The Gilbert-Elliot (GE) channel is a commonly-accepted model for packet erasures in networks. Streaming codes are a class of packet-level erasure codes designed to provide reliable communication over the GE channel. The design of a streaming code may be viewed as a two-step process. In the first, a more tractable, delay-constrained sliding window (DCSW) channel model is considered as a proxy to the GE channel. The streaming code is then designed to reliably recover from all erasures introduced by the DCSW channel model. Simulation is typically used to evaluate the performance of the streaming code over the original GE channel, as analytic performance evaluation is challenging. In the present paper, we take an important first step towards analytical performance evaluation. Recognizing that most, efficient constructions of a streaming code are based on the diagonal embedding or horizontal embedding of scalar block codes within a packet stream, this paper provides upper and lower bounds on the block-erasure probability of the underlying scalar block code when operated over the GE channel.
翻译:Gilbert-Elliot (GE) 频道是一个通用的网络封隔模式。 串流代码是一组包级封隔代码, 旨在为 GE 频道提供可靠的通信。 串流代码的设计可被视为一个两步过程。 首先, 一个更可移动的、 延迟限制的滑动窗口( DCS) 频道模式被视为 GEE 频道的代理。 然后, 流流码的设计可以可靠地从 DCS 频道模型引入的所有封隔中恢复。 模拟通常用于评估原始 GE 频道流线代码的性能, 因为分析性能评估具有挑战性。 在本文中, 我们迈出了一个重要的第一步, 分析性绩效评估。 认识到流码的多数有效构建是基于包流流流流流的单层嵌入或水平嵌入, 本文在运行 GE 频道时, 提供了底片块块隔断概率的上限和下限。