Abstract
Admission control s oftware i s used t o make accept or deny decisions about incoming s ervice requests to avoid overload. Existing media streaming s oftware
includes only limited support f or admission control by allowing for predefined static rules. Such rules limit for example t he number of requests that are allowed to enter the s ystem during a certain time or define thresh-olds concerning the utilization l evel of a single resource such as network bandwidth. In media s treaming appli-cations, however, the bottleneck resource (CPU, Disk I/O, network bandwidth, etc.) might change over time depending on the current demand for different types of audio or video files. This paper proposes a m odel for adaptive admission control i n t he presence of multiple scarce resources. Opportunity costs for a service r equest are determined at the m oment of an incoming r equest and compared to the r evenue of a r equest i n order to make an accept/deny decision. Opportunity costs are based on resource utilization, service r esource require-ments, expected future demand for s ervices, and the r ev-enue per accepted service. The m odel allows rejection of service r equests early to reserve capacity required to perform future service r equests with higher revenues. We describe a number of experiments t o illustrate the benefits of adaptive admission control m odels over static admission control rules.


摘要 接納控制軟件是用來接受或拒絕決定輸入的服務請求軟件，從而避免過重的負擔。現(xiàn)有的流媒體只支持有限的軟件，其中就包括接納控制軟件。接納控制軟件使用他們預先定義的靜態(tài)的規(guī)則。這種規(guī)則限制例如靜態(tài)數(shù)據(jù)要求在一定時間內(nèi)都不準進入系統(tǒng)，或定義有關閾值利用水平的一個單一的資源，如：網(wǎng)絡帶寬。然而，流媒體應用的瓶頸資源(CPU，磁盤I / O，網(wǎng)絡帶寬等) 依靠目前的需求可能會隨時間改變?yōu)椴煌愋偷囊纛l或視頻文件。本文提出了一種新的模型自適應接納控制，在他面前的數(shù)倍稀缺的資源按機會成本來確定服務請求是當前請求，對照收入請求形成一個接受/拒絕的決定。機會成本是基于資源利用、服務資源需求預計未來需求與收入接受的服務。該模型允許拒絕服務請求過早的備用容量來完成未來服務請求與更高的收入。我們描述了一系列實驗現(xiàn)象，說明了在靜態(tài)的接納控制模型的自適應的接納控制規(guī)則的好處。