H. Zhang and D. Ferrari. Rate controlled static-priority queueing. Proceedings of IEEE Infocom '93, pages 227-- 236, March 1993.  Home/Search   Document Details and Download   Summary   Related Articles   Check

....by the traffic behavior of the other sessions sharing a server. This guaranteed data rate generally requires some admission control mechanism to allocate the finite link capacity of the servers. Several rate based service disciplines have been proposed: Delay EDD [5] Jitter EDD [15] RCSP [17], PGPS [11, 12, 13, 14] Stop and Go [6, 7, 8] Hierarchical Round Robin [10] and VirtualClock [20] Among them, Zhang s VirtualClock in [20] is the only work that does not present an upper bound on end to end delay. Because of this, VirtualClock has been classified as a service discipline that ....

....test at connection establishment time [5] to avoid scheduling saturation, which can now occur even if bandwidth is not overbooked [5, 19] The schedulability test is then a compromise on the looser coupling between reserved rate and delay bound. Rate Controlled Static Priority Queueing (RCSP) [17] is a service discipline that avoids both framing strategies (as in Stop and Go and HRR) and sorted priority queues (that are used in all the other service disciplines studied here, including VirtualClock) by the separation of rate control and delay control in the design of the server, which ....

[Article contains additional citation context not shown here]

H. Zhang and D. Ferrari, "Rate-Controlled Static-Priority Queueing," In Proceedings of IEEE INFOCOM 1993, pp. 227-236.

.... putting traffic policing function at all network access points will prevent misbehaving users from affecting other users, traffic distortions due to network fluctuations suggest that protection should be implemented within the network using a rate based service discipline [10] In previous work [7, 8, 9], we have shown that, by having a server with two components, a schedule that can provide a local delay bound can be combined with a rate controller, and end to end delay bounds can obtained from arbitrary topology network with these types of servers. The second reason that FCFS is not enough is ....

....is that an FCFS server can only offer a single value of delay bound for all the channels. However, the performance requirements for integrated services networks will be diverse. It is important to support multiple classes of Quality of Service. More advanced service disciplines can achieve that [5, 6, 9]. The third reason that FCFS may not be enough is that more sophisticated service disciplines can provide better bounds than the FCFS discipline. 6 Conclusion In this paper, we showed that it is possible to provide deterministic delay bounds even when the sum of the peak rates of all the ....

Hui Zhang and Domenico Ferrari. Rate-controlled static priority queueing. In Proceedings of IEEE INFOCOM '93, San Francisco, California, April 1993.

....Virtual Clock [8] WFQ [18] PGPS [19] SCFQ [9] and WF Q [21] are classified as work conserving disciplines. Examples of non work conserving disciplines are Jitter EarliestDue Date (Jitter EDD) 30] Stop and Go [31] Hierarchical Round Robin [32] and Rate Controlled Static Priority (RCSP) [33]. Based on the internal architecture, the service disciplines can be also classified as frame based or sorted priority [6] In a frame based algorithm, there are frames with either fixed or variable length in time domain. Reservations for each session are made to determine the maximum amount of ....

H. Zhang and D. Ferrari, "Rate-Controlled Static-Priority Queueing," in Proc. of IEEE INFOCOM, (San Francisco, CA), pp. 227--236, Apr. 1993.

....the traffic behavior of the other sessions sharing the same server. This guaranteed data rate generally requires some admission con trol mechanism to allocate the finite link capacity of the servers. Several rate based service disciplines have been proposed: Delay EDD [5] Jitter EDD [22] RCSP [26], VirtualClock [29] PGPS [17, 18, 19, 20] Stop and Go [9, 10, 11] and Hierarchical Round Robin [13] All of these service disciplines provide an upper bound on end to end delay (this includes VirtualClock for which an upper bound on end to end delay was unknown until This work was supported in ....

....RCSP, see below) must be constrained to a scheme more restrictive than a token bucket filter. The traffic characterization specifies a minimum packet interarrival time Xn n, a minimum average packet interarrival time Xav e over an averaging interval of time 1, and a maximum packet length P. In [26], bandwidth is reserved at the peak rate implied by Xn n. This admission control is refined in [27] where both Xn n and Xav e are taken into consideration. In the Leave in Time scheme, band width may be reserved at the average rate, although a session may need to reserve more bandwidth than its ....

[Article contains additional citation context not shown here]

H. Zhang and D. Ferrari, "Rate-Controlled Static-Priority Queueing," In Proceedings of lEEE 1NFOCOM '93, pp. 227- 236, March 1993.

....behavior at its source node. Specifying modeling a real time application s traffic pattern is a challenging problem especially when source traffic has variable bit rate (VBR) characteristics, e.g. MPEG coded video. The leaky bucket model [1, 2] and the (X min ; X ave ; I ; Smax ) model [3, 4] are prototypical example input traffic specifications. In the leaky bucket model, the amount of connection i s traffic generated during time interval [t; is assumed to be upper bounded by oe i ae i ( Gamma t) That is, the size of an instantaneous traffic burst is limited by oe i and the ....

....delay requirements. So, the hardware implementation of a service discipline must be scalable with respect to the number and the type of real time channels. Recently, there have been several service disciplines proposed for per connection end toend delay guarantees in packet switching networks [3, 4, 6 12]. Virtual Clock [6] is a ratebased traffic control algorithm that can be used in packet switching networks. Although it can provide delay guaranteed service with an appropriate connection admission control [13] it tends to penalize a connection that has received better than guaranteed service ....

[Article contains additional citation context not shown here]

H. Zhang and D. Ferrari, "Rate-controlled static-priority queueing," in Proc. of IEEE INFOCOM, pp. 227--236, 1993.

.... best effort service: Integrated Services (IntServ) 12] and Differentiated Services (DiffServ) In the network data plane of the IntServ architecture, scheduling schemes such as Weighted Fair Queueing (WFQ) 3] Virtual Clock (VC) 17] and RateControlled Earliest Deadline First (RC EDF) [15] have been proposed to support guaranteed QoS. In the control plane, a signaling protocol RSVP [16] is required for admission control and resource reservation. While IntServ provides QoS guarantees, it requires per flow management at core routers, which places an unbearable burden on core routers. ....

H. Zhang and D. Ferrari, "Rate-controlled Static-priority Queueing", Proceedings of IEEE INFOCOM'93, April, 1993.

....function of the traffic controller is to keep the cell arrival rate at the scheduler below the token generation rate ae i . It holds the incoming cells until their supposed arrival times, and then transfers them into the scheduler. This supposed arrival time is called the logical arrival time in [6, 9]. The logical arrival time of an incoming cell is calculated based on that of the previous cell of the same channel. Thus, the logical arrival time of the k cell at the n node, X k;n , is calculated as: X k;n = A 1;n k = 1 max(X k Gamma1;n ; A k;n ) k 2; 2.2) where A k;n is the ....

....UNI. The second and third terms are the sum of the maximumqueueing delays and the sum of propagation delays, respectively. Our approach shares the same idea of splitting the traffic controllers and the scheduler to provide the bounded link delays in Rate Controlled Static Priority Queueing (RCSP) [6]. Since TCRM allocate a different priority level to each real time channel, they can provide arbitrary link delay bounds, whereas RCSP provides a preset finite number of link delay bounds. This enables the allocation of a wide range of bandwidths in TCRM at the expense of more complex channel ....

[Article contains additional citation context not shown here]

H. Zhang and D. Ferrari, "Rate-controlled staticpriority queueing", in Proc. of IEEE INFOCOM 93, pp. 227--236, 1993.

....deadlines are considered lost. Using the user specified delay bound D, the maximum packet size S max , the maximum burst size B max , and the maximum packet arrival rate G max , one can establish a hard real time channel in a pointto point network using one of the schemes in [9] 12] 13] [31]. Considering the differences between multiaccess bus networks and point to point networks, we will first develop a new scheme for establishing hard real time channels on multiaccess buses with a more general traffic model than that used for point to point networks. We will then identify the ....

H. Zhang and D. Ferrari, "Rate-Controlled Static-Priority Queueing, " Technical Report TR-92-003, Computer Science Division, Univ. of California at Berkeley, Jan. 1992.

....networks, because their packet multiplex ers do not differentiate between real time and nonreal time traffic, nor among real time connections themselves. Several packet multiplexing techniques have been proposed to provide different QoS for different real time applications [1] 2] 3] 4] [5], 6] 7] 8] 9] In particular, we proposed Traffic Controlled Rate Monotonic Priority Scheduling (TCRM) 9] as a cell multiplexing scheme to achieve deterministic real time communication in ATM networks. The prime attraction of the TCRM lies in its simplicity and, hence, makes it suitable ....

H. Zhang and D. Ferrari, "Rate-Controlled Static-Priority Queueing, " Proc. IEEE INFOCOM, pp. 227-236, 1993.

....the transmission of all the data, the end user requests the (virtual) connection to be terminated, and then the network disconnects it and releases its reserved resources. For efficient real time communication service, we must have good packet scheduling, source traffic characterization [5, 11], and efficient QoS routing which is the focus of this paper. Without efficient QoS routing, a network may fail to find a route and reject the request for establishing a new connection even when there exists a route with enough resources to provide the requested QoS. This QoS routing is therefore ....

H. Zhang and D. Ferrari, "Rate-controlled static-priority queueing," in Proc. of IEEE INFOCOM, pp. 227--236, 1993.

....where traffic from different connections is multiplexed. Drawing on the similarities with CPU scheduling, classical real time scheduling policies such as the First Come First Serve (FCFS) Earliest Deadline First (EDF) Generalized Processor Sharing (GPS) Fair Queuing (FQ) etc. are employed [1, 13, 20, 21, 22]. However, to reduce the cost and the complexity of design and implementation, most of the currently available switches use simple scheduling policies such as First Come First Serve (FCFS) with hardly any provision for implementing more complex scheduling policies. In this paper, we focus on the ....

....increases with a proper choice of leaky bucket parameters at the UNI. While we focus on traffic regulation for meeting end to end deadlines, our work also complements much of the previous studies which essentially concentrate on designing and analyzing scheduling policies for ATM switches [1, 4, 5, 7, 9, 10, 13, 17, 18, 20, 21, 22, 23]. A modified FCFS scheduling scheme called F IFO was proposed and studied in [1] The switch scheduling policy called Stop and Go is presented in [7] A virtual clock scheduling scheme in which cells are prioritized by a virtual time stamp assigned to them, is discussed in [22] The use of ....

[Article contains additional citation context not shown here]

H. Zhang and D. Ferrari. Rate-controlled static priority queueing. In Proceedings of IEEE Infocom'93, pages 227--236, Mar. 1993.

No context found.

H. Zhang and D. Ferrari. Rate-controlled static-priority queueing. In IEEE INFOCOM'93, pages 227-236, April 1993.

....chooses packets in FCFS order from the highest priority non empty queue. Non real time packets are serviced only when there are no real time packets; their order is not specified. The combination of a ratecontroller and an SP scheduler is called a Rate Controlled Static Priority (RCSP) server [91], which is shown in Figure 3.6. Non Real Time Queue Output m 1 Priority Level Real time Packet Queues . Real time) Input Traffic Regulator 1 Regulator 2 Regulator n (Non Real time) Regulated Traffic One regulator for each of the n connections Rate Controller Scheduler ....

....connections Rate Controller Scheduler Figure 3.6: Rate Controlled Static Priority Queueing By limiting the number of connections at each priority level via certain admission control conditions, the waiting time of each packet at a priority level can be bounded. The following theorem is used in [91] to give the admission control condition for an SP scheduler. Theorem 3.3 Let d ) be the delay bounds associated with each of the n priority levels, respectively, in a Static Priority scheduler. Assume that C q is the set of connections at level q, and, that the j j ) Also ....

Hui Zhang and Domenico Ferrari. Rate-controlled static priority queueing. In Proceedings of IEEE INFOCOM'93, pages 227--236, San Francisco, California, April 1993.

No context found.

H. Zhang and D. Ferrari. Rate controlled static-priority queueing. Proceedings of IEEE Infocom '93, pages 227-- 236, March 1993.

No context found.

H. Zhang, and D. Ferrari, "Rate-Controlled Static-Priority Queueing", INFOCOM, pp. 227--236", 1993

No context found.

H. Zhang and D. Ferrari. Rate controlled static-priority queueing. Proceedings of IEEE Infocom '93, pages 227-- 236, March 1993.

No context found.

H. Zhang, D. Ferrari. Rate-Controlled Static Priority Queueing, Proc. IEEE INFOCOM, San Francisco, CA, September 1993.

No context found.

H. Zhang and D. Ferrari, "Rate-Controlled Static Priority Queueing," in Proc. of the IEEE INFOCOM 1993.

No context found.

H. Zhang and D. Ferrari, "Rate-controlled static priority queueing," IEEE Conference on Computer Communications '93, 1993, pp. 227-236.

No context found.

H. Zhang and D. Ferrari. Rate-controlled static-priority queueing. In Proceedings of IEEE INFOCOM '93, volume 1, pages 227--236, San Francisco, CA, March 1993.

No context found.

H. Zhang and D. Ferrari, "Rate-controlled static priority queueing," in Proc. IEEE INFOCOM '93, San Francisco, CA, Apr. 1993, pp. 227-236.

No context found.

Zhang H, Ferrari D (1993) Rate-controlled static priority queueing. In: Proceedings of INFOCOM '93, volume 2, pp 227--236

No context found.

H. Zhang and D. Ferrari. Rate-controlled static priority queueing. In Proceedings of IEEE INFOCOM'93, pages 227--236, San Francisco, April 1993.

No context found.

H. Zhang and D. Ferrari, "Rate-controlled static-priority queueing," in Proc. of IEEE INFOCOM, pp. 227--236, 1993.

No context found.

H. Zhang and D. Ferrari, "Rate-controlled static-priority queueing," in Proc. of IEEE INFOCOM 93, pp. 227--236.

First 50 documents  Next 50

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