基于隨機(jī)檢測(cè)算法的網(wǎng)絡(luò)擁塞控制系統(tǒng)設(shè)計(jì)

李新煒

摘 ?要： 為解決由傳輸數(shù)據(jù)堆積而引發(fā)的網(wǎng)絡(luò)擁塞問(wèn)題，達(dá)到縮減網(wǎng)絡(luò)滯后延遲的目的，設(shè)計(jì)一種基于隨機(jī)檢測(cè)算法的新型網(wǎng)絡(luò)擁塞控制系統(tǒng)。利用LAIDS/LIDS主機(jī)隨機(jī)檢測(cè)架構(gòu)，對(duì)擁塞控制器、網(wǎng)絡(luò)數(shù)據(jù)過(guò)濾器進(jìn)行調(diào)節(jié)，完成新型系統(tǒng)的硬件運(yùn)行環(huán)境搭建。通過(guò)傳輸數(shù)據(jù)擁塞定義的方式對(duì)網(wǎng)絡(luò)中可能出現(xiàn)的擁塞情況進(jìn)行分類(lèi)處理，再根據(jù)具體判斷結(jié)果選擇合適的網(wǎng)絡(luò)擁塞控制方式，實(shí)現(xiàn)新型系統(tǒng)的軟件運(yùn)行環(huán)境搭建。結(jié)合軟、硬件結(jié)構(gòu)，完成基于隨機(jī)檢測(cè)算法的網(wǎng)絡(luò)擁塞控制系統(tǒng)設(shè)計(jì)。對(duì)比實(shí)驗(yàn)結(jié)果表明，與現(xiàn)有控制系統(tǒng)相比，應(yīng)用新型控制系統(tǒng)后單位時(shí)間內(nèi)傳輸數(shù)據(jù)堆積量的最大值不超過(guò)5.5×1013 TB，網(wǎng)絡(luò)滯后延遲平均值始終低于60 ms，網(wǎng)絡(luò)擁塞現(xiàn)象得到有效緩解。

關(guān)鍵詞： 隨機(jī)檢測(cè); 擁塞控制; LAIDS/LIDS架構(gòu); 運(yùn)行環(huán)境搭建; 擁塞數(shù)據(jù)分類(lèi); 系統(tǒng)設(shè)計(jì)

中圖分類(lèi)號(hào)： TN876?34; TP273 ? ? ? ? ? ? ? ? ? ?文獻(xiàn)標(biāo)識(shí)碼： A ? ? ? ? ? ? ? ? ? ?文章編號(hào)： 1004?373X（2019）09?0055?05

Design of network congestion control system based on random detection algorithm

LI Xinwei

（Dongchang College of Liaocheng University， Liaocheng 252000， China）

Abstract： In order to solve the network congestion problem caused by transmission data accumulation and reduce the network delay， a new network congestion control system based on random detection algorithm is designed. The LAIDS/LIDS host random detection architecture is utilized to adjust the congestion controller and network data filter to complete the hardware running environment establishment of the new system. The congestion situations possibly appeared in the network are classified by defining the transmission data congestion modes， and then the appropriate network congestion control mode is selected according to the specific judgment results to realize the software running environment establishment of the new system. The design of network congestion control system based on random detection algorithm is completed by combining the hardware and software structures. The contrast experimental results show that， in comparison with the available control systems， the new system′s maximum accumulation amount of transmission data in unit time is less than 5.5×1013 TB， and the average delay of network is always lower than 60 ms， and the network congestion phenomenon is effectively relieved.

Keywords： random detection; congestion control; LAIDS/LIDS architecture; running environment establishment; congestion data classification; system design

網(wǎng)絡(luò)擁塞是一種嚴(yán)重的網(wǎng)絡(luò)傳輸性能下降情況，當(dāng)分組交換網(wǎng)絡(luò)中的傳輸任務(wù)過(guò)于繁雜時(shí)，存儲(chǔ)轉(zhuǎn)發(fā)節(jié)點(diǎn)的傳輸能力會(huì)受到資源數(shù)據(jù)的嚴(yán)格限制，進(jìn)而導(dǎo)致鏈路帶寬急劇下降，網(wǎng)絡(luò)開(kāi)始出現(xiàn)存儲(chǔ)空間滿額、數(shù)據(jù)處理能力受損等問(wèn)題。從數(shù)據(jù)調(diào)節(jié)角度來(lái)看，當(dāng)網(wǎng)絡(luò)長(zhǎng)時(shí)間處于過(guò)載負(fù)荷狀態(tài)時(shí)，數(shù)據(jù)傳輸所必需的固有處理能力和信道容量始終不能得到滿足，而用戶在不了解網(wǎng)絡(luò)情況的狀況下，會(huì)對(duì)核心處理網(wǎng)絡(luò)發(fā)出大量的連接請(qǐng)求，這也是導(dǎo)致網(wǎng)絡(luò)擁塞現(xiàn)象惡性循環(huán)的主要原因[1?2]。當(dāng)網(wǎng)絡(luò)出現(xiàn)擁塞情況時(shí)，傳輸數(shù)據(jù)會(huì)大量堆積在物理信道中，進(jìn)而導(dǎo)致網(wǎng)絡(luò)出現(xiàn)明顯的滯后延遲[3]?，F(xiàn)有技術(shù)手段為解決此問(wèn)題，利用主動(dòng)隊(duì)列管理方法在發(fā)生擁塞情況前，截取大量的數(shù)據(jù)包信息，并通過(guò)調(diào)整執(zhí)行設(shè)備神經(jīng)元權(quán)值的方式，迫使數(shù)據(jù)丟包概率持續(xù)下降，達(dá)到抑制網(wǎng)絡(luò)擁塞的目的。但隨著科學(xué)技術(shù)手段的進(jìn)步，這種傳統(tǒng)方式始終不能達(dá)到理想的處理效果，因此引入隨機(jī)檢測(cè)算法，在對(duì)傳輸數(shù)據(jù)進(jìn)行定義、分類(lèi)處理的基礎(chǔ)上，建立一種新型的網(wǎng)絡(luò)擁塞控制系統(tǒng)。從實(shí)用性角度來(lái)看，這種新型系統(tǒng)有效解決了現(xiàn)有系統(tǒng)的遺留問(wèn)題，具備較強(qiáng)的實(shí)際推廣價(jià)值。