B.-K. Choi, D. Xuan, R. Bettati, et al., "Scalable QoS guaranteed communication services for real-time applications," in Proc. of IEEE ICDCS'00, Apr. 2000, pp. 180--187.  Home/Search   Document Details and Download   Summary   Related Articles   Check

....compared to FCFS, and by as much 28.1 compared to deadline based scheduling policy. 5. Related Work There are significant research results on real time communications on single hop wired LANs (e.g. 24] 25] multi hop wired LANs (e.g. 14] ATM (e.g. 17] 18] 19] and the Internet (e.g. [6][13] 22] 21] A good survey about real time network architecture for packet switched network is [3] However, there have been few published works on real time multi hop sensor networks, which has significant different constraints from previous real time networks. Directed diffusion [10] is a ....

B.Choi,D.Xuan,C.Li,R.Bettati,andW.Zhao,"Scalable QoS Guaranteed Communication Services for RealTime Applications,"inProc. of the IEEE International Conf. on Distributed Computing Systems, (ICDCS), April 2000.

....performance of individual packets, it must be carefully addressed. In the DiffServ domain, applications are differentiated by their classes. Accordingly, many previous studies assume that priorities are assigned on a class basis only, where all the flows in a class are assigned the same priority [4]. We study more generalized priority assignment algorithms, where the flows in a class may be assigned different priorities and flows from different classes may share a same priority. While the proposed algorithms are still relatively simple and efficient, we find that they are effective in ....

....at Server #. ### ### # is the available bandwidth for traffic with priority no higher than #. Derivation of Inequality (3) will be discussed in Subsection VB. At this point we would like to make the following observations on Theorem 1: # In a previously derived delay computation formula in [4], there was an implicit limitation on the relationship of traffic classes and priorities: One traffic class can only have a single priority, and one priority can only be assigned to a single class traffic. The new delay formula removes this limitation, and gives more flexibility when ....

B. Choi, D. Xuan, C. Li, R. Bettati and W. Zhao, Scalable QoS Guaranteed Communication Services for Real-Time, IEEE Proceedings of ICDCS, 2000.

....performance of individual packets, it must be carefully addressed. In Diffserv domain, applications are differentiated by their classes. Accordingly, many previous studies assume that priorities are assigned based on classes only, typically, the flows in a class are assigned by the same priority [4]. Flows from different classes are given different priority. We study more generalized priority assignment algorithms. We allow that the flows in a class may be assigned by different priorities and flows from different classes may share a same priority. While the proposed algorithms are still ....

....this kind of information and just depend on , p;k , and L. However, the values of these parameters are available at the time when the system is (re )configured. Hence, the delay computation formula is insensitive to the flow population information. In the delay computation formula in [4], there was an implicit limitation on the relationship of traffic classes and priorities, that is, one traffic class can only have a single priority, and one priority can only be assigned to a single class traffic. This limitation puts a constraint on differentiating service. The new delay formula ....

[Article contains additional citation context not shown here]

B. Choi, D. Xuan, C. Li, R. Bettati and W. Zhao, Scalable QoS Guaranteed Communication Services for Real-Time, Proc. of ICDCS, 2000.

....the soft state and refresh. This makes Boomerang inappropriate for hard real time services, because the soft state and refresh can not guarantee the allocated resources over the life time of the connection. Some work has been done on guarantees of end to end delay deadlines within DS architecture [27, 30]. The basic idea in [27] is to move per ow information from core routers to edge routers by relying on estimation. The key points of the estimation technique are: 1) each core router relies on dynamic packet status information for admission control and packet scheduling, 2) dynamic packet status ....

....router relies on dynamic packet status information for admission control and packet scheduling, 2) dynamic packet status information is carried with each packet. With this information, the core router estimates each ow s dynamic information such as end to end delay. In contrast to this, work in [30], signi cantly reduces the run time overhead of admission control by doing some of the computation o line, that is, during network design or con guration. Also, by using a simple packet forwarding mechanism such as class based static priority scheduling, its runtime overhead is lower than that ....

[Article contains additional citation context not shown here]

B. Choi, D. Xuan, C. Li, R. Bettati, W. Zhao, \Scalable QoS Guaranteed Communication Services for Real-Time Applications," The 20th IEEE International Conference on Distributed Computing Systems, Taipei, Taiwan, pp. 180-187, April 2000.

....delay requirement. In the networks with rate based schedulers, such as weighted fair queuing (WFQ) 18, 19] virtual clock (VC) 20] etc. delay requirement can be directly mapped to bandwidth requirement. In the networks with priority driven schedulers, we can follow the approach presented in [21, 22] to transform delay requirement into an upper bound on bandwidth that need to be reserved, hence converting the delay related QoS to bandwidth based QoS. 18 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 0 5 10 15 20 25 30 35 40 45 50 Flow Arrival Rate Figure 7: Average Number of Retrials A ....

B. Choi, D. Xuan, C. Li, R. Bettati, W. Zhao, \Scalable QoS Guaranteed Communication Services for Real-Time Applications", Proc. of IEEE ICDCS'00, 2000.

....performance of individual packets, it must be carefully addressed. In the DiffServ domain, applications are differentiated by their classes. Accordingly, many previous studies assume that priorities are assigned on a class basis only, where all the flows in a class are assigned the same priority [4]. We study more generalized priority assignment algorithms, where the flows in a class may be assigned different priorities and flows from different classes may share a same priority. While the proposed algorithms are still relatively simple and efficient, we find that they are effective in ....

....at Server k. k p;k k is the available bandwidth for traffic with priority no higher than p. Derivation of Inequality (3) will be discussed in Subsection VB. At this point we would like to make the following observations on Theorem 1: In a previously derived delay computation formula in [4], there was an implicit limitation on the relationship of traffic classes and priorities: One traffic class can only have a single priority, and one priority can only be assigned to a single class traffic. The new delay formula removes this limitation, and gives more flexibility when ....

B. Choi, D. Xuan, C. Li, R. Bettati and W. Zhao, Scalable QoS Guaranteed Communication Services for Real-Time, IEEE Proceedings of ICDCS, 2000.

No context found.

B. Choi, D. Xuan, C. Li, R. Bettati, W. Zhao, "Scalable QoS Guaranteed Communication Services for Real-Time Applications," The 20th IEEE International Conference on Distributed Computing Systems, Taipei, Taiwan, pp. 180-187, April 2000.

....methods in the sink tree paradigm. A simulation study is provided in Section 6. Finally, conclusions and future work are described in Section 7. 2 Previous Work Some work has been done on admission control of real time applications with end to end delay constraint within the DS architecture [15, 22, 23, 25]. The basic idea in [15] is to move per flow information from core routers to edge routers by relying on dynamic packet status carried with each packet. So the core router estimates each flow s dynamic information such as end to end delay for admission control and packet scheduling based on the ....

....or marking assumed routing overhead not mentioned zero side effect thrashing signaling for resource sharing end to end delay not mentioned calculated off line deploy ability probing sw at host sink tree sw at router Table 1. Host Based vs. Network Based Endpoint Admission Control work in [22] significantly reduces the run time overhead of admission control by doing some of the computation off line, that is, during network design or configuration. Other than these, the maximum achievable resource utilization in the DS model has been studied in [23] with a heuristic route selection ....

[Article contains additional citation context not shown here]

B. Choi, D. Xuan, C. Li, R. Bettati, W. Zhao, "Scalable QoS Guaranteed Communication Services for Real-Time Applications," The 20th IEEE International Conference on Distributed Computing Systems, Taipei, Taiwan, pp. 180187, April 2000.

.... network, given the information about the amount of bandwidth reserved for the aggregated trac of each class on the links in the network and the source ow model of that trac, we also derive a delay formula to estimate the worst case queueing delay su ered by packets belonging to a particular class [4]. A wealth of other delay formulas for other scheduling policies is available in the literature [1, 5, 6, 15, 19, 20] 5 NetCamo: Camou aging Network Trac 5.1 Motivation Based on the approach described in the previous sections and [7] we have realized NetCamo, a system NetCamo System ....

B. Choi, D. Xuan, C. Li, R. Bettati and W. Zhao, \Scalable QoS Guaranteed Communication Services for Real-Time Applications," Proceedings of ICDCS'00, 2000

....but infers it based on the information carried with the packet. We believe that to be scalable over a wide variety technologies, the core router should serve packets with class based information only, thus rely on flow aggregation. We have reported some preliminary work on this subject in [17]. We extend that work as follows: 1) We formally introduce the concept of safe utilization level of a server, and use configuration time test to determine this utilization level. 2) We discuss how appropriate route selection improves utilization levels, and design a safe route selection ....

....over a given link. In this section, we compare our heuristic search algorithm described in Section 5.2 with the Shortest Path based routing algorithm (SP) The measurement is the achieved maximum utilization assignment level of the compared algorithms. We selected the MCI ISP backbone network [17]. In this network, L, the diameter of the topology is 4, N , the maximum number of links for a router is 6. In the experiments, all routers can act as edge routers, and flows can be established between any two routers. All links in the simulated network have the same capacity of 100 Mbps. We use ....

B. Choi, D. Xuan, C. Li, R. Bettati and W. Zhao, "Scalable QoS Guaranteed Communication Services for Real-Time Applications," Proc. of ICDCS, 2000.

No context found.

B.-K. Choi, D. Xuan, R. Bettati, et al., "Scalable QoS guaranteed communication services for real-time applications," in Proc. of IEEE ICDCS'00, Apr. 2000, pp. 180--187.

No context found.

B. Choi, D. Xuan, C. Li, R. Bettati, and Wei Zhao, Scalable QoS Guaranteed Communication Services for Real-Time Applications, Proc. of the IEEE International Conf on Distributed Computing Systems, April 2000.

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