Presentation on theme: "Chapter 11: Internet Operation"— Presentation transcript:
1Chapter 11: Internet Operation Business Data Communications, 7e
2Objectives Internet Addressing Internet Routing Protocols The Need for Speed and Quality of serviceDifferentiated Services
3Internet Addressing32-bit global internet address for source & destination in the IP header (base on IPv4)Includes a network identifier and a host identifierDotted decimal notation(binary)(decimal)
4Class-Based IP Addresses Rightmost bits of the 32-bit IP address designate a hostThe leftmost bits of the 32-bit address designate a networkClass-based, or classful, IP addressing was adopted to allow for a variable allocation of bits to specify network and hostThe first few leftmost bits specify how the rest of the address should be separated into network and host fieldsThis provides flexibility in assigning addresses to hosts and allows a mix of network sizes on an internetIn general terms, the rightmost, or least significant,bits of the 32-bit IP address designate a host, and the leftmost, or most significant,bits designate a network. A fixed allocation of bits, such as 16 bits for networknumber and 16 bits for host, was deemed inadequate to handle the global Internet,where some organizations might have a few networks, each with many hosts andsome organizations might have many networks, each with a few hosts. Therefore, ascheme known as class-based , or classful , IP addressing was adopted.Class-based IP addresses allow for a variable allocation of bits to specifynetwork and host. For this scheme, the first few leftmost bits specify how the restof the address should be separated into network and host fields. This encoding providesflexibility in assigning addresses to hosts and allows a mix of network sizes onan internet.
5Network ClassesClass A: Few networks, each with many hosts All addresses begin with binary 0Class B: Medium networks, medium hosts All addresses begin with binary 10Class C: Many networks, each with few hosts All addresses begin with binary 110
6Format of IP AddressShort for multicast backbone. A small set of Internet sites, each of which can transmit real-time audio and video simultaneously to all the others. MBONE sites are equipped with special software to send and receive packets at high speed using the IP one-to-many multicasting protocol. The MBONE has been used for video conferencing and even for a Rolling Stones concert in 1994.
7Network Classes (cont.) IP addresses are usually written in: “Dotted Decimal Notation”, i.e. a decimal number represent each byte of the 32-bit address.Example: Binary representation of an IP is : Decimal representation is: (decimal).
8Network Classes (cont.) Class A Network begins with 0Note: Network addresses ( ) and ( ) are reserved Therefore Class A contains: ( = = 126) network numbersRange of the 1st decimal number for Class A: 1.***.***.*** to 127.***.***.***
9Network Classes (cont.) Class B begin with binary 10 starts from (128) ends to (191) i.e. Range of the 1st decimal number for Class B: 128.***.***.*** to 191.***.***.*** the 2nd Byte is also part of class B i.e. there are 214 = 16,384 Class B addressesClassB
10Network Classes (cont.) Class C begin with binary 110 starts from (192) ends to (223) Range of the 1st decimal number for class C: 192.***.***.*** to 223.***.***.*** the 2nd & 3rd Byte is also part of class C There are 221 = 2,097,152 Class C addresses
11Subnets & Subnet MasksAllows for subdivision of internets within an organization and add a number of LANs to the internet and insulate their internal complexity within their organization by assigning a single “network number” to all the LANsEach LAN can have a subnet number, allowing routing among networksHost portion is partitioned into subnet and host numbersFrom the point of view of the rest of the internet, there is a single network at that site.This simplifies addressing and routing.
12Subnets & Subnet Masks (Cont.) Then to allow the Routers within the site to function properly, each LAN is assigned a subnet number.32-bit Source Address32-bit Source Address
13Subnets & Subnet Masks (Cont.) To include the subnet number, the host portion of the internet address is partitioned into a subnet number and a host number to accommodate this new level of addressing.Network Portion:Class A: 7 + 1bitsClass B: 14+2 bitsClass C: bitsHost Portion:Class A: 24bitClass B: 16 bitClass C: 8 bitExtended Network Number or Address Mask:NetworkHostWithin the subnetted network, the local Routers must route on the basis of an extended network numberNetworkSubnetHost
14Subnets & Subnet Masks (Cont.) The use of address mask allows the host to determine whether an outgoing datagram is destined for a host on the same LAN (send directly) or another LAN (send datagram to router)Some methods (manual config.) are used to create address masks and make them known to the local routers
15Subnets & Subnet Masks (Cont.) The effect of the subnet mask is to erase the portion of the host field that refers to an actual host on a subnet. What remains is the network number and the subnet number.
17Subnets & Subnet Masks (Cont.) A local complex consisting of 3 LANs and 2 Routers. To the rest of the internet, this complex is a single network with a class C address of the form X, where 192 ( ) is the network number and x the host number.Example of Subnetworking:
19Subnets & Subnet Masks (Cont.) IP Address:Host number:25IP Address:Host number:1Net ID/subnet ID:Subnet number:1Net ID/subnet ID :Subnet number:2IP Address:Net ID/subnet ID :Subnet number:3IP Address:Example1: A datagram with the destination address arrives at R1 from the rest of the internet or from LAN Y. R1 has addresses of LAN X, LAN Y, LAN Z. R1 doesn’t know about hosts internal to these LANs.In order to determine where R1 should send the datagram with receiver address R1 bitwise AND the subnet mask:( ) i.e. ( ) and IP address ( ) to determine that destination address refers to subnet:( ) i.e. 1, which is LAN X, and so forward the datagram to LAN X.For both R1 & R2 RoutersThe effect of the subnet mask is to erase the portion of the host field that refers to an actual host on a subnet. What remains is the network number and the subnet number.
20IP Address & Subnet Masks Binary RepresentationDotted DecimalIP AddressSubnet Mask for both R1 & R2 RoutersBitwise AND of address and mask (resultant network/subnet number)Subnet number1Host number251.25
22Hosts must also employ a subnet mask to make routing decisions. Example2: If a datagram with destination address ( ) arrives at R2 from LAN Z, R2 applies the mask and then determines from its forwarding database that datagrams destined for subnet 1 should be forwarded to R1Hosts must also employ a subnet mask to make routing decisions.The default subnet mask for a give class of addresses is a null mask, which yields the same network and host number as the non-subnetted address.IP Address:Host number:25IP Address:Host number:1Net ID/subnet ID:Subnet number:1Net ID/subnet ID :Subnet number:2IP Address:Net ID/subnet ID :Subnet number:3IP Address:Subnets &Subnet Masks (Cont.)
23Classless Inter-Domain Routing (CIDR) Makes more efficient use of the 32-bit IP address than the class-based methodDoes away with the class designation and with the use of leading bits to identify a classEach 32-bit address consists of a leftmost network part and a rightmost host part, with all 32 bits used for addressingAssociated with each IP address is a prefix value that indicates the length of the network portion of the addressA CIDR IP address is written as a.b.c.d/pa is the value of the first byte of the addressb the value of the second bytec the value of the third byted the value of the fourth bytep is in the range of 1 through 32 and indicates the length of the network portion of the addressBy the mid-1990s, it became evident toInternet designers and administrators that the 32-bit class-based addressing schemewas woefully inadequate for the growing demand for IP addresses. The long-termsolution to this problem, as described in Chapter 8, was the development of IPv6,which includes 128-bit address fields. The use of 128-bit addresses increases the numberof possible unique addresses by a factor of almost compared to the use of32-bit addresses.However, the deployment of IPv6 would take many years so, as an interimmeasure, CIDR was adopted. CIDR makes more efficient use of the 32-bit IPaddress than the class-based method primarily because it makes more efficientuse of the address space. With class-based addressing, an organization can requesta block of addresses that provides 8, 16, or 24 bits for host addresses. BecauseInternet addresses were typically only assigned as blocks of a certain class, therewere a lot of wasted addresses.CIDR does away with the class designation and with the use of leading bitsto identify a class. Instead, each 32-bit address consists of a leftmost network partand a rightmost host part, with all 32 bits used for addressing. Associated with eachIP address is a prefix value that indicates the length of the network portion of theaddress. A CIDR IP address is written as a .b .c .d /p , where a is the value of the firstbyte of the address, b the value of the second byte, c the value of the third byte, andd the value of the fourth byte. Each of these values is in the range of 0 to 255. Theprefix value p is in the range of 1 through 32 and indicates the length of the networkportion of the address.In CIDR notation, a prefix is shown as a 4-octet quantity, just like a traditionalIPv4 address or network number, followed by the “/” (slash) character, followedby a decimal value from 0 through 32. For example, the legacy “Class B” network, with an implied network mask of , is defined as the prefix/16, the “/16” indicating that the mask to extract the network portion of theprefix is a 32-bit value where the most significant 16 bits are ones and the least significant16 bits are zeros. Similarly, the legacy “Class C” network numberis defined as the prefix /24; the most significant 24 bits are ones and theleast significant 8 bits are zeros.Note that each 32-bit address still has (and must have) a unique interpretation.That is, each IP address must have associated with it a prefix value p for proper routingto the correct network and delivery to the correct host. However, the IP address fieldonly provides space for the 32-bit IP address and not for the prefix value. Accordingly,each CIDR routing table entry in each Internet router contains a 32-bit IP address anda 32-bit network mask, which together give the length of the IP prefix. Clearly, it wouldbe impractical to have an entry for each of the 232 possible IP addresses together witha mask at each router. Instead, multiple IP addresses referring to a block of CIDRaddresses can be identified with a single mask, a process known as supernetting.Examples:Class B Network with an implied network mask is defined as /1616 bits 1 and 16 bits 0Class C Network with /2424 bits 1 and 8 bits 0Supernetting: Multiple IP addresses referring to a block of CIDR addresses can be identified with a single mask.
24IPv6 AddressesIPv6 addresses are 128 bits in length. Addresses are assigned to individual interfaceson nodes, not to the nodes themselves. A single interface may have multipleunique unicast addresses. Any of the unicast addresses associated with a node’sinterface may be used to uniquely identify that node. As with IPv4, IPv6 addressesuse CIDR rather than address classes.IPv6 addresses are 128 bits in length. Addresses are assigned to individual interfaceson nodes, not to the nodes themselves. A single interface may have multipleunique unicast addresses. Any of the unicast addresses associated with a node’sinterface may be used to uniquely identify that node. As with IPv4, IPv6 addressesuse CIDR rather than address classes.The combination of long addresses and multiple addresses per interfaceenables improved routing efficiency over IPv4. Longer internet addresses allowfor aggregating addresses by hierarchies of network, access provider, geography,corporation, and so on. Such aggregation should make for smaller routing tablesand faster table lookups. The allowance for multiple addresses per interface wouldallow a subscriber that uses multiple access providers across the same interface tohave separate addresses aggregated under each provider’s address space.IPv6 allows three types of addresses (Figure 11.2):• Unicast: An identifier for a single interface. A packet sent to a unicast addressis delivered to the interface identified by that address.• Anycast: An identifier for a set of interfaces (typically belonging to differentnodes). A packet sent to an anycast address is delivered to one of the interfacesidentified by that address (the “nearest” one, according to the routingprotocols’ measure of distance).• Multicast: An identifier for a set of interfaces (typically belonging to differentnodes). A packet sent to a multicast address is delivered to all interfacesidentified by that address.The notation for an IPv6 address uses eight hexadecimal number to representthe eight 16-bit blocks in the 128-bit address, with the numbers divided by colons.For example:FE80:0000:0000:0000:0001:0800:23E7:F5DBTo make the notation more compact, leading zeroes in any hexadecimal number areomitted. For the preceding example, the result is:FE80:0:0:0:1:800:23E7:F5DBTo further compress the representation, a zero or any contiguous sequence of zeroesis replaced by a double colon. For our example, the result is:FE80::1:800:23E7:F5DBAnycast Address
25Internet Routing Protocols Routers are responsible for receiving and forwarding packets between interconnected networksRouters make decisions based on the knowledge of the topology and traffic/delay conditions of the Internet. (based on topology leads to a static -permanent- route based on the traffic makes it a dynamic route)Must dynamically adapt to changing network conditions to avoid congested and failed portions of the network.Two key concepts to distinguish in routing function:Routing information RI: Information about topology & delaysRouting algorithm: The algorithm used to make a routing decision for a particular datagram, based on the current RIThe routers in an internet are responsible for receiving and forwarding packetsthrough the interconnected set of networks. Each router makes routing decisionsbased on knowledge of the topology and traffic/delay conditions of the internet.In a simple internet, a fixed routing scheme is possible, in which a single,permanent route is configured for each source–destination pair of nodes in thenetwork. The routes are fixed, or at most only change when there is a change inthe topology of the network. Thus, the link costs used in designing routes cannotbe based on any dynamic variable such as traffic. They could, however, be basedon estimated traffic volumes between various source–destination pairs or thecapacity of each link.In more complex internets, a degree of dynamic cooperation is needed amongrouters. In particular, routers must avoid portions of the network that have failedand should avoid portions of the network that are congested. To make such dynamicrouting decisions, routers exchange routing information using a special routingprotocol for that purpose. Information is needed about which networks can bereached by which routes, and the delay characteristics of various routes.In considering the routing function, it is important to distinguish two concepts:• Routing information: Information about the topology and delays of theinternet• Routing algorithm: The algorithm used to make a routing decision for a particulardatagram, based on current routing information
26Autonomous Systems (AS) To proceed with Routing Protocol let’s introduce AS:Key characteristics of an ASSet of routers and networks managed by a single organizationSet of routers exchanging information via a common routing protocolConnected (in a graph-theoretic sense); that is, there is a path between any pair of nodes (except in times of failure).Interior Router Protocol (IRP) passes information between routers within an ASExterior Router Protocol (ERP) passes information between routers in different ASsThe protocol used within the AS does not need to be implemented outside of the systemThis flexibility allows IRPs to be custom tailored to specific applications and requirementsTo proceed with our discussion of routing protocols, we need to introduce the conceptof an autonomous system (AS) . An AS exhibits the following characteristics:1. An AS is a set of routers and networks managed by a single organization.2. An AS consists of a group of routers exchanging information via a commonrouting protocol.3. Except in times of failure, an AS is connected (in a graph-theoretic sense);that is, there is a path between any pair of nodes.A shared routing protocol, which we shall refer to as an interior routerprotocol (IRP) , passes routing information between routers within an AS. Theprotocol used within the AS does not need to be implemented outside of thesystem. This flexibility allows IRPs to be custom tailored to specific applicationsand requirements.
27Application of Interior and Exterior Routing Protocols It may happen, however, that an internet will be constructed of more thanone AS. For example, all of the LANs at a site, such as an office complex or campus,could be linked by routers to form an AS. This system might be linked througha wide area network to other ASs. The situation is illustrated in Figure Inthis case, the routing algorithms and information in routing tables used by routersin different ASs may differ. Nevertheless, the routers in one AS need at least aminimal level of information concerning networks outside the system that canbe reached. We refer to the protocol used to pass routing information betweenrouters in different ASs as an exterior router protocol (ERP) .In general terms, IRPs and ERPs have a somewhat different flavor. An IRPneeds to build up a rather detailed model of the interconnection of routers withinan AS in order to calculate the least-cost path from a given router to any networkwithin the AS. An ERP supports the exchange of summary reachability informationbetween separately administered ASs. Typically, this use of summary informationmeans that an ERP is simpler and uses less detailed information than an IRP.In the remainder of this section, we look at what are perhaps the most importantexamples of these two types of routing protocols: BGP and OSPF.Autonomous System 1Autonomous System 2Interior router ProtocolExterior router protocol
28IRP & ERP IRP: Interior router protocol ERP: Exterior router protocol Needs to build up a detailed model of the interconnection of routers within an AS in order to calculate the least-cost path from a given router to any network within the ASERP: Exterior router protocolSupports the exchange of summary reachability information between separately administered ASs. Use of summary information means that an ERP is simpler and uses less detailed information than an IRP
29Border Grouping Protocol (BGP) BGP was designed to allow routers (called gateways) in different AS to cooperate in the exchange of routing information.BGP has become the preferred ERP (Exterior Router Protocol) for the internets that employ TCP/IP suite.BGP has 3 functional procedures:1. Neighbor acquisition2. Neighbor reachability3. Network reachabilityBPG: Preferred exterior router protocol for the Internet.Neighbors: Two routers are considered as Neighboring routers if attached to the same network or share the same network.Routers in different autonomous systems may wish to exchange routing information.For this purpose, it is necessary to perform neighbor acquisition.Neighbor acquisition occurs when two neighboring routers in different autonomous systems agree to exchange routing information regularly.Protocol does not address how one router knows the address/existence of another, it is decided at the configuration time by network manager.To perform neighbor acquisition one router send an open message to another. If the target router accepts the request, it returns a keepalive message.Neighbor reachability procedure is used after neighbor relationship is established in order to maintain the relationship by each partner periodically send keepalive messages to each other.Network reachability is that each router maintains a database of the networks that it can reach and the preferred route for reaching each network. When a change is made to this database, the router broadcast an update message and all BGP routers can build up and maintain their routing information.
30Open Shortest Path First (OSPF) Widely used as IRP (Interior Router Protocol) in TCP/IP networksUses link state routing algorithmRouters maintain topology database of ASTopology is express as directed graph consisting of:RouterNetworkCarry data that neither originates nor terminates on an end system attached to this networkVertices or Nodes:Transit:Stub:Open Shortest Path First is an interior router protocol (IRP) for TCP/IP.Each router maintains descriptions of the state of its local links to networks,and transmits updates from time to time.OSPF computes a route through the internet that incurs the least cost based on a user-configurable metric of cost base on either delay, data rate, dollar cost, etc.OSPF is able to equalize loads over multiple equal-cost paths.Each router maintains a database that reflects the known topology of the autonomous system of which it is a part. The topology is expressed as a directed graph, consisting of the Vertices,…If it is not a transit networkConnecting router vertices of two router connected by point-to-point link.Connecting router vertex to network vertex of directly connected.Edges
31Open Shortest Path First (OSPF)Cnt’d An Autonomous SystemDirected Graph of the Autonomous SystemRouters 6 & 10: Joined by a point-to-point link are represented by a pair of edges directly connected by a pair of edges in the graph.Routers 1,2,3, and 4 to network 3: Multiple routers are attached to a network. The directed graph shows all routers bi-directionally connected to the network vertex.Network 7: A single router is attached to a network, the network will appear in the graph as a stub.Host 1: an end system is directly connected to a router.Networks 12 to 15: A router is connected to other autonomous systems, then the path cost to each network in the other system must be obtained by some exterior routing protocol (ERP).Each such network is represented on the graph by a stub and an edge to the router with the known path cost.
32Open Shortest Path First (OSPF)Cnt’d Routers 6 & 10: Joined by a point-to-point link are represented by a pair of edges directly connected by a pair of edges in the graph.Routers 1,2,3, and 4 to network 3: Multiple routers are attached to a network. The directed graph shows all routers bi-directionally connected to the network vertex.Network 7: A single router is attached to a network, the network will appear in the graph as a stub.Host 1: an end system is directly connected to a router.Networks 12 to 15: A router is connected to other autonomous systems, then the path cost to each network in the other system must be obtained by some exterior routing protocol (ERP).Each such network is represented on the graph by a stub and an edge to the router with the known path cost.Directed Graph of the Autonomous SystemSPF tree for R6An Autonomous System
33SPF tree & Routing Table for Router R6 -A cost is associated with the output side of each router interface.-This cost is configurable by the system administrator.-Arcs on the graph are labeled with the cost of the corresponding router output interface (NOT INPUT, e.g. cost of Arc between R6 to R5 =6 but R6 to R5 = 7 therefore 6 should be taken not 7).Arcs with no label cost have a cost of 0.Arcs from networks to routers always have a cost of 0.A database of the directed graph is maintained by each router.This database is pieced together from link state messages from other routers in the internet.A router uses an algorithm to calculate the least-cost path to all destination networks.SPF tree for R6
34MulticastingSending a packet from a source to the members of a multicast groupMulticast addressesAddresses that refer to a group of hosts on one or more networksPractical applications include:MultimediaTeleconferencingDatabaseDistributed computationReal-time workgroupTypically, an IP address refers to an individual host on a particular network. IP alsoaccommodates addresses that refer to a group of hosts on one or more networks.Such addresses are referred to as multicast addresses , and the act of sending a packetfrom a source to the members of a multicast group is referred to as multicasting .Multicasting has a number of practical applications. For example,• Multimedia: A number of users “tune in” to a video or audio transmissionfrom a multimedia source station.• Teleconferencing: A group of workstations form a multicast group such that atransmission from any member is received by all other group members.• Database: All copies of a replicated file or database are updated at the sametime.• Distributed computation: Intermediate results are sent to all participants.• Real-time workgroup: Files, graphics, and messages are exchanged amongactive group members in real time.Multicasting done within the scope of a single LAN segment is straightforward.IEEE 802 and other LAN protocols include provision for MAC-levelmulticast addresses. A packet with a multicast address is transmitted on a LANsegment. Those stations that are members of the corresponding multicast grouprecognize the multicast address and accept the packet. In this case, only a singlecopy of the packet is ever transmitted. This technique works because of the broadcastnature of a LAN: a transmission from any one station is received by all otherstations on the LAN.
35Illustration of Multicasting In an internet environment, multicasting is a far more difficult undertaking. Tosee this, consider the configuration of Figure 11.5, in which a number of LANs areinterconnected by routers. Routers connect to each other either over high-speedlinks or across a wide area network (network N4). A cost is associated with eachlink or network in each direction, indicated by the value shown leaving the routerfor that link or network. Suppose that the multicast server on network N1 is transmittingpackets to a multicast address that represents the workstations indicatedon networks N3, N5, and N6. Suppose that the server does not know the locationof the members of the multicast group. Then one way to assure that the packetis received by all members of the group is to broadcast a copy of each packet toeach network in the configuration, over the least-cost route for each network. Forexample, one packet would be addressed to N3 and would traverse N1, link L3, andN3. Router B is responsible for translating the IP-level multicast address to a MAC levelmulticast address before transmitting the MAC frame onto N3.
36Traffic Generated by Various Multicasting Strategies Table 11.2summarizes the number of packets generated on the various links and networks inorder to transmit one packet to a multicast group by this method. In this table, thesource is the multicast server on network N1 in Figure 11.5; the multicast addressincludes the group members on N3, N5, and N6. Each column in the table refers tothe path taken from the source host to a destination router attached to a particulardestination network. Each row of the table refers to a network or link in the configurationof Figure Each entry in the table gives the number of packets thattraverse a given network or link for a given path. A total of 13 copies of the packetare required for the broadcast technique.Now suppose the source system knows the location of each member of themulticast group. That is, the source has a table that maps a multicast address intoa list of networks that contain members of that multicast group. In that case, thesource need only send packets to those networks that contain members of the group.We could refer to this as the multiple unicast strategy. Table 11.2 shows that in thiscase, 11 packets are required.Both the broadcast and multiple unicast strategies are inefficient because theygenerate unnecessary copies of the source packet. In a true multicast strategy, thefollowing method is used:1. The least-cost path from the source to each network that includes members ofthe multicast group is determined. This results in a spanning tree of the configuration.The spanning tree is a set of all the networks that include multicastmembers, plus sufficient links between networks to establish a route from asource to all multicast members.2. The source transmits a single packet along the spanning tree.3. The packet is replicated by routers only at branch points of the spanning tree.
37Multicast Routing Protocols At the local level, individual hosts need a method of joining or leaving a multicast groupInternet Group Management Protocol (IGMP)Used between hosts and routers on a broadcast network such as Ethernet or a wireless LAN to exchange multicast group membership informationSupports two principal operations:Hosts send messages to routers to subscribe to and unsubscribe from a multicast group defined by a given multicast addressRouters periodically check which multicast groups are of interest to which hostsFor multicasting to work, the source of a multicast packet, together with Internetrouters, must identify networks that include hosts with the given multicast addressand determine a route that will reach all hosts in the group. For this purpose, anumber of address discovery and routing protocols are used at different levels of theInternet architecture.At the local level, individual hosts need a method ofjoining or leaving a multicast group. The host needs to be able to alert a routeron its local network of its membership status in a multicast group. On a broadcastnetwork, such as Ethernet or a wireless LAN, the Internet Group ManagementProtocol (IGMP) is used between hosts and routers to exchange multicastgroup membership information. IGMP takes advantage of the broadcastnature of a LAN to provide an efficient technique for the exchange of informationamong multiple hosts and routers. In general, IGMP supports two principaloperations:1. Hosts send messages to routers to subscribe to and unsubscribe from a multicastgroup defined by a given multicast address.2. Routers periodically check which multicast groups are of interest to whichhosts.
38Interior Routing Protocols Routers must cooperate across an organization’s internet or across the Internet to route and deliver multicast IP packetsRouters need to know which networks include members of a given multicast groupRouters need sufficient information to calculate the shortest path to each network containing group membersMulticast Extensions to OSPF(open shortest path first) (MOSPF)Enhancement to OSPF for the exchange of multicast routing informationProtocol Independent Multicast (PIM)Designed to extract needed routing information from any unicast routing protocol and may support routing protocols that operate across multiple ASs with a number of different unicast routing protocolsIGMP enables a router to know of hosts on anattached network that are using a particular multicast IP address. Next, routersmust cooperate across an organization’s internet or across the Internet to routeand deliver multicast IP packets. Routers must exchange two sorts of information.First, routers need to know which networks include members of a given multicastgroup. Second, routers need sufficient information to calculate the shortest path toeach network containing group members. These requirements imply the need for amulticast routing protocol.Within an AS, a number of alternative multicast routing protocols havebeen developed. We mention two here. Multicast Extensions to OSPF (MOSPF)is an enhancement to OSPF for the exchange of multicast routing information.Periodically, each router floods information about local group membership to allother routers in its AS. The result is that all routers in an AS are able to build upa complete picture of the location of all group members for each multicast group.Each router constructs the shortest-path spanning tree from a source network toall networks containing members of a multicast group.Protocol Independent Multicast (PIM) provides a more general solution tomulticast routing than MOSPF. As the name suggests, PIM is a separate routingprotocol, independent of any existing unicast routing protocol. PIM is designedto extract needed routing information from any unicast routing protocol and maysupport routing protocols that operate across multiple ASs with a number ofdifferent unicast routing protocols.
39Emergence of High-Speed LANs In recent years two significant trends altered the role of the personal computer and therefore the requirements on the LAN:The more powerful platforms of personal computers support graphics-intensive applications and ever more elaborate graphical user interfaces to the operating systemInformation technology (IT) organizations have recognized the LAN as a viable and essential computer platform, resulting in the focus on network computingTraditionally, office LANs provided basic connectivity services—connectingpersonal computers and terminals to mainframes and midrange systems that ran corporateapplications and providing workgroup connectivity at the departmental ordivisional level. In both cases, traffic patterns were relatively light, with an emphasison file transfer and electronic mail. The LANs that were available for this type ofworkload, primarily Ethernet and token ring, were well suited to this environment.In recent years, two significant trends altered the role of the personal computerand therefore the requirements on the LAN:1. The speed and computing power of personal computers continued to enjoyexplosive growth. These more powerful platforms support graphics-intensiveapplications and ever more elaborate graphical user interfaces to the operatingsystem.2. IT (information technology) organizations have recognized the LAN as aviable and essential computing platform, resulting in the focus on networkcomputing. This trend began with client/server computing, which has become adominant architecture in the business environment and the more recent Web-focusedintranet trend. Both of these approaches involve the frequent transferof potentially large volumes of data in a transaction-oriented environment.The effect of these trends has been to increase the volume of data to be handledover LANs and, because applications are more interactive, to reduce the acceptabledelay on data transfers. The earlier generation of 10-Mbps Ethernets and 16-Mbpstoken rings is simply not up to the job of supporting these requirements.
40The need for speed and QoS The Emergence of High-Speed LANsRole of PCs & requirements of LANs in need for High-speed:More powerful PCs, graphical applications & GUI-MIS Recognition of LAN as a viable computing platform, -C/S computing in business, -Graphics in transaction, -interactive applications on the Internet, -need to reduce the acceptable delay on data transfer creating large volume of data to be handled over LANs. So that 10Mbps Ethernets and 16 Mbps token rings are not adequate for High-speed LANs.Effect has been to increase volume of traffic over LANs:Examples of requirements calling for high speed LANCentralized server farm (e.g. color publishing operation)Power workgroup (e.g. software developers, CAD users transferring huge files across the Internet to share with piers.)High-speed local backbone (i.e. interconnection of these LANs)Convergence and unified communications (voice/video, and collaborative applications have increased the LAN traffic)
41The need for speed and QoS Corporate Wide Area NetworkingGreater dispersal of employee baseChanging application structuresIncreased client/server and intranetWide deployment of GUIsDependence on Internet accessMore data must be transported off premises and into the wide areaDigital ElectronicsMajor contributors to increased image and video trafficDigital Versatile Disc (DVD)Increased storage means more information to transmitDigital Still CameraCamcordersStill Image Cameras
42Quality of Service (QoS) Real-time voice and video don’t work well under the Internet’s “best effort” delivery serviceBest effort?fair delivery service, internet treats all packets equally. During congestion packet delivery slows down. In severe congestions, packets are dropped at random to ease congestion. No distinction is made in terms of the relative importance or timeliness of traffic/packets. (ATM)-”Asynchronous Transfer Mode”, a packet switching with fix size cells of 53 octetQoS provides for varying application needs in Internet transmission
43Categories of Traffic Elastic Inelastic Can adjust to changes in delay and throughput accessExamples: File transfer, , web accessInelasticDoes not adapt well, if at all, to changesExamples: Real-time voice, audio and video
44Inelastic Traffic Requirements ThroughputRequires a firm minimum value for throughputDelayresult in acting late to disadvantage (e.g. stock trading)Delay VariationRT applications (e.g. teleconferencing) require an upper bound. As the allowable delay gets larger, real delay in delivering the data gets longer and a larger delay buffer is required at the receiversPacket lossRT applications can sustain packet loss with varying amountThese requirements are difficult to meet in an environment with variable queuing delay and congestion losses.
45Requirements of Inelastic Applications 1. Application need to state their requirements either:In advance by service requeston the fly by means of fields in the IPThe 1st approach is preferred because the network can anticipate demands and deny new requests if the resources are limited.2. During congestion, elastic traffic need still be supported by:introducing a reservation protocol to deny service requests that would leave too few resources available to handle current elastic traffic
46A Comparison of Application Delay Sensitivity and Criticality in an Enterprise Sensitivity ==> demand Qos to provide TIMELY and HIGH data rateCriticality ==> QoS to provide RELIABILITY
47Differentiated Services (DS) Key characteristics:No change is required to IPExisting applications need not be modified to use DSProvides a built-in aggregation mechanism – all traffic with the same DS octet is treated the same by the network serviceRouters deal with each packet individually and do not have to save state information on packet flowsProvide QoS on the basis of the needs of different groups of usersMost widely accepted QoS mechanism in enterprise networksAs the burden on the Internet grows, and as the variety of applications grows, thereis an immediate need to provide differing levels of QoS to different users. Thedifferentiated services (DS) architecture is designed to provide a simple, easy-to-implement,low-overhead tool to support a range of network services that are differentiatedon the basis of performance. In essence, differentiated services do notprovide QoS on the basis of flows but rather on the basis of the needs of differentgroups of users. This means that all the traffic on the Internet is split into groupswith different QoS requirements and that routers recognize different groups on thebasis of a label in the IP header.Several key characteristics of DS contribute to its efficiency and ease ofdeployment:• IP packets are labeled for differing QoS treatment using the 6-bit DS field inthe IPv4 and IPv6 headers (Figure 8.7). No change is required to IP.• A service level agreement (SLA) is established between the service provider(internet domain) and the customer prior to the use of DS. This avoids theneed to incorporate DS mechanisms in applications. Thus, existing applicationsneed not be modified to use DS.• DS provides a built-in aggregation mechanism. All traffic with the same DSoctet is treated the same by the network service. For example, multiple voiceconnections are not handled individually but in the aggregate. This providesfor good scaling to larger networks and traffic loads.• DS is implemented in individual routers by queuing and forwarding packetsbased on the DS octet. Routers deal with each packet individually and do nothave to save state information on packet flows.Today, DS is the most widely accepted QoS mechanism in enterprise networks.
48ServicesA DS framework document lists all the following detailed performance parameters that might be included in an SLAService performance parameters, such as expected throughput, drop probability, and latencyConstraints on the ingress and egress points at which the service is provided, indicating the scope of the serviceTraffic profiles that must be adhered to for the requested service to be providedDisposition of traffic submitted in excess of the specified profileThe DS type of service is provided within a DS domain, which is defined as acontiguous portion of the Internet over which a consistent set of DS policies areadministered. Typically, a DS domain would be under the control of one administrativeentity. The services provided across a DS domain are defined in a service levelagreement, which is a service contract between a customer and the service providerthat specifies the forwarding service that the customer should receive for variousclasses of packets. A customer may be a user organization or another DS domain.Once the SLA is established, the customer submits packets with the DS octet markedto indicate the packet class. The service provider must assure that the customer getsat least the agreed QoS for each packet class. To provide that QoS, the service providermust configure the appropriate forwarding policies at each router (based onDS octet value) and must measure the performance being provided to each class onan ongoing basis.If a customer submits packets intended for destinations within the DS domain,then the DS domain is expected to provide the agreed service. If the destination isbeyond the customer’s DS domain, then the DS domain will attempt to forward thepackets through other domains, requesting the most appropriate service to matchthe requested service.A DS framework document lists the following detailed performance parametersthat might be included in an SLA:• Service performance parameters, such as expected throughput, drop probability,and latency• Constraints on the ingress and egress points at which the service is provided,indicating the scope of the service• Traffic profiles that must be adhered to for the requested service to be provided• Disposition of traffic submitted in excess of the specified profile
49DS Services ProvidedTraffic offered at service level A will be delivered with low latencyTraffic offered at service level B will be delivered with low loss90% of in-profile traffic delivered at service level C will experience no more than 50 ms latency95% of in-profile traffic delivered at service level D will be deliveredTraffic offered at service level E will be allotted twice the bandwidth of traffic delivered at service level FTraffic with drop precedence X has a higher probability of delivery than traffic with drop precedence YThe framework document also gives some examples of services that might beprovided:1. Traffic offered at service level A will be delivered with low latency.2. Traffic offered at service level B will be delivered with low loss.3. Ninety percent of in-profile traffic delivered at service level C will experienceno more than 50 ms latency.4. Ninety-five percent of in-profile traffic delivered at service level D will bedelivered.5. Traffic offered at service level E will be allotted twice the bandwidth of trafficdelivered at service level F.6. Traffic with drop precedence X has a higher probability of delivery than trafficwith drop precedence Y.The first two examples are qualitative and are valid only in comparison to othertraffic, such as default traffic that gets a best-effort service. The next two examplesare quantitative and provide a specific guarantee that can be verified by measurementon the actual service without comparison to any other services offered at thesame time. The final two examples are a mixture of quantitative and qualitative.
50Value of field is “codepoint” Packets are labeled for handling in 6-bit DS field in the IPv4 header, or the IPv6 headerValue of field is “codepoint”6-bits allows 64 codepoints in 3 poolsForm xxxxx0 - reserved for assignment as standardsForm xxxx11 - reserved for experimental or local useForm xxxx01 - also reserved for experimental or local use, but may be allocated for future standards action as neededPrecedence subfield indicates urgencyRoute selection, Network service, Queuing disciplineRFC 1812 provides two categories of recommendations for queuing disciplineQueue ServiceCongestion ControlDS FieldPackets are labeled for service handling by means of the 6-bit DS field in the IPv4header or the IPv6 header (Figure 8.7). The value of the DS field, referred to as theDS codepoint , is the label used to classify packets for differentiated services.With a 6-bit codepoint, there are, in principle, 64 different classes of trafficthat could be defined. These 64 codepoints are allocated across three pools ofcodepoints, as follows:• Codepoints of the form xxxxx0, where x is either 0 or 1, are reserved forassignment as standards.• Codepoints of the form xxxx11 are reserved for experimental or local use.• Codepoints of the form xxxx01 are also reserved for experimental or local usebut may be allocated for future standards action as needed.
51Differentiated Services (DS) Functionality in the internet and private internets to support specific QoS requirements for a group of users, all of whom use the same service label in IP packets.All the traffic on the Internet is split into groups with different QoS requirements and that routers recognize different groups on the basis of a label in the IP header.In DSTraffic on the internet is split into groups with different QoS requirementsRouters recognize different groups based on the label in the IP heads.IPv4 or IPv6 uses Type of Service
52Differentiated Services (DS)-Cont. Provides QoS based on “user group needs” rather than traffic flowsKey characteristics of DS:Differing QoS are labeled using the “6-bit DS field” in the IPv4 and IPv6 headersService-Level Agreements (SLA) govern DS, eliminating need for application-based assignmentDS provides a built-in aggregation mechanism. All traffic with the same DS octet is treated the same by the network serviceDS is implemented in individual router by queuing and forwarding packets based on the DS octet
53Allows the user to guide IP and router. Ipv4 HeaderAllows the user to guide IP and router.This field was not used until recent introduction of Differentiated ServicesType ofService Field
54Ipv4 Type of Service Field Explicit congestion notification fieldDifferentiated service fieldIpv4 Type of Service FieldThe DS type of service is provided within a DS domain, which is defined as a contiguous portion of the Internet over which a consistent set of DS policies are administered. Typically, a DS domain would be under the control of one administrative entity. The services provided across a DS domain are defined in a service-level agreement (SLA), which is a service contract between a customer and the service provider that specifies the forwarding service that the customer should receive for various classes of packets. A customer may be a user organization or another DS domain. Once the SLA is established, the customer submits packets with the DS octet marked to indicate the packets class. The service provider must assure that the customer gets at least the agreed QoS for each packet class. To provide that QoS, the service provider must configure the appropriate forwarding policies at each router (based on DS octet value) and must mearure the performance being provided to each class on an ongoing basis.If a customer submits packets intended for destinations within the DS domain, then the DS domain is expected to provide the agreed service. If the destination is beyond the customer’s DS domain, then the DS domain will attempt to forward the packets through other domain, requesting the most appropriate service to match the requested service.DS/ECN (8 bits): Prior to the introduction of differentiated services, this field was referred to as the Type of Service field and specified reliability, precedence, delay, and throughput parameters. This interpretation has now been superseded.The first 6 bits of the TOS field are now referred to as the DS (differentiated services) field.The remaining 2 bits are reserved for an ECN (explicit congestion notification) field.
55DS Framework DocumentA DS framework document lists the following detailed performance parameters that might be included in an SLA:Service performance parameters (e.g. expected throughput, drop probability, and latency)Constraints on the ingress (right to enter) and egress (right of going out) points at which the service is provided, indicating the scope of the serviceTraffic profiles that must be adhered to for the requested service to be provided, such as token bucket parametersDisposition of traffic submitted in excess of the specified profile
56DS Framework DocumentThe framework document also gives some examples of services that might be provided:Qualitative Examples:Traffic offered at service level A will be delivered with low latencyTraffic offered at service level B will be delivered with low lossQuantitative Examples:90% of in-profile traffic delivered at service level C will experience no more than 50 ms latency95% of in-profile traffic delivered at service level D will be delivered.Mixed Qualitative and Quantitative Examples:Traffic offered at service level E will be allotted twice the bandwidth of traffic delivered at service level FTraffic with drop precedence X has a higher probability of delivery than traffic with drop precedence Y
57DS OctetPackets are labeled for service handling by means of the DS octet, which is placed in the Type of Service field of an IPv4 header or the Traffic Class field of IPv6 header.IP Header
58DS Octet IPv4 Type of Service Field Packets are labeled for service handling by means of the DS octet, which is placed in the Type of Service field of an IPv4 header or the Traffic Class field of IPv6 header.IP Header
59DS FieldDS Field: Packets are labeled for service handling by means of the 6-bit DS field, in the IPv4 or IPv6. The value of the DS field, referred to as the DS codepoint, is the label used to classify packet for differentiated services.IP Header
60DS Field 6 bit DS field is used to label packets for service handling. The value of the DS field is referred to as the DS codepoint.6 bits provide 64 (i.e. 26 = 64) classes of traffic.6 bit code point is divided into 3 categories.
61DS Field/DS Octet Format Request For Comments 2474 defines the DS octet as having the following format:The left most 6 bits form a DS codepoint and the rightmost 2 bits are currently unused.The DS codepoint is the DS label used to classify packets for differentiated services.With a 6-bit codepoint, there are, in principle, 64 different classes of traffic that could be defined.These 64 codepoints are allocated across 3 pools (categories) of codepoints, as follows:
62DS Octet Format (x is either 0 or 1) 1. StandardDefault Packet Class (best-effort forwarding)Backward Compatibility(or equivalent) with the IPv4 precedence service2. Experimental/Local Use3. Experimental/Local Useor Future Standards
63DS Octet Format (x is either 0 or 1) 1. StandardDefault Packet Class (best-effort forwarding), in order they are received, and as soon as link capacity becomes available.2. Experimental/Local Use3. Experimental/Local Useor Future Standards
64DS Fieldxxx 000 Backward Compatibility (or equivalent) with the IPv4 precedence service.To explain the requirement of Codepoints, precedence field of IPV4 should be described.The original IPv4 includes “type of service” field which has two subfields:a 3-bit precedence subfield, anda 4-bit TOSThese subfields serve complementary functions:The precedence subfield provides guidance about the relative allocation of router resources for the datagram.TOS provides guidance to the IP entity in the source or router on selecting the next hop for each datagram.
65What is Precedence Field? Precedence field is set to indicate the degree of urgency or priority to be associated with a datagram. If a router supports the precedence subfield, there are 3 approaches to responding:Route selection: A particular route may be selected if the router has a smaller queue for that route or if the next hop on that route supports network precedence or priority (e.g. a token ring network supports priority).Network service: If the network on the next hop supports precedence, then that service is invokedQueuing discipline: A router may use precedence to affect how queues are handled. For example a router may give preferential treatment in queues to datagrams with higher precedence.
66Request For Comments 1812RFC 1812 ( Requirementes for IPV4) provides recommendations for queuing discipline that falls into 2 categories.Queue ServiceCongestion Control• Queue service(a) Routers SHOULD implement precedence-ordered queue service. Precedence-ordered queue service means that when a packet is selected for out- put on a (logical) link, the packet of highest precedence that has been queued for that link is sent.Any router MAY implement other policy-based throughput management procedures that result in other than strict precedence ordering, but it MUST be configurable to suppress them (i.e., use strict ordering).• Congestion control. When a router receives a packet beyond its storage capacity, it must discard it or some other packet or packets.(a) A router MAY discard the packet it has just received; this is the simplest but not the best policy.(b) Ideally, the router should select a packet from one of the sessions most heavily abusing the link, given that the applicable OoS policy permits this. A recommended policy in datagrarn environments using FIFO queues is to discard a packet randomly selected from the queue. An equivalent algorithm in routers using fair queues is to discard from the longest queue. A router MAY use these algorithms to determine which packet to discard.(c) If precedence-ordered queue service is implemented and enabled, the router MUST NOT discard a packet whose IP precedence is higher than that of a packet that is not discarded.(d) A router MAY protect packets whose IP headers request the maximize reliability TOS, except where doing so would be in violation of the previous rule.(e) A router MAY protect fragmented IP packets, on the theory that dropping a fragment of a datagram may increase congestion by causing all fragments of the datagrarn to be retransmitted by the source.(f) To help prevent routing perturbations or disruption of management func- tions, the router MAY protect packets used for routing control, link control, or network management from being discarded. Dedicated routers (i.e., routers that are not also general purpose hosts, terminal servers, etc.) can achieve an approximation of this rule by protecting packets whose source or destination is the router itself.The DS codepoints of the form xxxOOO should provide a service that at rninimum is equivalent to that of the lPv4 precedence functionality.
67DS Configuration & Operation A DS domain consists of a set of contiguous routers, that is, it is possible to get from any router in the domain to any other router in the domain by a path that does not include routers outside the domain. Within a domain interpretation of DS codepoints is uniform, so that a uniform, consistent service is provided.
68DS Configuration & Operation Figure 11.6 illustrates the type of configuration envisioned in the DS documents. ADS domain consists of a set of contiguous routers; that is, it is possible to get fromany router in the domain to any other router in the domain by a path that doesnot include routers outside the domain. Within a domain, the interpretation of DScodepoints is uniform, so that a uniform, consistent service is provided.
69DS Configuration & Operation In a DS domainRouters are either boundary nodes or interior nodesInterior nodes use per-hop behavior (PHB) rulesRouters in a DS domain are either boundary nodes or interior nodes. Typically the interior nodes implement simple mechanisms for handling packets based on their DS codepoint values. This includes a queuing discipline to give preferential treatment depending on codepoint value, and packet-dropping rules to dictate which packets should be dripped first in the event of buffer saturation. The DS specifications refer to the forwarding treatment provided at a router as per-hop behaviour (PHR). This PHB must be available at all routers, and typically PHB is the only part of DS implementation in interior routers.
70DS Configuration & Operation The boundary nodes include PHB mechanisms but also more sophisticated traffic conditioning mechanisms required to provide the desired service. Thus interior routers have minimal functionality and minimal overhead in providing the DS service, while most of the complexity is in the boundary nodes. The boundary node function can also be provided by a host system attached to the domain, on behalf of the applications at that host system.Routers in a DS domain are either boundary nodes or interior nodes. Typically the interior nodes implement simple mechanisms for handling packets based on their DS codepoint values. This includes a queuing discipline to give preferential treatment depending on codepoint value, and packet-dropping rules to dictate which packets should be dripped first in thed event of buffer saturaion. The DS specifications refer to the forwarding treatment provided at a router as per-hop behaviour (PHR). This PHB must be available at all routers, and typically PHB is the only part of DS implementation in interior routers.
71Elements of Traffic Conditioning Functions Boundary nodes have PHB (per-hop behavior) & traffic conditioning.The traffic conditioning function consists of five elements:Classifier: Classifies based on DS codepointsMeter: Measures that the packet traffic meets packet class or exceedsMarker: re-marking packets that exceed the profile for the best-effortShaper: Delaying packet stream as necessary.Dropper: Drops packets if the rate of packets exceeds profile specification.
72Traffic Conditioning Function Elements: ClassifierSeparates submitted packets into different classesMeterMeasures submitted traffic for conformance to a profileMarkerRe-marks packets with a different codepoint as neededShaperDelays packets as necessary so that the packet stream in a given class does not exceed the traffic rate specified in the profile for that classDropperDrops packets when the rate of packets of a given class exceeds that specified in the profile for that classThe traffic conditioning function consists of five elements:• Classifier: Separates submitted packets into different classes. This is thefoundation of providing differentiated services. A classifier may separatetraffic only on the basis of the DS codepoint (behavior aggregate classifier)or based on multiple fields within the packet header or even the packetpayload (multifield classifier).• Meter: Measures submitted traffic for conformance to a profile. The meterdetermines whether a given packet stream class is within or exceeds theservice level guaranteed for that class.• Marker: Re-marks packets with a different codepoint as needed. This may bedone for packets that exceed the profile; for example, if a given throughput isguaranteed for a particular service class, any packets in that class that exceedthe throughput in some defined time interval may be re-marked for best-efforthandling. Also, re-marking may be required at the boundary between two DSdomains. For example, if a given traffic class is to receive the highest supportedpriority, and this is a value of 3 in one domain and 7 in the next domain,then packets with a priority 3 value traversing the first domain are re-markedas priority 7 when entering the second domain.• Shaper: Delays packets as necessary so that the packet stream in a given classdoes not exceed the traffic rate specified in the profile for that class.• Dropper: Drops packets when the rate of packets of a given class exceeds thatspecified in the profile for that class.
73Relationships Between the Elements of Traffic Conditioning After a flow is classified, its resource consumption must be measured. The metering function measures the volume of packets over a particular time interval to determine a flow’s compliance with the traffic agreement. If the host is bursty, a simple data rate or packet rate may not be sufficient to capture the desired traffic characteristics.A token bucket scheme is an example of a way to define a traffic profile to take into account both packet rate and burstiness.
74Traffic Conditioning Diagram Figure 11.7 illustrates the relationship between the elements of trafficconditioning. After a flow is classified, its resource consumption must be measured.The metering function measures the volume of packets over a particular timeinterval to determine a flow’s compliance with the traffic agreement.If a traffic flow exceeds some profile, several approaches can be taken.Individual packets in excess of the profile may be re-marked for lower-qualityhandling and allowed to pass into the DS domain. A traffic shaper may absorba burst of packets in a buffer and pace the packets over a longer period of time.A dropper may drop packets if the buffer used for pacing becomes saturated.
75Service Level Agreements (SLA) Contract between the network provider and a customer that defines specific aspects of the service to be providedTypically includes:A description of the nature of service to be providedExpected performance level of the serviceProcess for monitoring and reporting the service levelA service level agreement (SLA) is a contract between a network provider and acustomer that defines specific aspects of the service that is to be provided. The definitionis formal and typically defines quantitative thresholds that must be met. AnSLA typically includes the following information:• A description of the nature of service to be provided: A basic service wouldbe IP-based network connectivity of enterprise locations plus access to theInternet. The service may include additional functions such as Web hosting,maintenance of domain name servers, and operation and maintenance tasks.• The expected performance level of the service: The SLA defines a number ofmetrics, such as delay, reliability, and availability, with numerical thresholds.• The process for monitoring and reporting the service level: This describes howperformance levels are measured and reported.
76Typical Framework for SLA Figure 11.8 shows a typical configuration that lends itself to an SLA. In thiscase, a network service provider maintains an IP-based network. A customer has anumber of private networks (e.g., LANs) at various sites. Customer networks areconnected to the provider via access routers at the access points. The SLA dictatesservice and performance levels for traffic between access routers across the providernetwork. In addition, the provider network links to the Internet and thus providesInternet access for the enterprise.An SLA can be defined for the overall network service. In addition, SLAs canbe defined for specific end-to-end services available across the carrier’s network,such as a virtual private network, or differentiated services.
78Table 11.3 (a) Sampled Metrics Table 11.3 lists the metrics that have been defined in RFCs at the time of thiswriting. Table 11.3a lists those metrics which result in a value estimated based on asampling technique. The metrics are defined in three stages:• Singleton metric: The most elementary, or atomic, quantity that can be measuredfor a given performance metric. For example, for a delay metric, asingleton metric is the delay experienced by a single packet.• Sample metric: A collection of singleton measurements taken during a giventime period. For example, for a delay metric, a sample metric is the set ofdelay values for all of the measurements taken during a one-hour period.• Statistical metric: A value derived from a given sample metric by computingsome statistic of the values defined by the singleton metric on the sample.For example, the mean of all the one-way delay values on a sample might bedefined as a statistical metric.Src = IP address of a hostDst = IP address of a host
79Table 11.3(b) Other Metrics The measurement technique can be either active or passive. Active techniquesrequire injecting packets into the network for the sole purpose of measurement.There are several drawbacks to this approach. The load on the network is increased.This in turn can affect the desired result. For example, on a heavily loaded network,the injection of measurement packets can increase network delay, so that the measureddelay is greater than it would be without the measurement traffic. In addition,an active measurement policy can be abused for denial-of-service attacks disguisedas legitimate measurement activity. Passive techniques observe and extract metricsfrom existing traffic. This approach can expose the contents of Internet traffic tounintended recipients, creating security and privacy concerns. So far, the metricsdefined by the IPPM working group are all active.Table 11.3b lists two metrics that are not defined statistically. Connectivitydeals with the issue of whether a transport-level connection is maintained by thenetwork. The current specification (RFC 2678) does not detail specific sampleand statistical metrics but provides a framework within which such metrics couldbe defined. Connectivity is determined by the ability to deliver a packet across aconnection within a specified time limit. The other metric, bulk transfer capacity,is similarly specified (RFC 3148) without sample and statistical metrics but beginsto address the issue of measuring the transfer capacity of a network service withthe implementation of various congestion control mechanisms.
80Model for Defining Packet Delay Variation Figure 11.9 illustrates the packet delay variation metric. This metric is usedto measure jitter, or variability, in the delay of packets traversing the network. Thesingleton metric is defined by selecting two packet measurements and measuringthe difference in the two delays. The statistical measures make use of the absolutevalues of the delays.
81Summary Chapter 11: Internet Operation Quality of service Emergence of high-speed LANsCorporate WAN needsInternet trafficDifferentiated servicesDS fieldDS configuration and operationSLAsIP performance metricsInternet addressingIPv4 addressingIPv6 addressingInternet routing protocolsAutonomous systemsBorder gateway protocolOSPF protocolMulticastingMulticast transmissionMulticast routing protocolsChapter 11 summary.Chapter 11: Internet Operation
83Service Level Agreements (SLA) Contract between the network provider and customer that defines specific aspects of the service provided.Typically includes: -Service description -Expected performance level -Monitoring and reporting process-Service descriptionA basic service would be IP-based network connectivity of enterprise locations plus access to the Internet. The service may include additional functions such as Webhosting, maintenance of domain name servers, and operation and maintenance tasks-Expected performance levelThe SLA defines a number of metrics, such as, delay reliability, and availability, with numerical thresholds.-Monitoring and reporting processThis describes how performance levels are measured and reported.
84SLA Example MCI Internet Dedicated Service 100% availabilityAverage round trip transmissions of ≤ 45 ms with the U.S.Successful packet delivery rate (reliability) ≥ 99.5%Denial of Service response within 15 minutesJitter performance will not exceed 1 ms between access routers
85IP Performance Metrics Three Stages of Metric Definitions -Singleton -Sample -StatisticalActive techniques require injecting packets into the networkPassive techniques observe and extract metrics
87Token Bucket SchemeBucket represents a counter, indicating allowable number of octetsBucket fills with octet tokenR := average data rate supportedB := Bucket sizeTherefore,During any time period T:The amount of data sent < RT +BIf a traffic flow exceeds some profile, several approaches can be taken.Individual packets in excess of the profile may be re-marked for lower-quality handling and allowed to pass into the DS domain.A traffic shaper may absorb a burst of packets in a buffer and pace the packets over a longer period of time.A dropper may drop packets if the buffer used for pacing becomes saturated.R:=input rateM:=output rateT: Duration of the max-rate burstB+RT = MTT = B/(M-R) sec