Slide 1Types of ServicesSlide 3Slide 4Objective of paperMeasurement Based Admission ControlMeasurement Based Admission ControlWorst case delayComputation of new delay boundEquivalent Token bucket filterAdmission control algorithmSlide 12Measurement mechanism: Time windowMeasuring delayMeasuring ratePerformance TuningCOPSSlide 18A Measurement Based Admission Control Algorithm for Integrated Services Packet Networks ( extended version)S. Jamin, P. Danzig, S. Shenker, L. Zhang2Types of ServicesGuaranteed service a absolute bound on delay of every packet is providedPredictive service a fair but not absolute bound on delay of every packet is provided3Guaranteed serviceProvides hard or absolute bound on delay of every packetSources are described by either peak or average rates or filter like token bucketDescriptions provide upper bounds on traffic generated by sourceAdmission control algorithms use these a priori characterizations to calculate worst case behavior of flowsNetwork utilization low when flows are burstyNetwork utilization acceptable when flows are smooth4Predictive ServiceDoes not provide absolute bound on delayservice commitment is somewhat less reliablesources characterized by token bucket filters at admission timebehavior of existing flows determined by measurement rather than a priori characterizations when only few flows present, unpredictability of individual flows causes measurement based approach to be conservativedelivers significant gain in utilization only when there is high degree of statistical multiplexing5Objective of paperDescribes a measurement based admission control algorithm for predictive serviceTo answer the following questions can reliable delay bounds be provided for measurement based admission control algorithm does predictive service increase network utilization6Measurement Based Admission ControlSet of criteria controlling whether to admit a new flow based on approximate model of traffic flows uses measured quantities as inputs Measurement process to measure the inputsset of criteria used in this case delay bound as a result of admitting the new flow link utilization as a result of admitting the new flow7Measurement Based Admission ControlSources requesting services characterize worst case behaviorto guarantee hard bound on delay worst case behavior considered conservative estimate of bound obtained new flow may not be admitted due to conservative estimateresults in low utilizationto achieve fairly reliable bound for predictive service approximate maximal delay by replacing worst case parameters with measured quantities results in higher utilization8Worst case delayTheorem by Parekh says the delay bound for a class j is the one time delay burst that accrues if the aggregate bucket of all classes 1 through j flows are simultaneously dumped into the switch and all classes 1 through j-1 sources continue to send at their reserved rates with j having lowest prioritythis worst case delay can change when flow of same class is admitted when flow of higher priority class is admitted when guaranteed flow is admittedclasses classified according to the delay bounds tolerated by that class9Computation of new delay boundThe delay bounds as a result of admitting the new flow is computedfor e.g., When new predictive flow of same class is admitted new delay bound = old delay bound + k( bucket size of similarly new delay bounds are calculated for the caseWhen new predictive flow of lower class is admittedWhen new guaranteed flow is admitted10Equivalent Token bucket filterIn equations replace worst case values by measured valuessubstitute the reserved rates by the measured ratessubstitute the worst case delays by the measured delaysthis results in describing the existing aggregate traffic by an equivalent token bucket filter with parameters determined from traffic measurement11Admission control algorithmNew predictive flow: if incoming flow requests predictive service in a particular class request denied if sum of flow’s requested rate and current usage exceeds link utilization levelrequest denied if admitting new flow violates delay bound of the same priority levelrequest denied if admitting new flow violates delay bound of the lower priority level12Admission control algorithmNew guaranteed flow: if incoming flow requests guaranteed service request denied if total bandwidth of the flows including new flow exceeds link utilization levelrequest denied if admitting new guaranteed flow violates delay bound of predictive services this is due to the decrease in the bandwidth available for predictive flows due to admission of a guaranteed flow13Measurement mechanism: Time windowTwo measurementsexperienced delayutilizationto estimate delays measure the queuing delays d of every packetto estimate utilization sample the usage rate of the guaranteed service and the predictive class over sampling period of S packet transmission units14Measuring delayMeasurement variable to estimate queuing delay is DMeasurement window of T packet transmission units used as basic measurement blockvalue of D updated on three occasions at end of measurement block D updated to reflect maximal packet delay seen in previous blockwhen individual delay measurement exceeds estimated maximum queuing delay D is updated to times the sampled delaywhenever a new flow is admitted D is updated to the value of projected delay from equationλ15Measuring rateMeasurement variable to estimate queuing delay is VMeasurement window of T packet transmission units used as basic measurement blockvalue of V updated on three occasions at end of measurement block V updated to reflect maximal sampled utilization seen in previous blockwhen individual utilization measurement exceeds estimated V, V is updated to the sampled valuewhenever a new flow is admitted V is updated16Performance Tuning Parameters , S, T can be varied for better performancesmaller S, more sensitive to bursts , larger S smoother trafficsmaller T means more adaptability to changes in traffic load, larger T results in more stability larger T results in fewer delay violations and lower link utilization λ17COPS A simple client/server model for
View Full Document