Advanced Computer NetworksImproving QOS in IP NetworksPrinciples for QOS GuaranteesPrinciples for QOS Guarantees (more)Slide 5Slide 6Summary of QoS PrinciplesScheduling And Policing MechanismsScheduling Policies: moreScheduling Policies: still moreSlide 11Policing MechanismsSlide 13Policing Mechanisms (more)IETF Integrated ServicesIntserv: QoS guarantee scenarioCall AdmissionIntserv QoS: Service models [rfc2211, rfc 2212]IETF Differentiated ServicesDiffserv ArchitectureSlide 21Classification and ConditioningForwarding - Per Hop Behavior (Core Router Functionality)Slide 24RSVP does not…Slide 26Slide 27Advanced Computer NetworksQuality of ServiceImproving QOS in IP NetworksThus far: “making the best of best effort”Future: next generation Internet with QoS guaranteesIntegrated Services: firm guaranteesRSVP: signaling for resource reservationsDifferentiated Services: differential guaranteessimple model for sharing and congestion studies:Principles for QOS GuaranteesExample: 1Mbps IP phone, FTP share 1.5 Mbps link. bursts of FTP can congest router, cause audio losswant to give priority to audio over FTPpacket marking needed for router to distinguish between different classes; and new router policy to treat packets accordinglyPrinciple 1Principles for QOS Guarantees (more)what if applications misbehave (audio sends higher than declared rate)policing: force source adherence to bandwidth allocationsmarking and policing at network edge:provide protection (isolation) for one class from othersPrinciple 2Principles for QOS Guarantees (more)Allocating fixed (non-sharable) bandwidth to flow: inefficient use of bandwidth if a flow doesn’t use its allocationWhile providing isolation, it is desirable to use resources as efficiently as possiblePrinciple 3Principles for QOS Guarantees (more)Basic fact of life: can not support traffic demands beyond link capacityCall Admission: flow declares its needs, network may block call (e.g., busy signal) if it cannot meet needsPrinciple 4Summary of QoS Principles Let’s next look at mechanisms for achieving this ….Packet ClassificationAssociate each packet with a reservation at the routersIsolation: Scheduling and policingManage queues so that packets receive the requested serviceHigh resource utilizationEfficient use of resources in the networkCall admissionDecide if new flow can be supported w/o sacrificing the QoS levels of the existing flowsScheduling And Policing MechanismsScheduling: choose next packet to send on linkFIFO (first in first out) scheduling: send in order of arrival to queuediscard policy: if packet arrives to full queue: who to discard?•Tail drop: drop arriving packet•Priority: drop/remove on priority basis•Random: drop/remove randomlyScheduling Policies: morePriority scheduling: transmit highest priority queued packet multiple classes, with different prioritiesclass may depend on marking or other header info, e.g. IP source/dest, port numbers, etc..Scheduling Policies: still moreRound robin scheduling:multiple classescyclically scan class queues, serving one from each class (if available)Scheduling Policies: still moreWeighted Fair Queuing: generalized Round Robineach class gets weighted amount of service in each cyclePolicing MechanismsGoal: limit traffic to not to exceed declared parametersThree common-used criteria: (Long term) Average Rate: how many pkts can be sent per unit time (in the long run)crucial question: what is the interval length: 100 packets per sec or 6000 packets per min have same average!Peak Rate: how many pkts can be sent over a short time, e.g. 100 pps peak rateBurst Size: max. number of pkts sent consecutively (with no intervening idle)Policing MechanismsToken Bucket: limit input to specified Burst Size and Average Rate. bucket can hold b tokenstokens generated at rate r token/sec unless bucket fullover interval of length t: number of packets admitted less than or equal to (r t + b).Policing Mechanisms (more)token bucket and WFQ combine to provide guaranteed upper bound on delay, i.e., QoS guarantee !WFQ token rate, rbucket size, bper-flowrate, RD = b/RmaxarrivingtrafficIETF Integrated Servicesarchitecture for providing QOS guarantees in IP networks for individual application sessionsresource reservation: routers maintain state info of allocated resources, QoS req’scall setup: admit/deny new call setup requestsQuestion: can newly arriving flow be admitted with performance guarantees while not violating QoS guarantees made to already admitted flows?Intserv: QoS guarantee scenarioResource reservationcall setup, signaling (RSVP)Traffic (T-spec), QoS (R-spec) declarationper-element admission controlQoS-sensitive scheduling (e.g., WFQ)request/replyCall AdmissionArriving session must :declare its QOS requirementR-spec: defines service requested from network•Controlled-load or guaranteed (delay target)characterize traffic it will send into network T-spec: defines traffic characteristics•Token bucket filter (avg bw + burstiness)signaling protocol: needed to carry R-spec and T-spec to routers (where reservation is required)RSVPIntserv QoS: Service models [rfc2211, rfc 2212]Guaranteed service:worst case traffic arrival: token bucket policed source simple (mathematically provable) bound on delay [Parekh 1992, Cruz 1988]Controlled load service:"a quality of service closely approximating the QoS that same flow would receive from an unloaded network element."WFQ token rate, rbucket size, bper-flowrate, RD = b/RmaxarrivingtrafficIETF Differentiated ServicesConcerns with Intserv:Scalability: signaling, maintaining per-flow router state difficult with large number of flows Flexible Service Models: Intserv has only two classes. Also want “qualitative” service classesrelative service distinction: Platinum, Gold, SilverDiffserv approach: simple functions in network core, relatively complex functions at edge routers (or hosts)Don’t define service classes, provide functional components to build service classesDiffserv ArchitectureEdge router:- per-flow traffic management- marks packets as in-profile and out-profile Core router:- per class traffic management- buffering and scheduling based on marking at edge- preference given to in-profile packetsscheduling...rbmarkingEdge-router Packet Marking class-based marking:
View Full Document