Advanced Traffic Management (QOS) Concepts 319

Copyright © 1998, Cisco Systems, Inc. All rights reserved. Printed in USA.Presentation_ID.scr 1

13191056_05F9_c2 © 1999, Cisco Systems, Inc. 1© 1999, Cisco Systems, Inc. 3191056_05F9_c2

2© 1999, Cisco Systems, Inc. 3191056_05F9_c2

Advanced TrafficAdvanced TrafficManagement and QoSManagement and QoS

ConceptsConceptsSession 319Session 319


33191056_05F9_c2 © 1999, Cisco Systems, Inc.

IntroductionIntroduction

• Traffic Management

• Applications and Transports

• So what Are the Issues forTCP

Voice on IP

Video (Broadcast and Teleconferencing)


Let's Talk about TrafficLet's Talk about TrafficManagementManagement

• Why it is a concern

• What the guiding principles are

• What tools are available

• What can be accomplished usingthose tools

• What cannot be accomplished



Why Traffic ManagementWhy Traffic ManagementIs a ConcernIs a Concern

• Needs of certain applicationsMail? Web? Transaction processing?

• Opportunities with certain transports


Guiding Principles forGuiding Principles forTraffic ManagementTraffic Management

• We want to achievePredictability

Reliability

Availability

• In a network thatKeeps intelligenceat the edges

Scales to necessarysizes andbandwidths

Minimizescomplexity

Uses cost-effectivetechnologies



What Tools Are Available forWhat Tools Are Available forTraffic ManagementTraffic Management

• Traffic path control

• Queue depth management

• Queue rate management

• Permission to use a link


How Well Will TrafficHow Well Will TrafficManagement Do?Management Do?

• We know we can do this:Management of latency

Management of bandwidth

• What cannot be accomplishedCreation of bandwidth that otherwisewould not be there



Source: Gartner Group Study, March 1997

WAN ProtocolBreakdown

WAN ProtocolBreakdown

1994 1996 1998E 2000E 2002E

80%70%60%50%40%30%20%10%0%

IP SNA IPXOthers RFC 1490

IP

Primarily a WAN IP TalkPrimarily a WAN IP Talk

• IP is the dominantinternet protocol

• TCP is the dominantdata transport

95% of Internet trafficuses TCP

• Voice is a growingmarket

But beware of hype

• Heterogeneouslink layers


Making NetworksMaking NetworksPredictablePredictable

The GrailThe Grail




This Is what You NeedThis Is what You Needto Understand:to Understand:

• TCP-based applications, voice,and video can be managed wellwith a little planning


Parekh and Gallagher’s PaperParekh and Gallagher’s Paper

• INFOCOMM ’93

• One must have at most a predictablepredictableamount of traffic in the network

• One must have predictablepredictable trafficdelay in each network element

• Given these, end-to-end delayend-to-end delay of ahost to host message is predictableis predictable



Definition of “Predictable”Definition of “Predictable”

• Does not mean“Fixed”, “Invariant”, or “Zero”

• Means that it has aMean value

Statistical distribution

Upper bound


Predictable Amount of TrafficPredictable Amount of Trafficin the Networkin the Network

• The source must pace trafficinitiation so that standing queuesare bounded

Queues form when arrival rate exceedsdeparture rate

• When congestion (too manymessages in one queue) sets in:

Sources must not increase their rateIdeally, sources decrease their rate



Examples of Source PredictabilityExamples of Source Predictability

• TCP will keep at most a certainamount of traffic in flight

We say it is “elastic”—rate isproportional to latency

• Voice will send only and exactly asfast as the coding algorithm permits

We say it is “inelastic”


Predictable Packet Treatment inPredictable Packet Treatment inRouters and SwitchesRouters and Switches

• Transit latency must be within limitsacceptable to the application

• Variation in transit latency must bewithin limits acceptable to theapplication

• No stream may be locked out apartfrom administrative policy

• Applicable policy must be observed



Examples of UnpredictabilityExamples of Unpredictability

• Queues change rapidly enough thatthe distribution cannot be described

• Discards happen frequently enoughthat there is effectively no upperbound on delivery time


Examples of PredictabilityExamples of Predictability

• Classes of queues get sufficientservice that ultimate arrival is timelyand normal

“Timely” is an application concept…



Quality of Service Issues inQuality of Service Issues inTraffic ManagementTraffic Management

• Predominantly TCP traffic

• Some specific applications

• Voice/video traffic


Managing TCP TrafficManaging TCP Traffic

Moving Mountains of DataMoving Mountains of DataWithout Incurring the WorldWithout Incurring the World

Wide WaitWide Wait




Transport Breakout TCP Applications

Backbone Traffic MixBackbone Traffic Mix

Source: MCI/NSF OC-3MON via http://www.nlanr.net, 1998


TCP Technology IssuesTCP Technology Issues

• Single drops communicate fromnetwork to sending host

“You need to slow down”

• Multiple drops in round trip triggertime-outs

“Something bad happened out here”



N+1

N

N+2N+3

Behavior of a TCP SenderBehavior of a TCP Sender

• Sends as much ascredit allows

• Starts credit smallAvoid overloadingnetwork queues

• Increases creditexponentially

To gauge network capability


Ack N+1

N+1

N

N+2N+3

Ack N+1

Ack N+1

Behavior of a TCP ReceiverBehavior of a TCP Receiver

• When in receipt of “nextmessage,” schedulesan ACK

• When in receipt ofsomething else,acknowledges all itcan immediately



Ack N+1

N+1

N

N+2N+3

Ack N+1

Ack N+1

Ack N+4

N+1

Sender Response to ACKSender Response to ACK

• If ACK acknowledgessomething

Update credit and send• If not, presume it indicates

a lost packetSend first unacknowledgedmessage right awayHalve current creditIncrease linearly to gaugenetwork throughput


Ack N+5

N+4

Ack N+1

N+1

N

N+2N+3

Ack N+1

Ack N+1

Ack N+4

N+1

N+4

WorldWorldWideWideWait!Wait!

Multiple Drops in TCPMultiple Drops in TCP

• In the event of multiple dropswithin the same session:

Current TCPs wait for time-out

Selective acknowledge maywork around (but seeINFOCOM ’98)

New Reno “fast retransmitphase” takes several RTTsto recover



Remember Parekh and GallagherRemember Parekh and Gallagher





How Can We Make TCP in aHow Can We Make TCP in aNetwork Act Predictably?Network Act Predictably?

• Predictable amount of trafficin the network:

Well-written TCP implementationsmanage their rates to the availablebandwidth

• Router needs toProvide predictable treatment of packets

Queue delay and drop characteristics



Fundamental FIFO QueueFundamental FIFO QueueManagement TechnologiesManagement Technologies

• Tail dropNetwork standard behavior

Causes session synchronization whenwaves of traffic experiencecorrelated drops

• Random Early Detection (RED)Random drops used to desynchronizeTCP sessions and control rates


Session SynchronizationSession Synchronization

• Sessionsynchronizationresults fromsynchronizedlosses

• Tail drop fromwaves of trafficsynchronizeslosses



RED EnabledRED EnabledCourtesy of Sean Doran, Ebone

Effect ofEffect ofRandom Early DetectionRandom Early Detection

• One day, below 100% throughputSimple FIFO with tail drop

• Starting 10:00 second day, 100%throughput

Random Early Detection enabled


RED EnabledRED EnabledCourtesy of Sean Doran, Ebone

Was that a Fluke?Was that a Fluke?

• No, here’s what happened that week…

• Session synchronization reducedthroughput until RED enabled



0

50

100

150

200

250

300

350

400

Elapsed Time

Ns

RT

T

Mean RTT Min RTT Max RTT STD DEV

MeanMeanLatencyLatencyCorrelatesCorrelateswithwithMaximumMaximumQueueQueueDepthDepth

FIFO Traffic TimingsFIFO Traffic Timings


0

50

100

150

200

250

300

350

400

Elapsed Time

Ms

RTT

Mean RTT Min RTT Max RTT STD DEV

MeanMeanLatencyLatencyCorrelatesCorrelateswithwithMinimumMinimumDropDropThresholdThreshold

AdditionalAdditionalCapacityCapacityto Absorbto AbsorbBurstsBursts

RED Traffic TimingsRED Traffic Timings



Therefore—TCP QoS Definition:Therefore—TCP QoS Definition:

• Normally at most one drop perround trip

• Mean variation in latency boundedby predictable network


TCP Flow StatisticsTCP Flow Statistics

• >90% of sessions have ten packetseach way or less

Transaction mode (mail, small web page)

• >80% of all TCP traffic results from<10% of the sessions, in highrate bursts

It is these that we worry about managing



An Interesting CommonAn Interesting CommonFallacy about RED:Fallacy about RED:

• “RED means you will have more drops”Statement derives from observed statistics

• RED means that you will haveCloser to 100% utilization of your line

Less average delay per packet

• But queuing theory?As a line approaches 100% utilization, drops willincrease, even though served load increaseseven though served load increases


TCP Traffic Management IssuesTCP Traffic Management Issues

• ApplicationsOften have site-specific policyassociated with them

Traffic often identifiable by port numbers

• SitesGenerally identifiable by address prefixor interface traffic is received on



TCP Bandwidth PolicyTCP Bandwidth PolicyQuestions to AnswerQuestions to Answer

• Particular site or application wants atleastleast a certain bandwidth

• Particular site or application wants atmostmost a certain bandwidth

• Particular site or application wants toaverageaverage about a certain bandwidth


ClassifierClassifier

Queues

InterfaceInterface

This Is Where “Classes”This Is Where “Classes”Come inCome in

• Classes can be for:Voice

Importantapplication/site

Unimportantapplication/site

Assuring at least a rate

Limiting to a rate



Managed Link

Left RightICU

UC Me

U Betcha

Some Class of Traffic Wants atSome Class of Traffic Wants atLeastLeast a Certain Bandwidth a Certain Bandwidth

• Example:Several organizations share cost of link

Distribute bandwidth proportional tofiscal responsibility


Class of Traffic Wants at Class of Traffic Wants at MostMosta Certain Bandwidtha Certain Bandwidth

• Traffic shaping

• Similar queuing technology to class-based weighted fair queuing

• Rate assigned toInterface or sub-interface

Frame Relay circuit

ATM virtual channel (in hardware)



T-1

64KBPS

Examples of Rate ControlExamples of Rate Control

• Intranet exposureLimit rate of web surfingoutside the company

• Frame Relay networkAccess rate exceedsPVC rate—limit rate torate of PVC


Some Class of Traffic Wants toSome Class of Traffic Wants toAverageAverage a Certain Bandwidth a Certain Bandwidth

• Service provider or largeenterprise model

• Designed forCost containment

Managed response toconflicting demands



Marking TCP Traffic at EdgeMarking TCP Traffic at Edge

• A useful technique:

• Mark traffic at a network edge withsimple classifier

• This allows network toDo the right thing without having to fullyclassify everywhere

Use more effective markings


Serving TCP Traffic with theServing TCP Traffic with theAssured ServiceAssured Service

• Presumes service level agreementFlat rate for traffic meeting a rate/burst profile

Usage charging for traffic out of profile

• Drop management (weighted RED)All traffic subject to loss

Traffic out of profile much more subject to loss

Enhances ISP traffic engineering

(Good for service provider and consumer)



Assured Service

0%

10%

20%

30%

40%

50%

60%

70%

0 10 20 30 40 50 60 70 80 90

TimeU

sage

Structure of Presumed ServiceStructure of Presumed ServiceLevel AgreementLevel Agreement

• Up to raterate over intervalintervalis “in profile”

• Traffic within profilegets some guarantees

• Traffic out of profilehas no guaranteesPotentially dropped byWRED at bottleneckUsage pricingUsage pricing of overage


Line Congested?Drop at Some Rate!

Best Effort Service in SimpleBest Effort Service in SimpleIP NetworksIP Networks



Line Congestedand Packet Out ofand Packet Out ofProfileProfile? Drop at

Some Rate!

Assured Service in SimpleAssured Service in SimpleIP NetworksIP Networks


Best effort Service in anBest effort Service in anATM-Based NetworkATM-Based Network



Assured Service in anAssured Service in anATM-Based NetworkATM-Based Network


So, for TCPSo, for TCP

• Traffic can be contained to a rate in amanner consistent with good qualityof service

• Traffic can be managed well with alittle foresight and planning



Convergence with VoiceConvergence with VoiceNetworksNetworks

“It’s about Internet Telephony!”“It’s about Internet Telephony!”



Again, the Premise:Again, the Premise:

• TCP-based applications,voicevoice, and video can bemanaged well with a littleplanning



0

20

40

60

80

100

1996 1998 2000

MultimediaDynamic WWWStatic WWWFTP and TelnetEmail and NewsOther

Changing Corporate NetworkChanging Corporate NetworkApplication PredominanceApplication Predominance

2%2%7%7%

27%27%

17%17%

39%39%

8%8%

2%2%

7%7%

27%27%

17%17%

39%39%

8%8%

13%13%

28%28%

15%15%

17%17%

12%12%

14%14%

Numbersin Percent

Source: The Yankee Group, 1996


• Email

• Informationsearch/access

• Subscriptionservices/“Push”

• Conferencing/multimedia

• Video/imaging

250

200

150

100

50

1997 1998 1999 2000 2001

Traffic Projectionsfor Voice and Data

Rel. BitVolume

Circuit Switched Voice

Data(IP)

“From 2000 on, 80% of Service“From 2000 on, 80% of ServiceProvider Profits Will Be DerivedProvider Profits Will Be Derivedfrom IP-Based Services.”from IP-Based Services.”Source: CIMI Corp.Source: CIMI Corp.

Growth of IP TrafficGrowth of IP Traffic

Source: Multiple IXC Projections



IP orIP orVoiceVoice

ATMATM

OpticalOptical

B-ISDNB-ISDN IP or VoiceIP or Voiceover ATMover ATM

IP overIP overSONET/SDHSONET/SDH

IP overIP overOpticalOptical

IPIP

SONET/SDHSONET/SDH

OpticalOptical

ATMATM

SONET/SDHSONET/SDH

IP orIP orVoiceVoice

OpticalOptical

VoiceVoice

IPIP

OpticalOptical

High End IP Transport AlternativesHigh End IP Transport Alternatives

Lower Cost, Complexity and OverheadLower Cost, Complexity and Overhead

Multiplexing, Protection and Management at Every LayerMultiplexing, Protection and Management at Every Layer


H.323 Voice/VideoH.323 Voice/Video

• VoiceConstant bit rate when sending

Relatively small messages (44-170 bytes)

• VideoGenerally high variable bit rate

Controlled by codec efficiency on picture

Message size is generally the MTU



Delta Frames

KeyFrame

KeyFrame

Video: Traffic PatternVideo: Traffic Pattern


Video: Effect of DelayVideo: Effect of Delay

Delta Frames

KeyFrame

KeyFrame



Distribution of Deliveries in Time

Tran

smis

sion

Tim

ePlaybackPoint

Unless it’sToo Late…

TypicalDelivery

Application Buffers Datato Ensure Consistency

Preferred DeliveryPreferred DeliveryIntervalInterval

Video: Playback PointVideo: Playback Point


Synchronization ofSynchronization ofVoice and VideoVoice and Video

• McGurk effect: voice can soundgarbled to human ear when notsynchronized with video

• Therefore, we have tosynchronize these



QoS Definition for Voice:QoS Definition for Voice:

• Low loss rate

• Low absolute delay in two-waysituations

Broadcast voice doesn’t havethis problem…

• Low variation in delay


Key Issue for Voice QoS:Key Issue for Voice QoS:

• Silent periods must not be randomlyinserted or removed so as to makeother sounds unintelligible

• End to end delay must becomprehended by human listener



QoS Definition for Video:QoS Definition for Video:

• Low loss rate

• Low absolute delay in two-waysituations

• Low variation in delay


Key Issues for Video:Key Issues for Video:

• All packets that comprise a videoframe must arrive during the sameframe interval

OK if it’s the last millisecond of thatinterval…

• Audio and video must besynchronized when shown to user



How Can We Make Internet VoiceHow Can We Make Internet VoiceAct Predictably?Act Predictably?

• Predictable amount of traffic inthe network

• Predictable treatment of packets inrouters and switches

• Planning to support these aspectsresults in a predictable network


Voice/Video TrafficVoice/Video TrafficManagement IssuesManagement Issues

• The fundamental problems withVoice/video traffic are

It doesn’t slow down in response todelay or loss

It requires minimal variation in delay



Predictable Amount ofPredictable Amount ofTraffic in the NetworkTraffic in the Network

• The implication is that we have tocontrol used capacity

Capacity that individual calls consume“If you experience poor quality, use amore compact encoding or a lowerframe rate”

Capacity that total call volume canconsume

“If there isn’t capacity, refusenew calls”


Predictable Treatment of PacketsPredictable Treatment of Packetsin Routers and Switchesin Routers and Switches

• We have to place voice in queuesthat give it high priority

Maintain tight delay budgets

Application of class-based WFQ



Planning for a Predictable NetworkPlanning for a Predictable Network

• Enable CB—WFQ on all relevant linksConfigure voice queue with more bandwidththan traffic will need, or

For low bandwidth, priority queue [12.0(6)T]

• Low speed links should useLink Fragmentation or FRF.12

RTP compression for voice

• Enable RSVP call negotiation“Refuse excess calls”


FRF.12, and Link FragmentationFRF.12, and Link Fragmentationand Interleavingand Interleaving

• Premise:Reducing voice packet size reducessession requirements on network

So compress out IP, UDP, and RTPheaders as much as possible

• Limits jitter on lower bandwidth links



Resource ReservationResource Reservation

• Current deployment

• Current extensions

• Extensions being developed


Current DeploymentCurrent Deployment

• RSVP version 1Call control for individual sessions

Deployed

Cisco 11.2

Microsoft Windows ’98 (service pack)

Microsoft Windows NT 2000

• Appropriate to edge networks



Current ExtensionsCurrent Extensions

• Policy management via COPS

• LAN management via subnetbandwidth manager


Policy Management Via COPSPolicy Management Via COPS

• Local or centralpolicy server canauthorize decisions

• Local policy:Simple policies

• Central policyserver:

Certificates,

Complex policies



LAN Management Via SubnetLAN Management Via SubnetBandwidth ManagerBandwidth Manager

• Subnet bandwidthmanager is RSVPin a switch

• Controls aggregatereservations ona LAN


Extensions Being DevelopedExtensions Being Developed

• Rapid deployment of calls

• Aggregate classification inedge networks

• Aggregate classification andadmission in service providernetworks



Rapid Deployment of CallsRapid Deployment of Calls

• Problem: need acknowledgedreservation installation

• Solution: acknowledge it…


PSTN

PSTN

Aggregate Classification inAggregate Classification inEdge NetworksEdge Networks

• Use differentiatedservices codepoints to identifytraffic

Rather thanspecific flows



Aggregate Classification inAggregate Classification inEdge NetworksEdge Networks

• Reservation requestedby host in the usualway (RFC 2205)

• Flow classificationand policing at firsthop router

• Flow admission alongend to end path

• Aggregate classificationand policing atsubsequent routers


• Voice/video callsPlaced across aggregationdomain boundary

Aggregate Classification Aggregate Classification and Admissionand AdmissionAcross Service Provider NetworksAcross Service Provider Networks

• Why?Otherwise, you don’tknow that bandwidthexists on a path

• Aggregate reservationsPlaced from ingressto egress for DSCP used

Use expeditedforwarding service

Limited rate of change



Solving Voice/Video Issues UsingSolving Voice/Video Issues Usingthe Expedited Forwarding Servicethe Expedited Forwarding Service

• Rate controlApplication at source

Reservation in network

• Jitter controlWFQ’s priority queue (low speed)

Statistically empty queue (CB-WFQ)


The Implications forThe Implications forVoice and VideoVoice and Video

• We can control call volumeAnd therefore traffic volume

• We can scalably prioritize traffic inthe system

And therefore deliver on latency issues

• So, voice and video can be managedwell with a little planning



Traffic Path ControlTraffic Path Control

What if IP Routing Isn’t QuiteWhat if IP Routing Isn’t QuiteGood Enough for Your Traffic?Good Enough for Your Traffic?



Traffic EngineeringTraffic Engineering

• Historical approachesLoad sharing

Routing metrics

• A new oneLabel switching



Load SharingLoad Sharing

• Multipath routingEqual and unequal cost

• Multilink PPP


RoutingRouting

• Administrative metricsDesigned to move traffic to statisticallylow volume links

• Load sensitive metricsDesigned to move data away fromcongested links

Tendency towards oscillation



Utility of These:Utility of These:

• While they basically work, they areNot deterministic, and

Tend to be hard to predict


MMulti-ulti-PProtocol rotocol LLabel abel SSwitchingwitching

• MPLS traffic engineeringVPNs and general engineering

• MPLS routing for resourcereservation

In the direction of QoS routing



Principles of Label SwitchingPrinciples of Label Switching

• Labeled paths:Multiple enumerated point to pointrelationships between pairs of routersSets of pair-wise relationships create alabeled tunnel

• Conceptually similar to ATM VCs orFrame Relay DLCs, but

Interface independentUsed to model network layer constructsVariable length packets


Notice: TwoLabels on One

Interface,Distinguishing

Routes

Network Layer Constructs…Network Layer Constructs…

• Types of traffic streamsDestination routes

Source-destination routes

AS pairs

BGP community pairs

• Tunnels can createAny routing that meetsengineering needs



Using Labeled Tunnels to CreateUsing Labeled Tunnels to CreateVirtual Private NetworksVirtual Private Networks

• Imagine edgenetwork with privateaddress space

• Stretch labeledtunnels across thenetwork

• Now, do it again• Disjoint networks

Same address spaceSeparate routing


MPLS Traffic EngineeringMPLS Traffic Engineering

• Same technology can dragspecific routes around

Several less-used paths vs afew denser paths…

• Initially seen asoff-line engineering

• Can use eitherLDP or RSVP to install routes



CoS in MPLS NetworksCoS in MPLS Networks

•• CClass oof SServiceRoughly similar to diff-serv code point

Eight values, not sixty-four

• Implements similar drop/delaymanagement within labeled tunnels

• Therefore, MPLS networks havefundamental TCP QoS support


The Obvious Hole…The Obvious Hole…

• Wouldn’t it be nice if engineeredlabeled tunnels could

Have specific bandwidths guaranteed?

Recover from network events quicklyand automatically using reasonable ifnot optimal routes?



MPLS Routing for ResourceMPLS Routing for ResourceReservationReservation

• Use OSPF/IS-IS to distributebandwidth availability information

• Edge router does SPF calculationwhen needed

• RSVP used to install labeled tunnelwhile checking for race events

• CoS field used to identify traffic forqueued rate support


Automated Reinstallation ofAutomated Reinstallation ofLabeled TunnelsLabeled Tunnels

• RSVP tears downaffected tunnels

• Edge devicesrecalculate routes

• RSVP used tore-install tunnels

• Bandwidth checksresult in retry



Q.E.D. MPLSQ.E.D. MPLS

• Traffic engineering for networklayer traffic can be managed wellwith a little planning


So, What Are YouSo, What Are Youto Do about It?to Do about It?

Here the Rubber Meets the RoadHere the Rubber Meets the Road




Plan Your Network forPlan Your Network forPredictabilityPredictability

• Network engineering

• Assured forwarding serviceTCP

• Expedited forwarding serviceVoice, implies some form of admission


Network EngineeringNetwork Engineering

• Capacity engineeringEngineered IP routes?

• May involve traffic engineeringLabeled tunnels?



Assured Forwarding ServiceAssured Forwarding Service

• Designed for TCPClasses control rates for SLAs

Drop controls trace effects backto sources

• Implement usingCommitted access rate,

Weighted Random Early Detection,

Class-based weighted fair queuing


Expedited Forwarding ServiceExpedited Forwarding Service

• Appropriate to voice/video

• RequiresUnder-subscribed traffic classes

Reservation of bandwidth

Policing



Guiding PrinciplesGuiding Principlesfor Predictabilityfor Predictability





In Your Network…In Your Network…

• TCP-based applications, voice,and video—and yourbandwidth—can be managedwell with a little planning



Please Complete YourPlease Complete YourEvaluation FormEvaluation Form

Session 319Session 319



Date post:	08-Dec-2016
Category:	Documents
Upload:	doantu
View:	217 times
Download:	2 times

Advanced Traffic Management (QOS) Concepts 319

Documents