C. Courcoubetis, G. Fouskas, and R.R. Weber. On the performance of an effective bandwidths formula. International Teletraffic Conference, 1994.  Home/Search   Document Details and Download   Summary   Related Articles   Check

....b. An admission control scheme should be employed for EF and AF traffic. 2. Bandwidth Model 2.1 Topology and traffic sources In our scenario we assume that VoIP and Videoconference sources generate EF traffic. For VoIP sources an ONOFF model with exponential ON and OFF durations can be used [1] [4]. As in [1] it is assumed that ON and OFF periods have an average of 1.004s and 1.587s respectively. This corresponds to a 38.53 talk spurt cycle, as recommended by ITU T specification for conversational speech. The peak transmission rate is 64kbps and the average is 24.8kbps. The model that is ....

....i 2 2 ) s s and e is the allowed Bit Error Rate (BER) You may note that (1) is also computationally simple and depends only on the BER and the means and variances of the bit rates generated by individual sources. There are also other approaches for the calculation of effective bandwidth [4] [5] but they are only valid for Markovian sources, assume the existence of very large buffers [4] in order to function and if the number of individual sources is not small, they give a greater overestimation of the actual demand than the bufferless model since they don t take into account the ....

[Article contains additional citation context not shown here]

Costas Courcoubetis, George Fouskas, Richard Weber. On the Performance of an Effective Bandwidths Formula. Proc. 14 th Int. Teletraffic Cong., June 1994

....with (k; j) entry given by E[exp( A i n ) Y i n 1 = j) j Y i n = k) Therefore, we deduce from Proposition 2.4 that P (X x) C( e Gamma x ; 8x 0 if N X i=1 log( i ( c: 4. 2) The quantity c i ( log( i ( is called the effective bandwidth of the process (A i n ) n [14, 24, 29, 32, 39, 29]. A similar result was presented by Chang and Cheng in [11, Example 3.4] but with a different coefficient C( The coefficient in [11] denoted as Gamma( is given by Gamma( max i;j r i ( r j ( where (r 1 ( r K ( T is the (positive) right eigenvector of the matrix ....

C. Courcoubetis, G. Fouskas, R. Weber, "On the Performance of an Effective Bandwidth Formula", Proc. of ITC'14, pp. 201-212, Antibes, June 1994, Elsevier, Amsterdam, eds. J. Labetoulle, J. Roberts.

....its most general form is not easy to compute. However, for most common arrival processes such as on off fluid sources or MMPP, simple expressions can be derived for efficient computation. Moreover, several approximations to the effective bandwidth function using a few parameters have been proposed [CFW94, Cha93]. These parameters can be estimated either off line or on line. Effective bandwidth approximation to the buffer loss probability (7) may not be good enough in many situations, for example, when the system load is low. This phenomenon has been pointed out even when the effective bandwidth ....

C. Courcoubetis, G. Fouskas and R. Weber, On the Performance of an Effective Bandwidths Formula, ITC 14, J. Labetoulle and J. W. Roberts Ed., 1994, pp. 201-212.

....that captures the behavior of the tail of the response time distribution at a multiplexer. The call admission problem is then solved by checking whether the effective bandwidth of the aggregate user population, including the new user, exceeds the service capacity. The reader is referred to [16, 23, 17, 12, 5] for treatments of this problem. As an application of our model we will obtain results for the source models commonly encountered in the study of network and multimedia server performance. We apply our results to several systems that have received considerable prior attention. These include: i) a ....

C. Courcoubetis, G. Fouskas, R. Weber, "On the Performance of an Effective Bandwidth Formula", Proc. of ITC'14, pp. 201-212, Antibes, June 1994, Elsevier, Amsterdam, Eds. J. Labetoulle, J. Roberts.

....by the algorithm. The results indicate that for these bursty variable bit rate video sources, the scheme is accurate enough to capture most of the available statistical multiplexing gain and consequently, attains utilizations considerably higher than those of effective bandwidth techniques [4, 9]. The remainder of this paper is organized as follows. In Section 2, we develop the rate variance admission control algorithm and resource allocation scheme. Next, in Section 3, we investigate our scheme s implications for traffic models and admission control tests in multi service networks. In ....

.... the RV (t) characterization and fi, we calculate the maximum number of admissible flows for the QoS (d; p) pair using Equation (6) To compare our approach with previous resource allocation schemes, we also perform admission control experiments using effective bandwidth admission control tests of [4]. An effective bandwidth scheme as in [4, 9] reserves a bandwidth for each flow according to its stochastic properties as well as the required loss probability p. Once the effective bandwidth of a flow is determined, which we denote by E j (p) the admission control test checks that P j E j (p) ....

[Article contains additional citation context not shown here]

C. Courcoubetis, G. Fouskas, and R. Weber. On the performance of an effective bandwidths formula. In Proceedings of 14th International Teletraffic Congress, pages 201--212, June 1994.

....its mean and peak rate, its amount of burstiness, of the QoS required by the connection, and is in general less than its peak rate (and of course above its mean rate) It is beyond the scope of the present paper to derive explicit formulas of effective bandwidths. Sophisticated mathematical tools [CW95,CFW94] show that for large systems (large in the amount of network resources available, i.e. bandwidth and buffer at the switches, and hence capable of a large amount of multiplexing) effective bandwidths 8 can be effectively computed for a large variety of traffic streams, where the QoS criterion ....

C. Courcoubetis, G. Fouskas, and R. Weber. On the Performance of an Effective Bandwidth Formula. In proceedings of ITC'94, 1994.

....may be associated with stationary sources. These bandwidths have the advantage that they are simple functions of the average rate of a source and its index of dispersion. The index of dispersion may be estimated from data, and evaluated for simple Markov modulated fluid models using theorem 3. In [4], we have presented, with Fouskas, experimental evidence that in the cases of Markov modulated and autoregressive source models the use of these bandwidths can achieve a desired quality of service and a good utilization of the switch. There are many other issues relating to the use of effective ....

C. Courcoubetis, G. Fouskas, and R.R. Weber. On the performance of an effective bandwidths formula. International Teletraffic Conference, 1994.

....may be associated with stationary sources. These bandwidths have the advantage that they are simple functions of the average rate of a source and its index of dispersion. The index of dispersion may be estimated from data, and evaluated for simple Markov modulated fluid models using theorem 3. In [4], we have presented, with Fouskas, experimental evidence that in the cases of Markov modulated and autoregressive source models the use of these bandwidths can achieve a desired quality of service and a good utilization of the switch. There are many other issues relating to the use of effective ....

C. Courcoubetis, G. Fouskas, and R.R. Weber. On the performance of an effective bandwidths formula. International Teletraffic Conference, 1994.

....isn t too overestimated, especially when the buffer size is large. In [7] the CLP is approximated in terms of the first and the second moment of the aggregate arrival rate and the size of the buffer. The approximation is effective only when the aggregate arrival rate is normally distributed. In [3], the CLP is given by a formula involving the mean rate of each source, its index of dispersion and the buffer size. That formula takes the form of the first two terms in an asymptotic expansion that is accurate for large buffer sizes. In [4] the CLP is given by a function, whose form is known by ....

C. Courcoubetis, G. Fouskas, and R. Weber. On the Performance of an Effective Bandwidths Formula. In Proceedings of the ITC 14, 1994.

No context found.

C. Courcoubetis, G. Fouskas and R. Weber, On the Performance of an Effective Bandwidths Formula, ITC 14, J. Labetoulle and J. W. Roberts Ed., 1994, pp. 201-212.

Online articles have much greater impact   More about CiteSeer.IST   Add search form to your site   Submit documents   Feedback  

CiteSeer.IST - Copyright Penn State and NEC