DMCA
A Dynamic Priority Assignment Technique for Streams with (m,k)-Firm Deadlines (1994)
Venue: | IEEE Transactions on Computers |
Citations: | 178 - 2 self |
Citations
256 | Algorithms for scheduling imprecise computations,”
- Liu, Yu, et al.
- 1991
(Show Context)
Citation Context ...s percentage requirement can be met even if a large number of consecutive customers miss their deadlines. Streams with (m; k)-firm deadlines can also be serviced using the imprecise computation model =-=[4,7,6]-=-. In this model, customers are statically classified as either mandatory or optional. Mandatory customers are always serviced. Optional customers are serviced on a best-effort basis. A stream with (m;... |
241 |
Characterizing superposition arrival processes in packet multiplexers for voice and data
- Sriram, Whitt
- 1986
(Show Context)
Citation Context ... durations of the ON and the OFF states are exponentially distributed with averages ON ave and OFF ave , respectively. Such a stream is often used to model a stream of voice samples in a conversation =-=[9,11]-=-. The ON state corresponds to a talkspurt and the OFF state corresponds to a silence period. We first consider the case where all streams in the system have the same timing requirements. We also assum... |
139 |
Scheduling periodic jobs that allow imprecise results
- Chung, Liu, et al.
- 1990
(Show Context)
Citation Context ...s percentage requirement can be met even if a large number of consecutive customers miss their deadlines. Streams with (m; k)-firm deadlines can also be serviced using the imprecise computation model =-=[4,7,6]-=-. In this model, customers are statically classified as either mandatory or optional. Mandatory customers are always serviced. Optional customers are serviced on a best-effort basis. A stream with (m;... |
112 | RealTime Communication in Packet-Switched Networks,”
- Aras, Kurose, et al.
- 1994
(Show Context)
Citation Context ...their deadlines in a window of k j consecutive customers is increased. To accomplish Streams Server Service Queue Figure 1: System Model this, several different scheduling policies have been proposed =-=[1]-=-. Most of these policies can be abstracted as follows. There is a separate First-In First-Out queue for each stream (see Fig. 1). These queues ensure that customers from a given stream are serviced in... |
52 |
Scheduling policies for real-time and non-real time trac in a statisticalmultiplexer.
- Chipalkatti, Kurose, et al.
- 1989
(Show Context)
Citation Context ...o other work in the literature because of the new model. Many schemes have been proposed to meet a specified maximum allowable loss percentage, where the loss includes customers with missed deadlines =-=[2,3,5,8,10]-=-. As indicated earlier, however, the problem with this model is that the loss percentage requirement can be met even if a large number of consecutive customers miss their deadlines. Streams with (m; k... |
47 |
Survey of traffic control schemes and protocols
- Bae, Suda
- 1991
(Show Context)
Citation Context ...o other work in the literature because of the new model. Many schemes have been proposed to meet a specified maximum allowable loss percentage, where the loss includes customers with missed deadlines =-=[2,3,5,8,10]-=-. As indicated earlier, however, the problem with this model is that the loss percentage requirement can be met even if a large number of consecutive customers miss their deadlines. Streams with (m; k... |
17 | Congestion control for real-time traffic in high-speed networks
- Schulzrinne, Kurose, et al.
- 1990
(Show Context)
Citation Context ...o other work in the literature because of the new model. Many schemes have been proposed to meet a specified maximum allowable loss percentage, where the loss includes customers with missed deadlines =-=[2,3,5,8,10]-=-. As indicated earlier, however, the problem with this model is that the loss percentage requirement can be met even if a large number of consecutive customers miss their deadlines. Streams with (m; k... |
15 |
Optimal packet discarding: An ATM-oriented analysis model and initial results.
- Petr, Frost
- 1990
(Show Context)
Citation Context ...o other work in the literature because of the new model. Many schemes have been proposed to meet a specified maximum allowable loss percentage, where the loss includes customers with missed deadlines =-=[2,3,5,8,10]-=-. As indicated earlier, however, the problem with this model is that the loss percentage requirement can be met even if a large number of consecutive customers miss their deadlines. Streams with (m; k... |
6 |
Imprecise computations: A means to provide scheduling flexibility and enhance dependability
- Liu, Lin, et al.
- 1993
(Show Context)
Citation Context ...s percentage requirement can be met even if a large number of consecutive customers miss their deadlines. Streams with (m; k)-firm deadlines can also be serviced using the imprecise computation model =-=[4,7,6]-=-. In this model, customers are statically classified as either mandatory or optional. Mandatory customers are always serviced. Optional customers are serviced on a best-effort basis. A stream with (m;... |
6 |
Simulation analysis of a communication link with statistically multiplexed bursty voice sources
- Saleh, Habib, et al.
- 1993
(Show Context)
Citation Context ... durations of the ON and the OFF states are exponentially distributed with averages ON ave and OFF ave , respectively. Such a stream is often used to model a stream of voice samples in a conversation =-=[9,11]-=-. The ON state corresponds to a talkspurt and the OFF state corresponds to a silence period. We first consider the case where all streams in the system have the same timing requirements. We also assum... |
4 |
Congestion control for bursty video traffic in ATM networks
- Kubota, Murata, et al.
- 1992
(Show Context)
Citation Context ...o other work in the literature because of the new model. Many schemes have been proposed to meet a specified maximum allowable loss percentage, where the loss includes customers with missed deadlines =-=[2,3,5,8,10]-=-. As indicated earlier, however, the problem with this model is that the loss percentage requirement can be met even if a large number of consecutive customers miss their deadlines. Streams with (m; k... |