Exponential Decay Queue Thresholds for Optimal Buffer Sharing

نویسندگان

  • Boran Gazi
  • Zabih Ghassemlooy
چکیده

In an N x N shared buffer switch architecture, a common buffer space is shared among output ports, which provides flexibility in resource allocation. However the problem arises when the buffer space is dominated by one or more output ports in case of imbalanced input traffic. Thus this results in an increased packet loss rates due to unfair sharing of buffer space. For this reason buffer management policies are essential to regulate the buffer space occupied by active output ports. In this paper we introduce a dynamic threshold policy, which is inspired from the exponential decay. This policy suggests that the buffer space accessed by an output port decays exponentially with respect to its queue length. In this work we compare the packet loss rate performance of this scheme with the dynamic thresholds policy by carrying out a simulation study. We have shown that the performance of the exponential decay function policy can be improved by considering unused buffer space. It is found that the improved policy achieves the same packet loss rate as the dynamic thresholds policy with an improved buffer utilisation. The results suggest that the best threshold based buffer management policy is the one that allows buffer overflow in the case of hotspot packets.

برای دانلود متن کامل این مقاله و بیش از 32 میلیون مقاله دیگر ابتدا ثبت نام کنید

ثبت نام

اگر عضو سایت هستید لطفا وارد حساب کاربری خود شوید

منابع مشابه

Importance Sampling for Weighted-Serve-the-Longest-Queue

This paper considers buffer overflow probabilities for stable queueing systems with one server and different classes of arrivals. The service priority is given to the class of customers whose current weighted queue size is the largest (weighted-serve-the-longest-queue policy). We explicitly identify the exponential decay rate for the rare-event probabilities of interest, and construct asymptoti...

متن کامل

Large deviations for complex buffer architectures: the short-range dependent case

This paper considers Gaussian flows multiplexed in a queueing network, where the underlying correlation structure is assumed to be short-range dependent. Whereas previous work mainly focused on the FIFO setting, this paper addresses overflow characteristics of more complex buffer architectures. We subsequently analyze the tandem queue, a priority system, and generalized processor sharing. In a ...

متن کامل

Fast Simulation of a Queue Fed by a Superposition of Many (heavy-tailed) Sources

We consider a queue fed by a large number, say n, on–off sources with generally distributed onand off-times+ The queueing resources are scaled by n: The buffer is B[nb and the link rate is C[nc+The model is versatile+ It allows one to model both long-range-dependent traffic ~by using heavy-tailed on-periods! and short-rangedependent traffic ~by using light-tailed on-periods!+ A crucial performa...

متن کامل

Large Deviations for Small Buffers: An Insensitivity Result

This article focuses on a queue fed by a large number of ‘semi-Markov modulated fluid sources’, e.g., on/off sources with on and off-times that have general distributions. The asymptotic regime is considered in which the number of sources grows large, and the buffer and link rate are scaled accordingly. We aim at characterizing the exponential decay rate of the buffer overflow probability for t...

متن کامل

Asymptotic performance of queue length based network control policies

In a communication network, asymptotic quality of service metrics specify the probability that the delay or buffer occupancy becomes large. An understanding of these metrics is essential for providing worst-case delay guarantees, provisioning buffer sizes in networks, and to estimate the frequency of packet-drops due to buffer overflow. Second, many network control tasks utilize queue length in...

متن کامل

ذخیره در منابع من


  با ذخیره ی این منبع در منابع من، دسترسی به آن را برای استفاده های بعدی آسان تر کنید

برای دانلود متن کامل این مقاله و بیش از 32 میلیون مقاله دیگر ابتدا ثبت نام کنید

ثبت نام

اگر عضو سایت هستید لطفا وارد حساب کاربری خود شوید

عنوان ژورنال:

دوره   شماره 

صفحات  -

تاریخ انتشار 2005