R. Jain. The osu scheme for congestion avoidance using explicit rate indication. ATM Forum 94-0883, 1994. 426  Home/Search   Document Not in Database   Summary   Related Articles   Check

....this service does not provide any guarantees on quality of service (QoS) it attempts to minimize the cell loss at the expense of delay. The dynamic sharing of bandwidth between competing users has to be achieved via appropriate set of distributed controls. It has been researched in the ATM forum [2,11 13]. Under the current ATM standards, the source generates special cells, called resource manage ment (RM) cells, at a rate proportional to its current data cell rate. The feedback is available only when the source generates it. The RM cell travels to the destination, where it is turned around and ....

R.Jain, S. Kalyanaraman, and R.Viswanathan, "The OSU scheme for congestion avoidance using explicit rate indication," ATM Forum/94-883, September, 1994.

....to carry such traffic. ABR service was defined as a new service class for ATM networks to utilize the remaining resources not used by other types of services (e.g. CBR, VBR) There has been much effort devoted to the design of ABR flow control. For a survey on ABR rate based flow control see [12, 14, 28, 29, 56] and references therein. Several issues arise in reviewing the rate based control algorithms: ffl The explicit rate control mechanism has fast convergence characteristics. ffl A simple algorithm is preferred to make the complexity of explicit rate control reasonable. ffl In a large scale ....

R. Jain, S. Kalyanaraman, and R. Viswanathan. The OSU scheme for congestion avoidance using explicit rate indication. ATM Forum, 94-0883, Sep. 1994.

....entries represents the number of active connections. In this way, the rates of all sources are adjusted through RM cells in one round trip time when there is one congested switch in the network. EPRCA , proposed in [23] also uses this kind of scheme, and its simplified version can be found in [24]. In EPRCA , congestion is detected by estimating the traffic load at the switch rather than by using a threshold value in the cell buffer. For this, the switch is provided with an interval timer and counts the number of cells received during a fixed time interval. The source end system is also ....

R. Jain, S. Kalyanaraman, and R. Viswanathan, "The OSU scheme for congestion avoidance using explicit rate indication," ATM Forum Contribution 94-0883, September 1994.

....the authors study the effect of large propagation delays on feedback mechanisms. These papers provide much insight into the difficulties of using delayed feedback for flow control in high speed data networks. Feedback mechanisms for ABR mechanisms have also been studied extensively (see [10] 20] [12]) In [10] and [20] Hluchyj et.al. and Newman analyze increase decrease algorithms when the feedback is in the form of a CI bit. In [12] 2 Jain et.al. study explicit rate mechanisms. However, these papers study the effectiveness of feedback in a deterministic set up where the dynamics of the ....

....of using delayed feedback for flow control in high speed data networks. Feedback mechanisms for ABR mechanisms have also been studied extensively (see [10] 20] 12] In [10] and [20] Hluchyj et.al. and Newman analyze increase decrease algorithms when the feedback is in the form of a CI bit. In [12], 2 Jain et.al. study explicit rate mechanisms. However, these papers study the effectiveness of feedback in a deterministic set up where the dynamics of the VBR load on the system are not modeled explicitly. The outline of the paper is as follows. In Section 2 we describe a basic and show it be ....

R. Jain, S. Kalyanaramanan, and R. Vishwanathan. The OSU scheme for congestion avoidance using explicit rate indication. Technical report, Ohio State University, 1994.

.... dynamic networks, where there is often need to decrease the rates of some sessions in order to accomodate new entering sessions; moreover, the optimistic approach seems to get closer to speci cations for rate based ow control algorithms suggested in the ATM Forum on Trac Management (see, e.g. [26, 39]) A signi cant component of any rate based ow control algorithm, whether conservative or optimistic, is its scheduler, the abstract mechanism it uses to decide which session to increase In the model we consider, it is assumed that each session s lifetime is in nite; this assumption is ....

.... has been scheduled for an increase, a control message loops around its path and calculates on its way the minimum share of bandwidth the session may take from the excess capacities of links along the path; this may require possible adjustments to the rates of con icting sessions (see, e.g. [7, 8, 9, 26, 27]) Roughly speaking, the convergence complexity of a rate based ow control algorithm is the number of rate adjustments performed in the worst case till all sessions are terminated and max min fairness is reached. Since reaching max min fairness fast is vital for ecient utilization of the network ....

R. Jain, S. Kalyanaraman and R. Viswanathan, \The OSU Scheme for Congestion Avoidance Using Explicit Rate Indication," Technical Report ATM-FORUM/95-0883, The Ohio State University, September 1994.

....entries represents the number of active connections. In this way, the rates of all sources are adjusted through RM cells in one round trip time when there is one congested switch in the network. EPRCA , proposed in [23] also uses this kind of scheme, and its simplified version can be found in [24]. In EPRCA , congestion is detected by estimating the traffic load at the switch rather than by using a threshold value in the cell buffer. For this, the switch is provided with an interval timer and counts the number of cells received during a fixed time interval. The source end system is also ....

R. Jain, S. Kalyanaraman, and R. Viswanathan, "The OSU scheme for congestion avoidance using explicit rate indication," ATM Forum Contribution 94-0883, September 1994.

....size Figure 1: Early Packet Discard (EPD) algorithm schemes have been proposed, and they differ from one another by the various switch mechanisms employed to determine the congestion information and the mechanisms by which the source rates are adjusted. Many of the promising rate based proposals [1, 8, 9, 10, 7, 12, 15] use a technique called intelligent marking originally proposed in [13] see also [14] which achieves fairness without the need of per VC queueing nor per VC accounting. In this paper, we shall present simulation results comparing the performance of TCP over ABR service with rate based ....

....proposed in [11] in our simulation studies. For ABR congestion control schemes, we select one of the most promising proposals in the ATM Forum. There will be no significant difference in the simulation results if we choose to use other equally promising rate based schemes such as those proposed in [1, 4, 8, 9, 10, 12]. We would like to point out that in terms of hardware complexity, existing ABR rate based congestion control schemes proposed in the ATM Forum are more expensive to implement in ATM switches than UBR service with EPD schemes. The motivation for our work is to understand the degree of improvement ....

R. Jain, S. Kalyanaraman, and R. Viswanathan. The OSU scheme for congestion avoidance using explicit rate indication. ATM Forum Contribution 94-0883, September 1994.

....introduced by Anna Charny but requires only (1) operations. This is reached by keeping track of the available bandwidth for the ABR connections. For more detailed description of the required computations see [4] 2. 2 The OSU Congestion Avoidance Scheme This scheme was developed by Jain et al. [5] at the Ohio State University (OSU) In this scheme, the switches measure their input rates over a fixed averaging interval that is counted in cells [6] Thereby, the input rate is calculated as follows: Input rate = number of received cells length of measurement interval With this calculated ....

R. Jain, S. Kalyanaraman, and R. Viswanthan, "The OSU scheme for congestion avoidance using explicit rate indication," Tech. Rep. 94-883, ATM Forum, Sept. 1994.

....no first order linear approximation. It becomes extremely difficult to prove their stability and to explore conditions under which the control schemes have acceptable performance. Computer simulation hence becomes the only evaluation option. In the subsequent discussion, the OSU scheme proposed by [3] is studied as an example to illustrate the potential stability problem of the schemes designed in ad hoc manner. 4.1 Network Configuration S5 S1 D5 D1 VBR VBR . 10ms 50ms 10ms 50ms ATM Mux . Figure 2: Network with a single congestion link and five ABR connections As shown in Fig. 2, a ....

R. Jain, et al, "The OSU Scheme for Congestion Avoidance using Explicit Rate Indication," ATM Forum/94-0883, Sept. 1994.

No context found.

R. Jain, S. Kalyanaraman and R. Viswanathan, "The OSU Scheme for Congestion Avoidance using Explicit Rate Indication," Tech Report OSU-CISRC-1/96-TR02, Dept. of CIS, Ohio State University, 1996.

.... are treated as if the load level is z= 1 Delta) and the overloading sources are treated as if the load level is z= 1 Gamma Delta) If (OCR in cell FairShare) LAF in cell Max(LAF in cell, 1 Delta) g else LAF in cell Max(LAF in cell, 1 Gamma Delta) g We prove in [8] that this algorithm guarantees that the system consistently moves towards fair operation. We note that all the switch steps are O(1) w.r.t. the number of VCs. The value of OCR in the cell is corelated to z when the control cell enters the switch queue. The value of z may change before the ....

....sample configuration tests the steady state as well as the transient response of the scheme. It also shows convergence to fairshares. The TCRs of sources, bottleneck queue length and link utilization are shown in Figure 3. A complete set of simulation configurations and results may be found in [8]. 6 20 40 60 80 100 120 140 160 180 0 10000 20000 Transmitted Cell Rate Time in micro seconds TCR for S1 TCR for S2 (a) Transmitted Cell Rates 20 40 60 80 100 120 140 160 0 10000 20000 30000 Queue Length Time in micro seconds Cells in Q to Sw1 Sw2 link (b) Queue ....

R. Jain, Shiv Kalyanaraman and Ram Viswanathan, "The OSU Scheme for Congestion Avoidance using Explicit Rate Indication", OSU Technical Report OSU-CISRC-1/96TR02, Dept. of CIS, Ohio State University, January 1996. AF-TM refers to ATM Forum Traffic Management sub-working group contributions. All our papers and ATM Forum contributions are available through http://www.cis.ohio-state.edu/~jain/ 8

....complexity of switch design. We present one such scheme in this paper, the Ohio State University (OSU) scheme. The OSU scheme is named thus because it was a follow on to MIT scheme [1, 2] This paper provides an overview of the OSU scheme. The scheme is discussed and analysed in greater detail in [9]. In this paper, we first describe the control cell format, the source, switch and destination algorithms. We then highlight some of the unique features of the OSU scheme which have hence become commonly accepted parts of switch schemes or have been adopted by the standard. The simulation section ....

.... are treated as if the load level is z= 1 Delta) and the overloading sources are treated as if the load level is z= 1 Gamma Delta) If (OCR in cell FairShare) LAF in cell Max(LAF in cell, z (1 Delta) g else LAF in cell Max(LAF in cell, z (1 Gamma Delta) g We prove in [9, 10] that this algorithm guarantees that the system consistently moves towards fair operation. We note that all the switch steps are O(1) w.r.t. the number of VCs. 2.3.4 What Load Level to Use The OCR in the control cell is corelated to z when the control cell enters the switch queue. The value of ....

[Article contains additional citation context not shown here]

R. Jain, Shiv Kalyanaraman and Ram Viswanathan, "The OSU Scheme for Congestion Avoidance using Explicit Rate Indication," OSU Technical Report OSU-CISRC-1/96TR02, 4 Dept. of CIS, Ohio State University, January 1996.

No context found.

R. Jain, S. Kalyanaraman and R. Viswanathan, "The OSU Scheme for Congestion Avoidance using Explicit Rate Indication," Tech Report OSU-CISRC-1/96-TR02, Dept. of CIS, Ohio State University, 1996.

....However, the early rate based proposals were found to suffer problems of fairness [4] One approach to fixing the problems is by keeping track of the rate for each VC (per VC accounting) 7] though such an approach also increases the switch complexity. Many of the current rate based proposals [2, 10, 13, 14, 22, 26] use an alternative technique called intelligent marking, originally proposed in [24] see also [25] which achieves fairness without the need of per VC queueing nor per VC accounting. An approach that integrates the desirable features of the credit based proposals and the ratebased schemes was ....

R. Jain, S. Kalyanaraman, and R. Viswanathan. The OSU scheme for congestion avoidance using explicit rate indication. ATM Forum Contribution 94-0883, September 1994.

....relevant to them. These are specific to the OSU scheme. The detailed pseudo code of BECN option is given in appendix B. 5 Simulation Results The OSU scheme has been extensively simulated and the results have been presented in ATM Forum Contributions. A complete set of simulations can be found in [9]. Here we present one sample result for the basic scheme and its extensions. This configuration consists of four VCs and three switches as shown in Figure 4. The second link is shared by VC3 and VC4. However, because of the first link, VC3 is limited to a throughput of 1 3 the link rate. VC4 ....

R. Jain, S. Kalyanaraman and R. Viswanathan, "The OSU Scheme for Congestion Avoidance using Explicit Rate Indication," Tech Report OSU-CISRC-1/96-TR02, Dept. of CIS, Ohio State University, January 1996.

No context found.

R. Jain. The osu scheme for congestion avoidance using explicit rate indication. ATM Forum 94-0883, 1994. 426

No context found.

R. Jain. The osu scheme for congestion avoidance using explicit rate indication. ATM Forum94-0- , 1994. 426

No context found.

R. Jain. The osu scheme for congestion avoidance using explicit rate indication. ATM Forum94-( , 1994.

No context found.

R. Jain, S. Kalyanaraman, and R. Viswanathan, "The OSU Scheme for Congestion Avoidance Using Explicit Rate Indication," AF-TM 94-0883, September 1994.

No context found.

R. Jain, S. Kalyanaraman, and R. Viswanathan, "The OSU Scheme for Congestion Avoidance Using Explicit Rate Indication," AF-TM 94-0883, September 1994.

No context found.

R. Jain, et al, "The OSU Scheme for Congestion Avoidance Using Explicit Rate Indication," ATM Forum 94-0883.

No context found.

R. Jain. The osu scheme for congestion avoidance using explicit rate indication. ATM Forum 94-0883, 1994. 426

No context found.

R. Jain. The osu scheme for congestion avoidance using explicit rate indication. ATM Forum 94-0883, 1994. 426

No context found.

R. Jain, S. Kalyanarman, and R. Viswanathan, "The OSU Scheme for Congestion Avoidance Using Explicit Rate Indication", ATM Forum Contribution 94-0883, September 1994. 22

No context found.

R. Jain, S. Kalyanaraman, and R. Viswanathan, "The OSU Scheme for Congestion Avoidance Using Explicit Rate Indication," AF-TM 94-0883, September 1994.

First 50 documents

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