تمایز نرخ کاهشی تناسبی در یک صف FIFO
|کد مقاله||سال انتشار||تعداد صفحات مقاله انگلیسی||ترجمه فارسی|
|6392||2004||17 صفحه PDF||سفارش دهید|
نسخه انگلیسی مقاله همین الان قابل دانلود است.
هزینه ترجمه مقاله بر اساس تعداد کلمات مقاله انگلیسی محاسبه می شود.
این مقاله تقریباً شامل 4825 کلمه می باشد.
هزینه ترجمه مقاله توسط مترجمان با تجربه، طبق جدول زیر محاسبه می شود:
- تولید محتوا با مقالات ISI برای سایت یا وبلاگ شما
- تولید محتوا با مقالات ISI برای کتاب شما
- تولید محتوا با مقالات ISI برای نشریه یا رسانه شما
پیشنهاد می کنیم کیفیت محتوای سایت خود را با استفاده از منابع علمی، افزایش دهید.
Publisher : Elsevier - Science Direct (الزویر - ساینس دایرکت)
Journal : Computer Communications, Volume 27, Issue 18, 1 December 2004, Pages 1851–1867
The relative differentiated service model is one of several models proposed for service differentiation in networks [IEEE Network Sept/Oct (1999) 26]. In this model, an assurance is given that ‘higher classes will be better, or at least no worse than lower classes.’ This paper describes a relative loss rate differentiation scheme based on RED. The scheme is used for differentially dropping packets in a FIFO queue during times of congestion. The main idea is, if packet losses are unavoidable in the FIFO queuing system, then they should be distributed among the different service classes in the queue in inverse proportion to the service price or weight assigned to each class. The simulation studies using TCP traffic show that the scheme is very effective in ensuring relative loss rate differentiation between service classes.
In the relative differentiated service model, all traffic is grouped into N classes of service. For each class i, the service provided to class i (in terms of local (per-hop) performance measures of queuing delays and packet losses) will be better (or at least no worse) than the service provided to class (i−1), where 1<i≤N , ,  and . The ‘or at least no worse’ clause is included for levels of low network activity where all classes experience the same quality of service. The level of service in a class is relative to the other classes in the network and is not an absolute guarantee (in terms of end-to-end delay bound, bandwidth, etc.) since there is no admission control and resource reservation. The proportional differentiation model , ,  and  provides a way to control the quality spacing between classes locally at each hop, independent of the class loads. In this model, certain forwarding performance metrics are ratioed proportionally to the class differentiation parameters that the network operator chooses. Queuing delay and packet loss are two performance measures that can be used for proportional service differentiation. Let View the MathML sourcel¯i be the average loss rate for class i. Using this measure, the proportional differentiation model requires that the class loss rates be spaced as equation(1) View the MathML sourcel¯il¯j=σiσj,1≤i,j≤N. Turn MathJax on The parameters σi are the loss rate differentiation parameters and are ordered as σ1>σ2>…σN>0. In this particular definition, higher classes have better performance in terms of loss rates. The loss rate differentiation parameters σ1>σ2>…σN>0 provide the network operator with tuning knobs for adjusting the quality spacing between classes, independent of the class loads. Unlike other relative loss rate differentiation mechanisms , ,  and , this paper describes a relative loss rate differentiation mechanism designed around an active queue management (AQM) scheme. AQM schemes such as random early detection (RED)  and  provide network devices with some means to detect incipient congestion early and to convey congestion notification to the end-systems, allowing them to reduce their transmission rates before queues in the network overflow and packets are dropped. The basic RED scheme (and its newer variants) maintains an average of the queue length which it uses together with a number of queue thresholds to detect congestion. RED schemes drop incoming packets in a random probabilistic manner where the probability is a function of recent buffer fill history. The objective is to provide a more equitable distribution of packet loss, avoid the synchronization of flows  and , and at the same time improve the utilization of the network. The mechanism described in this paper is for relative loss rate differentiation in a FIFO queue only and does not cover delay differentiation or the coupling of the two. FIFO queuing has always been attractive because it is very simple to implement (packets are stored and served in the order in which they arrive). The limitation, however, is FIFO queuing provides little protection from high-bandwidth flows that consume a lot of bandwidth at the expense of other flows at the queue. A workaround to this problem is to rate-limit flows (e.g. using token buckets) at the network edge before the FIFO queuing point.
نتیجه گیری انگلیسی
We have described a PLR differentiation scheme that provides predictable and controllable short-term loss rate ratios between classes under low and high load. Loss rate differentiation in short timescales is not only beneficial to short-lived flows, but also provides assurance to the quality of service perceived by the higher service class user. The results show that service differentiation between classes is achieved independent of the total traffic load and the differences in traffic load in the classes. However, there were a few cases where the short-term variation in the loss rate ratios was not exactly met. These conditions tend to appear when the loss rate differentiation parameters have large differences, possibly making it more difficult to achieve the target values. The impact of user datagram protocol (UDP) traffic on the PLR differentiation mechanism is left as a future study. A prototype of the scheme is also in our plans. This will require modifying the software code of a device running our AQM scheme to include modules for detecting the class of a packet, computing the per-class drop probability pd,i(n), and dropping packets according to the packet drop routine described in Section 2.2.