevalence of Internet video streaming challenges the design and operation modern networks. Content centric networking (CCN) has been proposed to address through ubiquitous in-network caching. While expected benefits include higher performance lower bandwidth consumption, CCN introduces new privacy issues at layer 3. This is because adversaries could infer content consumed by others checking cached data in routers. In this paper, we first analyze space improve both caching cache for...

TCP is an important factor affecting user-perceived performance of Internet applications. Diagnosing the causes behind issues in wild essential for better understanding current shortcomings TCP. This paper presents a flow analysis framework that classifies stalls. The forms basis tool publicly available to research community. We use our analyze packet-level traces three services (cloud storage, software download and web search) deployed by popular Chinese service provider. find as many 20%...

Mobile Live Streaming (MLS) services are now one of the most popular types mobile apps. They involve a (often amateur) user broadcasting content to potentially large online audience via unreliable networks (e.g., LTE). Although prior work has focused on viewer-side behavior, it is equally important study and improve broadcaster operations. Using detailed logs obtained from major MLS provider, we first conduct an in-depth measurement uploading behavior. Our key findings include wasteful...

Most congestion control algorithms (CCAs) are designed for specific network environments. As such, there is no known algorithm that achieves uniformly good performance in all scenarios flows. Rather than devising a one-size-fits-all (which likely impossible task), we propose system to dynamically switch between the most suitable CCAs flows This raises number of challenges, which address through design and implementation Antelope, can reconfigure stack use CCA individual We build machine...

Mobile messaging applications are already a part of everyone's daily life. Despite their prevalence, we have limited knowledge on user behavior and data transmission performance these services. In this paper, examine WeChat, one the largest mobile services with over 800 million active users. To end, analyze packet-level dataset captured in cellular network, containing 121GB WeChat 4.8M flows. Our analysis reveals several unique features including diurnal traffic pattern burst spikes,...

TCP is an important factor affecting user-perceived performance of Internet applications. Diagnosing the causes behind issues in wild essential for better understanding current shortcomings TCP. This paper presents a flow analysis framework that classifies stalls. The forms basis tool we use to analyze packet-level traces three services (cloud storage, software download, and web search) deployed by popular service provider. We find as many 20% flows are stalled half their lifetime....

Dual-band WiFi which supports both 2.4 GHz and 5 has been widely deployed, aiming to expand wireless capacity, eliminate serious interference in GHz. As the proportion of dual-band APs clients increase enormously, how select band access achieve considerable user experience is becoming essential network. Clients' native decisions that tend prefer will consequently cause leave notably idle. Through analyzing a unique dataset wild, we quantitatively study impact various factors on delay. We...

Large cloud service providers have built an increasing number of geo-distributed data centers (DCs) connected by Wide Area Networks (WANs). These DC-WANs carry both high-priority traffic from interactive services and low-priority bulk transfers. Given that a DC-WAN is expensive resource, often manage it via engineering algorithms rely on accurate predictions inter-DC (delay-sensitive) traffic. In this article, we perform large-scale measurement study Baidu. We measure how varies across their...

Video streaming is the most important mobile application today. To improve users' quality of experience (QoE), client player runs adaptive bitrate (ABR) algorithms that dynamically select for video chunks based on throughput or delivery time predictions. This paper aims to design an accurate predictor by investigating all its components, including input features, output target, and mapping function. We construct first theoretical framework reveals potential factors affecting chunk time....

Low-latency cloud rendering services use high-performance servers to provide mobile device users with exquisite graphics and convenient access experiences. Due the complexity of system diversity impacting factors, identifying bottlenecks has become a significant challenge. To demystify performance, we build an online measure latency distribution its key components.

A technique of linear networking utilizing Zigbee wireless sensor network is introduced in this paper, which applies the network, usually used short-distance area group, to long-distant monitoring system. The designed combined with relaying method under traditional communication mode, brings about multiple nodes. It able extend and relay transmission fast effectively datasheet. This can be utilized special field such as environment sensing highspeed railway track, power facility long-haul...