End-to-end latency

端到端延迟，从用户输入事件触发到系统画面响应显示之间的总延迟，由输入设备、操作系统、应用处理、网络、显示等多个环节的 partial delays 累加而成。

English Definition

"The total latency from the moment a user triggers an input event until the corresponding system response is displayed, encompassing all partial delays in the pipeline."

用法说明 · 针对中文母语者

强调「端到端」的整体性，常用来区别于单一环节的 partial latency。在产品/UX 写作中常简写为 E2E latency。

真实用例 · 3 条

"A system's end-to-end latency – the time between user input and system response – comprises several partial delays."

论文Pipeline extracted — review needed·Introduction 中正式定义术语

"The time from an input event being triggered by a user until a response by a system being displayed is commonly referred to as the system's end-to-end latency."

论文Pipeline extracted — review needed·Related Work 部分再次给出完整定义

"Therefore, end-to-end latency should not be viewed as a constant time delay, but rather a probability distribution of possible latencies for each individual interaction."

论文Pipeline extracted — review needed·强调 end-to-end latency 的统计分布特征

由 pipeline 自动采集，待人工 review