Data and Computer Data and Computer CommunicationsCommunications

Eighth EditionEighth Edition

by William Stallingsby William Stallings

Lecture slides by Lawrie BrownLecture slides by Lawrie Brown

Chapter 19 – Internetwork OperationChapter 19 – Internetwork Operation

Internetwork OperationInternetwork Operation

She occupied herself with studying a map on the opposite wall because she knew she would have to change trains at some point. Tottenham Court Road must be that point, an interchange from the black line to the red. This train would take her there, was bearing her there rapidly now, and at the station she would follow the signs, for signs there must be, to the Central Line going westward —King Solomon's Carpet, Barbara Vine (Ruth Rendell)

Internetwork OperationInternetwork Operation

consider mechanisms for handling growth in consider mechanisms for handling growth in network trafficnetwork traffic from low-volume text based terminal/emailfrom low-volume text based terminal/email to high volume multi-media web/voice/videoto high volume multi-media web/voice/video

historically IP nets gave best-effort datagram historically IP nets gave best-effort datagram delivery to all servicesdelivery to all services

now want variety of QoS in IP networksnow want variety of QoS in IP networks We will explore some new network services / We will explore some new network services /

functionsfunctions

MulticastingMulticasting

sending packet to addresses referring to sending packet to addresses referring to group of hosts on one or more networksgroup of hosts on one or more networks multimedia “broadcast”multimedia “broadcast” teleconferencingteleconferencing databasedatabase distributed computingdistributed computing real time workgroupsreal time workgroups

have design issues in addressing / routinghave design issues in addressing / routing

LAN MulticastLAN Multicast

LAN multicast is easyLAN multicast is easy send to IEEE 802 multicast MAC addresssend to IEEE 802 multicast MAC address since broadcast all stations will see packetsince broadcast all stations will see packet those in multicast group will accept itthose in multicast group will accept it only single copy of packet is neededonly single copy of packet is needed

but much harder in internetworkbut much harder in internetwork

Example Example ConfigConfig

Broadcast / Multiple Unicast / Broadcast / Multiple Unicast / MulticastMulticast

could broadcast packet to each networkcould broadcast packet to each network if server does not know members of groupif server does not know members of group requires 13 packetsrequires 13 packets

could send multiple unicast packetscould send multiple unicast packets to each net with members in multicast groupto each net with members in multicast group requires 11 packetsrequires 11 packets

or use true multicastor use true multicast which send single packets over any linkwhich send single packets over any link duplicating as needed to reach dest netsduplicating as needed to reach dest nets requires 8 packetsrequires 8 packets

True MulticastTrue Multicast

determine least cost path to each network determine least cost path to each network that has host in groupthat has host in group results in a spanning treeresults in a spanning tree of just those nets with members in groupof just those nets with members in group

transmit single packet along spanning treetransmit single packet along spanning tree routers replicate packets at branch points routers replicate packets at branch points

of spanning treeof spanning tree

Multicast ExampleMulticast Example

Requirements for Requirements for MulticastingMulticasting

router may have to forward more than one copy of router may have to forward more than one copy of packetpacket

need convention to identify multicast addresses (IPv4 need convention to identify multicast addresses (IPv4 Class D or IPv6 prefix)Class D or IPv6 prefix)

IPv4 : Class D - start 1110IPv4 : Class D - start 1110 IPv6 : 8-bit prefix, all 1, 4-bit flags field, 4-bit scope field,IPv6 : 8-bit prefix, all 1, 4-bit flags field, 4-bit scope field,

112-bit group identifier112-bit group identifier nodes translate between IP multicast addresses and nodes translate between IP multicast addresses and

list of networks containing group memberslist of networks containing group members router must translate between IP multicast address router must translate between IP multicast address

and network multicast addressand network multicast address Ethernet multicast addresses…Ethernet multicast addresses…

http://www.hep.ucl.ac.uk/~ytl/multi-cast/addresstranslation_01.html

Requirements for Requirements for MulticastingMulticasting

mechanism required for hosts to join and leave mechanism required for hosts to join and leave multicast groupmulticast group

routers must exchange inforouters must exchange info which networks include members of given groupwhich networks include members of given group sufficient info to work out shortest path to each sufficient info to work out shortest path to each

networknetwork

routing algorithm to work out shortest pathrouting algorithm to work out shortest path routers must determine routing paths based on routers must determine routing paths based on

source and destination addressessource and destination addresses

Spanning Tree from Router C Spanning Tree from Router C to Multicast Groupto Multicast Group

Internet Group Management Internet Group Management Protocol Protocol ((IGMPIGMP))

RFC RFC 33763376 to exchange multicast group info to exchange multicast group info between hosts & routers on a LANbetween hosts & routers on a LAN

hosts send messages to routers to subscribe hosts send messages to routers to subscribe to to and unsubscribe from multicast groupand unsubscribe from multicast group

routers check which multicast groups of interest routers check which multicast groups of interest to which hoststo which hosts

IGMP currently version 3IGMP currently version 3

Operation of IGMPv1 & v2Operation of IGMPv1 & v2 IGMPv1IGMPv1

hosts could join grouphosts could join group routers used timer to unsubscribe membersrouters used timer to unsubscribe members

IGMPv2IGMPv2 enabled hosts to unsubscribe enabled hosts to unsubscribe operational model:operational model:

receivers have to subscribe to groupsreceivers have to subscribe to groups sources do not have to subscribe sources do not have to subscribe to to groupsgroups any host can send traffic to any multicast groupany host can send traffic to any multicast group

problems:problems: spamming of multicast groupsspamming of multicast groups establishment of distribution trees is problematicestablishment of distribution trees is problematic finding globally unique multicast addresses difficultfinding globally unique multicast addresses difficult

IGMP v3IGMP v3

addresses weaknesses:addresses weaknesses: allowsallows hosts to specify list from which they hosts to specify list from which they

want to receive trafficwant to receive traffic traffic from other hosts blocked at routerstraffic from other hosts blocked at routers allowsallows hosts to block packets from sources hosts to block packets from sources

that send unwanted trafficthat send unwanted traffic

IGMP Message FormatsIGMP Message FormatsMembership QueryMembership Query

sent by multicast routersent by multicast router three types: general querythree types: general query, g, group-specific query, group-roup-specific query, group-

and-source specific queryand-source specific query

Membership Query FieldsMembership Query Fields

TypeType Max Response TimeMax Response Time ChecksumChecksum Group AddressGroup Address S FlagS Flag QRV (querier's robustness variable)QRV (querier's robustness variable) QQIC (querier's querier interval code)QQIC (querier's querier interval code) Number of SourcesNumber of Sources Source addressesSource addresses

IGMP Message FormatsIGMP Message FormatsMembership ReportMembership Report

IGMP Message FormatsIGMP Message FormatsGroup RecordGroup Record

1. IGMP Operation - Joining1. IGMP Operation - Joining

IGMP host IGMP host wants towants to make itself known as make itself known as group group member to other hosts and routers on LANmember to other hosts and routers on LAN

IGMPv3 IGMPv3 can can signal group membership with filtering signal group membership with filtering capabilities with respect to sourcescapabilities with respect to sources EXCLUDE modeEXCLUDE mode – all members except those listed – all members except those listed INCLUDE modeINCLUDE mode – only from group members listed – only from group members listed

to join send IGMP membership report message to join send IGMP membership report message address field multicast address of groupaddress field multicast address of group sent in IP datagram sent in IP datagram currentcurrent group members receive & learn new member group members receive & learn new member routersrouters listen to all IP multicast addresses to hear all listen to all IP multicast addresses to hear all

reportsreports

2. IGMP Operation – 2. IGMP Operation – Keeping Lists ValidKeeping Lists Valid

routersrouters periodically issue IGMP general query periodically issue IGMP general query messagemessage inin datagram with all-hosts multicast address datagram with all-hosts multicast address hostshosts must read such datagrams must read such datagrams hostshosts respond with respond with report message report message

router router don’t know every host in a groupdon’t know every host in a group needs to know at least one group member still activeneeds to know at least one group member still active each host in group sets timer with random delayeach host in group sets timer with random delay host hearing another report host hearing another report cancelscancels own own if timer expires, host sends reportif timer expires, host sends report onlyonly one member of each group one member of each group reports reports to routerto router

3. IGMP Operation - Leaving3. IGMP Operation - Leaving

hosthost leaves group leaves group by sendingby sending leave group leave group message to all-routers static multicast addressmessage to all-routers static multicast address sends asends a membership report message with membership report message with EXCLUDE EXCLUDE

optionoption and and null list of source addresses null list of source addresses

router router determines if have any remaining group determines if have any remaining group membersmembers using using group-specific query message group-specific query message

Group Membership with IPv6Group Membership with IPv6

IGMP defined for IPv4IGMP defined for IPv4 usesuses 32-bit addresses 32-bit addresses

IPv6 internets need functionalityIPv6 internets need functionality IGMP IGMP functions included in Internet functions included in Internet

Control Message ProtocolControl Message Protocol v 6 v 6 (ICMPv6) (ICMPv6) ICMPv6 has functionality of ICMPv4 & IGMPICMPv6 has functionality of ICMPv4 & IGMP

ICMPv6 includes group-membership query ICMPv6 includes group-membership query and group-membership report messageand group-membership report message

Routing ProtocolsRouting Protocols

Routing ProtocolsRouting Protocols

routers receive and forward packetsrouters receive and forward packets make decisions based on knowledge of make decisions based on knowledge of

topology and traffic/delay conditionstopology and traffic/delay conditions use dynamic routing algorithmuse dynamic routing algorithm distinguish between:distinguish between:

routing information - about topology & delaysrouting information - about topology & delays routing algorithm - that makes routing routing algorithm - that makes routing

decisions based on informationdecisions based on information

Autonomous Systems (AS)Autonomous Systems (AS)

is a group of routers and networks is a group of routers and networks managed by single organization managed by single organization

which exchange information via a common which exchange information via a common routing protocolrouting protocol

form a connected networkform a connected network at least one path between any pair of nodesat least one path between any pair of nodes except in times of failureexcept in times of failure

Interior Router Protocol & Interior Router Protocol & Exterior Routing ProtocolExterior Routing Protocol

interior router protocol (IRP)interior router protocol (IRP) passes routing information between routers within ASpasses routing information between routers within AS can be tailored to specific applicationscan be tailored to specific applications needs detailed model of network to functionneeds detailed model of network to function

may have more than one AS in internetmay have more than one AS in internet routing algorithms & tables may differ between themrouting algorithms & tables may differ between them

routers need info on networks outside own ASrouters need info on networks outside own AS use an exterior router protocol (ERP) for thisuse an exterior router protocol (ERP) for this

supports summary information on AS reachabilitysupports summary information on AS reachability

Application of IRP and ERPApplication of IRP and ERP

Approach #1 to Interior Approach #1 to Interior Routing : Routing : Distance-vector Distance-vector

each node (router or host) exchange information each node (router or host) exchange information with neighboring nodeswith neighboring nodes

first generation routing algorithm for ARPANETfirst generation routing algorithm for ARPANET eg. used byeg. used by Routing Information Protocol (RIP) Routing Information Protocol (RIP)

each node maintains vector of link costs for each node maintains vector of link costs for each directly attached network and distance and each directly attached network and distance and next-hop vectors for each destinationnext-hop vectors for each destination

requires transmission ofrequires transmission of much much info by routers info by routers distance vector & estimated path costsdistance vector & estimated path costs

changes take long changes take long time to propagatetime to propagate

Approach #2 to Interior Approach #2 to Interior Routing : Routing : Link-stateLink-state

designed to overcome drawbacks of distance-vectordesigned to overcome drawbacks of distance-vector each router determines link cost on each interfaceeach router determines link cost on each interface advertises set of link costs to all other routers in topologyadvertises set of link costs to all other routers in topology if link costsif link costs changechange,, router advertises router advertises newnew values values eacheach router constructs topology of entire configuration router constructs topology of entire configuration

can calculate shortest path to each destcan calculate shortest path to each dest use to construct use to construct routing table with first hop to each destrouting table with first hop to each dest

dodo not use distributed routing algorithm not use distributed routing algorithm, but any suitable , but any suitable alg to alg to determine shortest pathsdetermine shortest paths, eg. , eg. Dijkstra's algorithm Dijkstra's algorithm

Open Shortest Path First (OSPF) is a link-state protocolOpen Shortest Path First (OSPF) is a link-state protocol

What Exterior Routing What Exterior Routing Protocols are notProtocols are not

link-state and distance-vectorlink-state and distance-vector not not effective for effective for exterior router protocolexterior router protocol

distance-vectordistance-vector assumes routers share common distance metricassumes routers share common distance metric but different ASs may have different priorities & needsbut different ASs may have different priorities & needs but have no info on AS’s visited along routebut have no info on AS’s visited along route

link-statelink-state different ASs may use different metrics and have different ASs may use different metrics and have

different restrictionsdifferent restrictions flooding of link state information to all routers flooding of link state information to all routers

unmanageable unmanageable 

Exterior Router Protocols –Exterior Router Protocols –Path-vectorPath-vector

alternative alternative path-vectorpath-vector routing protocol routing protocol provides info about which networks can be reached provides info about which networks can be reached

by a given router and ASs crossed to get thereby a given router and ASs crossed to get there does not includedoes not include distance or cost estimatedistance or cost estimate hence dispenses with concept of routing metricshence dispenses with concept of routing metrics

have list of all ASs visitedhave list of all ASs visited on on a route a route enables router to perform policy routingenables router to perform policy routing

eg. eg. avoid path to avoid transiting particular ASavoid path to avoid transiting particular AS eg.eg. link speed, capacity, tendency to become link speed, capacity, tendency to become

congested, and overall quality of operationcongested, and overall quality of operation, security, security eg. eg. minimizing number of transit ASsminimizing number of transit ASs

Border Gateway Protocol Border Gateway Protocol (BGP)(BGP)

developed for use with TCP/IP internetsdeveloped for use with TCP/IP internets is preferred EGP of the Internetis preferred EGP of the Internet uses messages sent over TCP connectionuses messages sent over TCP connection current version is BGP-4 (RFC1771)current version is BGP-4 (RFC1771) functional proceduresfunctional procedures

neighbor acquisition - when agree to exchange infoneighbor acquisition - when agree to exchange info neighbor reachability - to maintain relationshipneighbor reachability - to maintain relationship network reachability - to update database of routesnetwork reachability - to update database of routes

BGP BGP MessagesMessages

OpenOpen UpdateUpdate Keep aliveKeep alive NotificationNotification

(Message Types -(Message Types -Open & KeepAlive)Open & KeepAlive)

router makes TCP connection to neighborrouter makes TCP connection to neighbor Open messageOpen message

sent by connection initiatorsent by connection initiator includes proposed hold timeincludes proposed hold time receiver uses minimum of own/sent hold time receiver uses minimum of own/sent hold time max time between Keepalive and/or Updatemax time between Keepalive and/or Update

Keep Alive messageKeep Alive message To tell other routers that this router is still hereTo tell other routers that this router is still here

(Message Types – Update)(Message Types – Update)

Update message conveys two info types:Update message conveys two info types: Info about single routes through internetInfo about single routes through internet List of routes being withdrawnList of routes being withdrawn

info on a route uses 3 fields:info on a route uses 3 fields: Network Layer Reachability Information (NLRI)Network Layer Reachability Information (NLRI) Total Path Attributes Length Total Path Attributes Length Path AttributesPath Attributes

withdraw route identified by dest IP addresswithdraw route identified by dest IP address

(Message Types – Update)(Message Types – Update)

Origin - IGP or EGPOrigin - IGP or EGP AS_Path - list of AS traversedAS_Path - list of AS traversed Next_hop - IP address of border routerNext_hop - IP address of border router Multi_Exit_Disc - info on routers internal to ASMulti_Exit_Disc - info on routers internal to AS Local_pref - inform routers in AS of route prefLocal_pref - inform routers in AS of route pref Atomic_Aggregate, Aggregator - implement Atomic_Aggregate, Aggregator - implement

route aggregation to reduce amount of inforoute aggregation to reduce amount of info

(AS_Path and Next_Hop Use)(AS_Path and Next_Hop Use)

AS_PathAS_Path used to implement routing policiesused to implement routing policies

• eg. to avoid a particular AS, security, performance, eg. to avoid a particular AS, security, performance, quality, number of AS crossedquality, number of AS crossed

Next_HopNext_Hop only a few routers implement BGPonly a few routers implement BGP responsible for informing outside routers of responsible for informing outside routers of

routes to other networks in ASroutes to other networks in AS

(Notification Message)(Notification Message)

sent when some error condition detected:sent when some error condition detected: Message header errorMessage header error Open message errorOpen message error Update message errorUpdate message error Hold time expiredHold time expired Finite state machine errorFinite state machine error CeaseCease

(BGP Routing Information (BGP Routing Information Exchange)Exchange)

within AS a router builds topology picture within AS a router builds topology picture using IGPusing IGP

router issues Update message to other router issues Update message to other routers outside AS using BGProuters outside AS using BGP

these routers exchange info with other these routers exchange info with other routers in other ASrouters in other AS AS_Path field used to prevent loopsAS_Path field used to prevent loops

routers must then decide best routesrouters must then decide best routes

Open Shortest Path First Open Shortest Path First (RFC2328)(RFC2328)

IGP of InternetIGP of Internet replaced Routing Information Protocol (RIP)replaced Routing Information Protocol (RIP) uses Link State Routing Algorithmuses Link State Routing Algorithm

each router keeps list of state of local links to networkeach router keeps list of state of local links to network transmits update state infotransmits update state info little traffic as messages are small and not sent oftenlittle traffic as messages are small and not sent often

uses least cost based on user cost metricuses least cost based on user cost metric topology stored as directed graphtopology stored as directed graph

vertices or nodes (router, transit or stub network)vertices or nodes (router, transit or stub network) edges (between routers or router to network)edges (between routers or router to network)

Example Example OSPF ASOSPF AS

Directed Directed Graph of Graph of

ASAS

SPF TreeSPF Treefor for

Router 6Router 6

(Integrates Services (Integrates Services Architecture)Architecture)

changes in traffic demands require variety changes in traffic demands require variety of quality of service of quality of service eg. internet phone, multimedia, multicasteg. internet phone, multimedia, multicast

new functionality required in routersnew functionality required in routers new means of requesting QoSnew means of requesting QoS IETF developing a suite of Integrated IETF developing a suite of Integrated

Services Architecture (ISA) standardsServices Architecture (ISA) standards RFC 1633 defines overall view of ISARFC 1633 defines overall view of ISA

(Internet Traffic Categories)(Internet Traffic Categories)

elastic trafficelastic traffic can cope with wide changes in delay and/or can cope with wide changes in delay and/or

throughputthroughput traditional TCP/IP traffictraditional TCP/IP traffic eg. FTP, email, telnet, SNMP, HTTPeg. FTP, email, telnet, SNMP, HTTP different sensitivity to throughput, delay, different sensitivity to throughput, delay,

congestioncongestion inelastic trafficinelastic traffic

does not easily adapt to variationsdoes not easily adapt to variations

(Inelastic Traffic (Inelastic Traffic Requirements)Requirements)

throughputthroughput delaydelay jitterjitter packet losspacket loss

need preferential treatment for some traffic need preferential treatment for some traffic typestypes

require elastic traffic to be supportedrequire elastic traffic to be supported

(ISA Approach)(ISA Approach)

IP nets control congestion byIP nets control congestion by routing algorithmsrouting algorithms packet discardpacket discard

ISA provides enhancements to traditional IPISA provides enhancements to traditional IP in ISA associate each packet with a flowin ISA associate each packet with a flow ISA functions:ISA functions:

admission controladmission control routing algorithmrouting algorithm queuing disciplinequeuing discipline discard policydiscard policy

(ISA (ISA in Router)in Router)

(ISA Services)(ISA Services)

GuaranteedGuaranteed assured data rateassured data rate upper bound on queuing delayupper bound on queuing delay no queuing lossno queuing loss

Controlled loadControlled load approximates best effort behavior on unloaded netapproximates best effort behavior on unloaded net no specific upper bound on queuing delayno specific upper bound on queuing delay very high delivery successvery high delivery success

Best EffortBest Effort traditional IP servicetraditional IP service

(Token Bucket Scheme)(Token Bucket Scheme)

(Queuing Discipline)(Queuing Discipline)

traditionally FIFOtraditionally FIFO no special treatment for high priority flow packetsno special treatment for high priority flow packets large packet can hold up smaller packetslarge packet can hold up smaller packets greedy connection can crowd out less greedy greedy connection can crowd out less greedy

connectionconnection

need some form of fair queuingneed some form of fair queuing multiple queues used on each output portmultiple queues used on each output port packet is placed in queue for its flowpacket is placed in queue for its flow round robin servicing of queuesround robin servicing of queues can have weighted fair queuingcan have weighted fair queuing

(FIFO and Fair Queue)(FIFO and Fair Queue)

(Resource Reservation: (Resource Reservation: RSVP)RSVP)

RFC 2205RFC 2205 unicast applications can reserve resources in unicast applications can reserve resources in

routers to meet QoSrouters to meet QoS if router can not meet request, application informedif router can not meet request, application informed

multicast more demanding, but may be reducedmulticast more demanding, but may be reduced some members of group may not require delivery some members of group may not require delivery

from particular source over given timefrom particular source over given time some group members may only be able to handle a some group members may only be able to handle a

portion of the transmissionportion of the transmission reservation means routers can decide in advance if reservation means routers can decide in advance if

can meet requirementscan meet requirements

(Soft State)(Soft State)

have different resource reservation needs have different resource reservation needs to traditional connection-oriented networksto traditional connection-oriented networks must dynamically changemust dynamically change

use concept of Soft Stateuse concept of Soft State set of state info in router that expires unless set of state info in router that expires unless

refreshedrefreshed applications must periodically renew applications must periodically renew

requests during transmissionrequests during transmission

(RSVP Characteristics)(RSVP Characteristics)

unicast and multicastunicast and multicast simplexsimplex receiver initiated reservationreceiver initiated reservation maintain soft state in the internetmaintain soft state in the internet provide different reservation stylesprovide different reservation styles transparent operation through non-RSVP transparent operation through non-RSVP

routersrouters support for IPv4 and IPv6support for IPv4 and IPv6

(Differentiated Services)(Differentiated Services)

simple, easily implemented, low overhead tool to simple, easily implemented, low overhead tool to support a range of differentiated network servicessupport a range of differentiated network services

IP Packets labeled for differing QoS using existing IP Packets labeled for differing QoS using existing IPv4 Type of Service or IPv6 DS fieldIPv4 Type of Service or IPv6 DS field

have service level agreement established have service level agreement established between provider and customer prior to use of DSbetween provider and customer prior to use of DS

have built in aggregationhave built in aggregation implemented by queuing and forwarding based on implemented by queuing and forwarding based on

DS octetDS octet most widely used QoS mechanism todaymost widely used QoS mechanism today

(DS Domains)(DS Domains)

(DS Services)(DS Services)

is defined within a DS domainis defined within a DS domain a contiguous portion of internet over which consistent a contiguous portion of internet over which consistent

set of DS policies are administeredset of DS policies are administered typically under control of one organizationtypically under control of one organization defined by service level agreements (SLA)defined by service level agreements (SLA) specify service received for classes of packetsspecify service received for classes of packets

once established customer submits packets with once established customer submits packets with DS marked indicating classDS marked indicating class service provider ensures agreed QoS within domainservice provider ensures agreed QoS within domain if transit other domains, provider chooses closest QoSif transit other domains, provider chooses closest QoS

(SLA Parameters)(SLA Parameters)

detailed service performance such as:detailed service performance such as: expected throughputexpected throughput drop probabilitydrop probability latencylatency

constraints on ingress and egress pointsconstraints on ingress and egress points traffic profilestraffic profiles disposition of traffic in excess of profiledisposition of traffic in excess of profile

(Example Services)(Example Services)

level A - low latencylevel A - low latency level B - low losslevel B - low loss level C - 90% of traffic < 50ms latencylevel C - 90% of traffic < 50ms latency level D - 95% in profile traffic deliveredlevel D - 95% in profile traffic delivered level E - allotted twice bandwidth of level Flevel E - allotted twice bandwidth of level F level F - with drop precedence X has level F - with drop precedence X has

higher probability of delivery than that of Yhigher probability of delivery than that of Y

(DS Field)(DS Field)

(DS Field - DS Codepoint)(DS Field - DS Codepoint)

6 bit field in IPv4 & IPv6 header6 bit field in IPv4 & IPv6 header 3 pools of code points3 pools of code points

xxxxx0 - assignment as standardsxxxxx0 - assignment as standards• 000000 - default best effort000000 - default best effort• xxx000 - IPv4 precedence compatibilityxxx000 - IPv4 precedence compatibility

xxxx11 - experimental or local usexxxx11 - experimental or local use xxxx01 - experimental or local but may be xxxx01 - experimental or local but may be

allocated for standards in futureallocated for standards in future

(IPv4 Precedence Service)(IPv4 Precedence Service)

IPv4 TOS field included subfields IPv4 TOS field included subfields precedence (3 bit) - datagram urgency/priorityprecedence (3 bit) - datagram urgency/priority TOS(4 bit) - guidance on selecting next hopTOS(4 bit) - guidance on selecting next hop

can respond withcan respond with route selection - smaller queue, has priorityroute selection - smaller queue, has priority network service - supports precedencenetwork service - supports precedence queuing discipline - support precedence queuing discipline - support precedence

ordered queueing & discard lower precedenceordered queueing & discard lower precedence

(DS Configuration and (DS Configuration and Operation)Operation)

within domain, interpretation of DS code within domain, interpretation of DS code points is uniformpoints is uniform

interior nodesinterior nodes implement simple mechanismsimplement simple mechanisms per-hop behavior (PHB) on all routersper-hop behavior (PHB) on all routers

boundary nodesboundary nodes have PHB & more sophisticated mechanismshave PHB & more sophisticated mechanisms hence most of complexityhence most of complexity

(DS Traffic Conditioner)(DS Traffic Conditioner)

(Per Hop Behavior – (Per Hop Behavior – Expedited Forwarding) Expedited Forwarding)

specific PHBs definedspecific PHBs defined expedited forwarding (EF) PHB (RFC 3246)expedited forwarding (EF) PHB (RFC 3246)

low-loss, low-delay, low-jitter, assured bandwidth, low-loss, low-delay, low-jitter, assured bandwidth, end-to-end service through DS domainsend-to-end service through DS domains

simulates a point-to-point connection or leased linesimulates a point-to-point connection or leased line difficultdifficult in in internet or packet-switching network internet or packet-switching network

queues queues onon node/router node/router r result in loss, delays, and jitteresult in loss, delays, and jitter unless internet grossly oversizedunless internet grossly oversized,, care care needed needed in in

handling premium service handling premium service traffic traffic EF PHB intent is to use empty/short queues to EF PHB intent is to use empty/short queues to

minimise delay, jitter & packet loss.minimise delay, jitter & packet loss.

(Expedited Forwarding(Expedited Forwarding RequirementsRequirements ) )

EF PHB designed to configure nodes so traffic EF PHB designed to configure nodes so traffic aggregate has minimum departure rateaggregate has minimum departure rate

border routers condition traffic aggregate (via border routers condition traffic aggregate (via policing / shaping) so arrival rate is less than policing / shaping) so arrival rate is less than minimum departure rate for nodesminimum departure rate for nodes

interior nodes treat traffic interior nodes treat traffic so noso no queuing effects queuing effects no specificno specific queuing policy set for interior nodes queuing policy set for interior nodes note a simple priority scheme can achieve note a simple priority scheme can achieve thisthis

EF traffic given absolute priorityEF traffic given absolute priority EF traffic EF traffic must notmust not overwhelm interior node overwhelm interior node but packet flows for other PHB traffic disruptedbut packet flows for other PHB traffic disrupted

(Assured Forwarding PHB)(Assured Forwarding PHB)

provide service superior to best-effortprovide service superior to best-effort without needing without needing reservation of resourcesreservation of resources or or

detailed flow discriminationdetailed flow discrimination based on explicit allocationbased on explicit allocation

users offered choice of classes of serviceusers offered choice of classes of service traffic monitored at boundary node, marked in/outtraffic monitored at boundary node, marked in/out inside network, no separation of traffic from different inside network, no separation of traffic from different

users or classesusers or classes when when congested,congested, drop out packets before in packets drop out packets before in packets different users will see different levels of servicedifferent users will see different levels of service

advantage is simplicityadvantage is simplicity

(AF(AF PHB PHB RFC 2597) RFC 2597)

four AF classes / traffic profiles are definedfour AF classes / traffic profiles are defined within each class, packets marked within each class, packets marked with twith three hree

drop precedence valuesdrop precedence values in congestion determines relative importancein congestion determines relative importance

simpler, more simpler, more flexibleflexible than resource reservation than resource reservation within interior DS node, traffic from within interior DS node, traffic from different different

classes is classes is treated separatelytreated separately different resources (buffer space, data rate)different resources (buffer space, data rate)

hence forwarding assurance depends on hence forwarding assurance depends on resources, current load & drop precedenceresources, current load & drop precedence

(Service Level Agreements)(Service Level Agreements)

is a contract between network provider is a contract between network provider and customer for aspects of serviceand customer for aspects of service

typically includes:typically includes: description of nature of servicedescription of nature of service expected performance level of serviceexpected performance level of service process for monitoring & reporting service process for monitoring & reporting service

levellevel similar to frame relay / ATM SLA’ssimilar to frame relay / ATM SLA’s but more difficult to realizebut more difficult to realize

(Service Level Agreements)(Service Level Agreements)

(IP Performance Metrics)(IP Performance Metrics)

IP Performance Metrics working group is IP Performance Metrics working group is developing a standard set of metricsdeveloping a standard set of metrics on quality, performance, reliabilityon quality, performance, reliability to provide common understandingto provide common understanding

3 stages of metrics3 stages of metrics singleton metric - elementary / atomic quantitysingleton metric - elementary / atomic quantity sample metric - taken over time periodsample metric - taken over time period statistical metric - derived from samplestatistical metric - derived from sample

active or passive measurementactive or passive measurement

(IP Performance Metrics)(IP Performance Metrics)Metric Name

One-Way Delay

Round-Trip Delay

One-Way Loss

One-Way Loss Pattern (distance / period)

Packet Delay Variation

Connectivity

Bulk Transfer Capacity

(Summary)(Summary)

reviewed various internetwork services & reviewed various internetwork services & functions to support varying servicesfunctions to support varying services

multicastingmulticasting routing protocolsrouting protocols integrated services architectureintegrated services architecture differentiated servicesdifferentiated services service level agreementsservice level agreements IP performance metricsIP performance metrics