Traffic Management for the Available Bit Rate (ABR) Service in ...
Traffic Management for the Available Bit Rate (ABR) Service in ... Traffic Management for the Available Bit Rate (ABR) Service in ...
In our simulations, a cycle consists of one request from the client and one response from the server. We use a small response burst size (16 cells), and vary the request burst size. Single-Client Con guration and Parameter Values The con guration we use is called the single-client con guration (Figure 7.9). It consists of a single client which communicates with the server, via a VC which traverses a bottleneck link. An in nite source is used in the background to ensure that the network is always loaded, and any sudden bursts of tra c manifest as queues. All the links run at 155 Mbps. Figure 7.9: Client-Server Con guration With In nite Source Background The response size is kept constant at 16 cells. The request size can be 16, 256 or 8192 for small, meduim or large bursts respectively. The inter-cycle time is chosen to be 1ms. All links are 500km long. The other source parameters are chosen to maximize ACR and disable the e ects of other source rules: ICR = 10 Mbps, TDF = 1/8, TCR = 10 cells/sec TRM = 100 ms, TBE = 512, CDF = 0 to disable SES Rule 6. The switch uses the ERICA algorithm to calculate rate feedback. The ERICA algorithm uses two key parameters: target utilization and averaging interval length. 247
The algorithm measures the load and number of active sources over successive averag- ing intervals and tries to achieve a link utilization equal to the target. The averaging intervals end either after the speci ed length or after a speci ed number of cells have been received, whichever happens rst. In the simulations reported here, the target utilization is set at 90%, and the averaging interval length defaults to 100 ABR input cells or 1 ms, represented as the tuple (1 ms, 100 cells). In the following sections, we pictorially describe the simulation results� a full set of graphs may be found in reference [60]. Small Bursts Small bursts are seen in LANE tra c. For example, the ethernet MTU, 1518 bytes is smaller than 32 (Nrm) cells. Since small bursts are smaller than Nrm cells, no RM cells are transmitted during certain bursts. As a result, no SES rules are triggered during these bursts. In other words, the entire burst is transmitted at one rate. However, when RM cells are nally transmitted, UILI is triggered which brings down the ACR to ICR. The source rate, S, is nearly zero due to the short burst time and long idle time. Hence, ICR + Sisapproximately equal to ICR. Figure 7.10: E ect of UILI on Small Bursts 248
- Page 223 and 224: counts the bursty source as active
- Page 225 and 226: 6.23 Summary of the ERICA and ERICA
- Page 227 and 228: (a) Transmitted Cell Rate (c) Link
- Page 229 and 230: (a) Transmitted Cell Rate (c) Link
- Page 231 and 232: (a) Transmitted Cell Rate (c) Link
- Page 233 and 234: (a) Transmitted Cell Rate (basic ER
- Page 235 and 236: (a) Transmitted Cell Rate (basic ER
- Page 237 and 238: (a) Transmitted Cell Rate (c) Link
- Page 239 and 240: (a) Transmitted Cell Rate (c) Link
- Page 241 and 242: (a) Transmitted Cell Rate (c) Link
- Page 243 and 244: (a) Transmitted Cell Rate (c) Link
- Page 245 and 246: (a) Transmitted Cell Rate (c) Link
- Page 247 and 248: (a) Transmitted Cell Rate (c) Link
- Page 249 and 250: (a) Transmitted Cell Rate (basic ER
- Page 251 and 252: RM cells. In this chapter, we devot
- Page 253 and 254: low. This tradeo was discovered and
- Page 255 and 256: in the speci cation. b) ACR shall n
- Page 257 and 258: 7.1.6 December 1995 Proposals There
- Page 259 and 260: Figure 7.2: Multiplicative vsAdditi
- Page 261 and 262: is never triggered. However, the PR
- Page 263 and 264: simulation results of bursty source
- Page 265 and 266: 1. The time-based proposal also ind
- Page 267 and 268: The switch maintains a local alloca
- Page 269 and 270: PNI = f0, 1g : f1 ) No rule 5b, 0 )
- Page 271 and 272: after reaching the goal. The time-b
- Page 273: Figure 7.8: Closed-Loop Bursty Tra
- Page 277 and 278: Medium Bursts Medium bursts are exp
- Page 279 and 280: count-based technique may be insu c
- Page 281 and 282: Figure 7.13: An event trace illustr
- Page 283 and 284: CHAPTER 8 SUPPORTING INTERNET APPLI
- Page 285 and 286: u ering which does not depend upon
- Page 287 and 288: 4 make it to the destination are ar
- Page 289 and 290: Figure 8.3: At the ATM layer, the T
- Page 291 and 292: issues and e ect of bursty applicat
- Page 293 and 294: from the loss and they trigger the
- Page 295 and 296: 8.8 Performance Metrics We measure
- Page 297 and 298: the performance is fair. Also, the
- Page 299 and 300: Figure 8.7(b) shows the rates (ACRs
- Page 301 and 302: Window Size in bytes vfive-tcp/opti
- Page 303 and 304: The e ect of large bu ers on CLR is
- Page 305 and 306: 8.13 Summary of TCP/IP performance
- Page 307 and 308: Feedback delay: Twice the delay fro
- Page 309 and 310: on a link, two cells are expected a
- Page 311 and 312: state only after the switch algorit
- Page 313 and 314: queue length is less susceptible to
- Page 315 and 316: to equalize rates for fairness, and
- Page 317 and 318: QB = Link bandwidth (RT T ; T ) and
- Page 319 and 320: u ered at the end-system, and not i
- Page 321 and 322: Part b): When ABR load goes away, t
- Page 323 and 324: All our simulations presented use t
The algorithm measures <strong>the</strong> load and number of active sources over successive averag-<br />
<strong>in</strong>g <strong>in</strong>tervals and tries to achieve a l<strong>in</strong>k utilization equal to <strong>the</strong> target. The averag<strong>in</strong>g<br />
<strong>in</strong>tervals end ei<strong>the</strong>r after <strong>the</strong> speci ed length or after a speci ed number of cells have<br />
been received, whichever happens rst. In <strong>the</strong> simulations reported here, <strong>the</strong> target<br />
utilization is set at 90%, and <strong>the</strong> averag<strong>in</strong>g <strong>in</strong>terval length defaults to 100 <strong>ABR</strong> <strong>in</strong>put<br />
cells or 1 ms, represented as <strong>the</strong> tuple (1 ms, 100 cells).<br />
In <strong>the</strong> follow<strong>in</strong>g sections, we pictorially describe <strong>the</strong> simulation results� a full set<br />
of graphs may be found <strong>in</strong> reference [60].<br />
Small Bursts<br />
Small bursts are seen <strong>in</strong> LANE tra c. For example, <strong>the</strong> e<strong>the</strong>rnet MTU, 1518<br />
bytes is smaller than 32 (Nrm) cells. S<strong>in</strong>ce small bursts are smaller than Nrm cells,<br />
no RM cells are transmitted dur<strong>in</strong>g certa<strong>in</strong> bursts. As a result, no SES rules are<br />
triggered dur<strong>in</strong>g <strong>the</strong>se bursts. In o<strong>the</strong>r words, <strong>the</strong> entire burst is transmitted at one<br />
rate. However, when RM cells are nally transmitted, UILI is triggered which br<strong>in</strong>gs<br />
down <strong>the</strong> ACR to ICR. The source rate, S, is nearly zero due to <strong>the</strong> short burst time<br />
and long idle time. Hence, ICR + Sisapproximately equal to ICR.<br />
Figure 7.10: E ect of UILI on Small Bursts<br />
248