Guanghui Zhang,Jack Y. B. Lee,Ke Liu,Haibo Hu,Vaneet Aggarwal
Abstract:Live video streaming has seen tremendous growth in the past decade. An important fact in live streaming is that the demand for low playback-latency inherently conflicts with the desire for high QoE. This requires different types of live services to seek different latency-QoE tradeoffs according to their service-requirements. However, our investigations revealed that it is fundamentally difficult for existing streaming algorithms to keep consistent latency in changing network conditions, let alone achieve the service-desired latency-QoE tradeoff. To tackle the challenge, this article develops a novel framework called Flexible Latency Aware Streaming (FLAS) that not only can achieve consistent low latency, but also control the latency-QoE tradeoff flexibly. Specifically, FLAS generates a set of adaptation logics offline, each optimized for a candidate tradeoff point, then selects the most appropriate one to run online. We first show how FLAS can be applied to optimizing the existing algorithms, then developed a novel Genetic Programming approach to fully exploit FLAS's potential. Extensive evaluations show that FLAS can precisely control latency all the way down to 1s and achieve substantially higher QoE than state-of-the-arts. FLAS can be readily implemented into real streaming platforms, offering a practical and reliable solution for live-streaming services.
What problem does this paper attempt to address?
### Problems the paper attempts to solve
This paper aims to solve the inherent conflict between playback latency and Quality of Experience (QoE) in live - video streaming. Specifically:
1. **The contradiction between low - latency requirements and high QoE**: In live - video streaming, low playback latency (usually required to be within a few seconds) is a very important performance metric, but low latency often leads to a decline in video quality, thus affecting the user experience. Conversely, the pursuit of high - quality video will increase the transmission latency, thereby increasing the playback latency.
2. **Differences in requirements for different services**: Different live - streaming services have different requirements for latency and QoE. For example, highly interactive live - streaming (such as live - selling, interactive live - shows) requires shorter latency but can tolerate lower video quality; while one - way broadcasting (such as news, concerts) can accept longer latency but requires higher video quality.
3. **Limitations of existing algorithms**: Existing live - streaming algorithms are difficult to maintain consistent low latency under changing network conditions, let alone achieve the latency - QoE trade - off required by services. The performance of these algorithms fluctuates greatly under different network conditions and cannot meet the needs in practical applications.
To address these challenges, the paper proposes a new framework - Flexible Latency - Aware Streaming (FLAS), which can not only achieve consistent low latency but also flexibly control the latency - QoE trade - off. Specifically, FLAS generates a series of adaptation logics offline, each logic is optimized for a candidate trade - off point, and then selects the most appropriate logic to execute according to network conditions at runtime. This enables FLAS to maintain stable latency under different network conditions and maximize QoE.
### Main contributions
1. **Revealing problems with existing algorithms**: Through extensive evaluation, the inconsistent performance of existing live - streaming algorithms under different network conditions has been revealed.
2. **Proposing the FLAS framework**: FLAS is a unified framework that can optimize existing learning - based algorithms. The paper shows how to apply FLAS to the L2AC algorithm, significantly improving performance.
3. **Introducing the genetic programming method**: To overcome the limitations of deep reinforcement learning, the paper proposes a method based on genetic programming to further improve the performance of FLAS.
### Experimental results
1. **Trade - off between QoE and latency**: FLAS - L2AC not only provides a continuous QoE - latency trade - off trajectory but also significantly improves QoE.
2. **High performance under low latency**: In the low - latency domain (< 3 seconds), FLAS - L2AC achieves a higher QoE than LAPAS.
3. **Spatial - temporal robustness**: FLAS - L2AC performs very stably under different network conditions and can maintain consistent low latency within 40 days.
In conclusion, FLAS provides a practical and reliable solution that can flexibly control latency and QoE in live - video streaming to meet the needs of different services.