Network Management Principles and Practice Mani Subramanian 2nd Edition Ch12

Chapter 12. Broadband Network Management: WAN > Broadband Network and Services

As new technologies emerge, service providers offer new services tocommercial and residential communities using those technologies. Inturn, offering of new services by service providers is propellinginformation technology to new heights. This is especially true inbroadband technology. Let us first define what broadband network andservices are, which we briefly introduced in Section 2.7.

The broadband network and the narrowband Integrated Services DigitalNetwork (ISDN) are multimedia networks that provide integrated analogand digital services over the same network. Narrowband ISDN islow-bandwidth network that can carry two 56 kilobaud rate channels.The broadband network can transport very high data rate signals. Thenarrowband ISDN is also known as Basic ISDN.

There are three types of information technology services: voice, video,and data. In the traditional terminology, voice and video services aretransported over the telecommunication network. The information maybe transported over telecommunication facilities in either an analog or adigital mode. The telecommunication network can be topologicallyseparated into a wide area network (WAN) and local loops. The formerserves transportation over long distance between switching offices, andthe latter covers the last mile from the switching office to thecustomer premises.

As we saw in Chapter 1, data services are transported over thecomputer network, which is made up of LANs and WANs. The switchesand multiplexers in the telecommunication network are replaced withbridges, routers, and gateways. The computer network uses thefacilities of the telecommunication network for transportation over WAN.

The broadband network has several interpretations. One of the chiefcharacteristics of broadband service is the integration of voice, video,and data services over the same transportation medium; in other words,it is multimedia transportation networking. Sometimes, the broadbandnetwork is confused with high-speed data network, either dedicated orcombined with real-time voice or video, especially in the data trafficarena. However, we limit our definition of the broadband network tothose that can handle multimedia service of voice, video, and data. The

Network Management: Principles and Practice > Review of Information ... file:///D:/TC Engg/8th Semester/TMN/Book Network Management Princ...

1 of 4 23-Sep-15 7:58 PM

broadband network is also called the Broadband Integrated ServicesDigital Network (B-ISDN).

The early form of the integrated services network is the Basic ISDN. Itconsists of two basic channels: B-channels, 56-kilobaud rate each,combined with an 8-kilobaud signaling channel, D-channel. Together,they are referred to as 2B+D. As we stated earlier, it is also callednarrowband ISDN. However, on-line video requires a much largerbandwidth. Besides, voice and video require low latency and latencyfluctuations, which are achieved by ATM technology. These necessitieshave led to the early implementation of broadband ISDN, moresuccinctly referred to as the broadband network.

The broadband network and service have contributed significantly toadvances in three network segments of WAN, access network, andhome/customer premises equipment (CPE) network. In the WANsegment, protocols used in addition to IP are the asynchronous transfermode (ATM), the Synchronous Optical Network (SONET)/theSynchronous Digital Hierarchy (SDH), and Multiprotocol Label Switching(MPLS). The ATM technology can be viewed as a hybrid of circuit- andpacket-switched transmission modes. As a switch, the ATM switchmakes a physical connection of a virtual circuit. However, the data aretransmitted as cells (or packets) unlike in a circuit-switched connection.In Section 2.6 we introduced ATM using cell technology. The data rateof SONET/SDH WAN is an integral multiple of basic OC-1 (OpticalCarrier)-1/STS (Synchronous Transport Signal), which is 51.84 Mbps.MPLS, as mentioned in Section 2.6, evolved as the broadband protocoland takes advantage of the high performance of ATM and the richness infeatures of IP and Ethernet.

Broadband access technology is implemented using one of fivetechnologies. Hybrid fiber coax (HFC) or cable modem technology is atwo-way interactive multimedia communication system using fiber andcoaxial cable facilities and cable modems. The second technology uses adigital subscriber line (DSL). There are several variations ofimplementing this, generically referred to as xDSL. For example, ADSLstands for Asymmetric DSL. We will learn about the various DSLtechnologies in Chapter 13. The third and fourth technologies usewireless transmission from the switching office or the head end to thecustomer premises. Transmission in the two cases is either terrestrial orvia a satellite. The fifth technology is the mobile wireless technology.Mobile wireless technology is deployed as either access technology usingGSM (Global System for Mobile Communications)/GPRS (General PacketRadio Service) or CDMA (Code Division Multiple Access), or as a


2 of 4 23-Sep-15 7:58 PM

home/CPE network using WiFi (IEEE 802.3) protocol. We will discussthese access technologies in Chapters 1315.

Figure 2.30, which is duplicated here as Figure 12.1 shows abroadband network. The WAN is MPLS/IP/ATM. The WAN is linked to thecustomer premises using either optical links, OC-n (Optical Carrier-n)/STS (Synchronous Transport Signal), or a broadband link withemerging access technology (HFC, xDSL, or wireless). The customernetwork consists of two classes, residential customers and corporatecustomers with campus-like network. The residential customers areeither residential homes or small corporations that use broadbandservices, but do not require the high-speed access network to WAN.Corporate customers need high-speed access and connect optical orsynchronous (E1/T1) links.

Figure 12.1. Broadband Service Networks

Radio, video (television), Internet Service Provider (ISP), and otherservice providers constitute the service providers. Multiple services aremultiplexed at the central office or the Multiple Service Operator (MSO)head end and are piped to the customer premises via common facilities.The service providers interface with WAN via gateways.

The management of a broadband network is more complex than that ofeither the conventional computer network, which is mostly based on IP,


3 of 4 23-Sep-15 7:58 PM

Related Content

or the telecommunication network. It is based on MPLS and ATM WANtechnology and broadband access technology. An ATM network is basedon switches with point-to-point connections (in contrast to one-to-manyconnections as in broadcast/multicast protocols). It is also a connection-oriented protocol and needs to be integrated in the connectionlessInternet environment. These provide challenges to the management ofan ATM network. The ATM network is slowly migrating to the MPLSnetwork. However, it is still in extensive use and hence is included hereas a broadband WAN. The MPLS protocol can be used with either ATM orIP; and the management requirements and the managementinformation bases (MIBs) are still evolving. We will discuss ATMtechnology in Section 12.2 and ATM network management in Section12.3.


4 of 4 23-Sep-15 7:58 PM

Chapter 12. Broadband Network Management: WAN > ATM Technology

The ATM has helped bring about the merger of computer andtelecommunication networks. There are five important conceptscomprising ATM technology [Keshav, 1997]. They are (1) virtualpathvirtual circuit (VPVC), (2) fixed packet size or cells, (3) smallpacket size, (4) statistical multiplexing, and (5) integrated services. Theimplementation of these concepts in a network that is made up of ATMswitches achieves high-speed network that can transport all threeservices (voice, video, and data). The desired quality of service isprovided to individual streams (unlike the current Internet) at the sametime. The network is also easily scaleable. The ATM Forum, anorganization that specifies standards for ATM implementation, has alsoprovided a framework for network management, which we will addressin Section 12.3.

The subject of ATM is full of acronyms. Hence, we have included aspecial acronym table, Table 12.1 that contains those that we usehere. There are many more acronyms associated with ATM.

Table 12.1. ATM Acronyms

AAL ATM Adaptation Layer

ABR Available Bit Rate

AMC Agent Management Entity

ATM Asynchronous Transfer Mode

BICI Broadband Intercarrier Interface

BISDN Broadband Integrated ServicesDigital Network

BISSI Broadband Inter Switching SystemInterface

CBR Constant Bit Rate

DS3 Digital Signal 3

DXI Digital Exchange Interface

ELAN Emulated Local Area Network


1 of 6 23-Sep-15 7:58 PM

ILMI Interim/Integrated LocalManagement Interface

IME Interface Management Entity

LATA Local Access and Transport Area

LCD Loss of Cell Delineation

LE LAN Emulation

LE ARP LAN Emulation Address ResolutionProtocol

LOF Loss of Frame

LOP Loss of Pointer

LOS Loss of Signal

QoS Quality of Service

SDH Synchronous Digital Hierarchy

UBR Unspecified Bit Rate

UNI User-Network Interface

VBR-rt Variable Bit Ratereal time

VBR-nrt Variable Bit Ratenon-real time

VC Virtual Channel

VCC Virtual Channel Connection

VCI Virtual Channel Identifier

VCL Virtual Channel Link

VP Virtual Path

VPC Virtual Path Connection

VPI Virtual Path Identifier

VPL Virtual Path Link

We learned about the cell transmission mode in Chapter 2. As shown inFigure 2.28(c) and discussed in Section 2.6.4, it combines the best ofthe circuit- and packet-switched modes of transmission. The packets areall of the same size and are small in size. Each cell has the fullbandwidth of the medium, and the cells are statistically multiplexed. Thepackets all take the same path using the VPVC concept. This mode oftransmission is called the asynchronous transfer mode (ATM) and is one


2 of 6 23-Sep-15 7:58 PM

of the fundamental concepts of ATM technology. Let us now look atother aspects of ATM technology.

We discussed in Section 2.4.10 how the VPVC concept is used in packetswitches. Figures 2.27(a) and (b) in Chapter 2 show the distinctionbetween the datagram and the virtual circuit configuration. In ATMtechnology, the virtual circuit configuration is used. A virtual circuit thathas been established between two stations, A and Z, is shown in Figure12.2. The routing tables in the ATM switches, A, B, and D, associatedwith this virtual circuit are shown in Figure 12.2. The virtual circuit isfirst established prior to sending the data. In our example, the virtualcircuit is the combination of virtual circuit links VCI-1, VCI-2, VCI-3, andVCI-4. Once the virtual circuit is established, all packets are transportedin the sequence in which they were transmitted by the source along thesame path for a given session. Thus, packets 1, 2, and 3 transmitted byStation A arrive at Station Z in the correct sequence traversing thesame links. Since the path is fixed for the entire session, thetransmission rate is considerably increased just as in circuit-switchedTDM transmission (Table 12.2).

Figure 12.2. Virtual Circuit Configuration

Table 12.2. AZ Virtual Circuit-RoutingTables

Switch Input VCI/Port OutputVCI/Port

A VCI-1/Port-1 VCI-2/Port-2

VCI-2/Port-2 VCI-1/Port-1

B VCI-2/Port-1 VCI-3/Port-2



3 of 6 23-Sep-15 7:58 PM

Switch Input VCI/Port OutputVCI/Port

D VCI-3/Port-1 VCI-4/Port-2


Although there is enhanced speed of transmission of packets, there isdelay associated with pre-establishing the links for the virtual circuit.This delay is reduced by pre-assigning links to a virtual circuit. This isdone by grouping a number of virtual circuits between two switches intoa virtual path. A virtual path identifier (VPI) comprises virtual circuitidentifiers (VCIs). Thus, establishing the route from Station A to StationZ in our example is to do a look-up of the VPIVCI tables. The price thatwe pay for adopting this mechanism is that some VCIs may remain idleduring non-busy traffic period, and thus waste the bandwidth. However,this wasted bandwidth is a lot less than that for dedicated physical linksin a circuit-switched transmission mode.

The VPVC can be established on a per session basis or on a permanentbasis between a pair of end stations that carry large volumes of traffic.In the former case, the circuit is established as and when needed andtorn down after the session is over. This is called the switched virtualcircuit (SVC). When the connection is established for long periods oftime and not switched between sessions, a permanent virtual circuit(PVC) is established.

ATM packets are of fixed size, each 53 bytes long. A fixed-size packetwas chosen so that fast and efficient switches can be built. Manyswitches can operate in parallel if they all perform switching on thesame-size packets.

The ATM packet size of 53 bytes has a header of 5 bytes and a payloadof 48 bytes. This size was arrived at by optimizing between two factors.The packet size should be as small as possible to reduce the delay inswitching and packetization. However, it should be large enough toreduce the overhead of the header relative to the payload.

The main challenge in integrating the three services is to meet thedifferent requirements of each. Voice and video traffic require lowtolerance on variations in delay and low end-to-end (roundtrip) delaysfor good interactive communication. Once voice data are lost or delayed,


4 of 6 23-Sep-15 7:58 PM

real-time communication is garbled and we cannot reproduce it. Thus, ithas to be given the highest priority of service in transmission. This istrue with the voice portion of the information in video transmission.Further, the voice and video have to be synchronized. Otherwise, it willbe like watching a movie with the conversation lagging behind themouth motion due to wrong threading of the film. Pure video withoutsound can have less priority than audio.

Data traffic can have a much higher tolerance on latency. It is primarilya store and forward technology and the traffic itself is inherently burstyin nature. However, data speed is important for large data transmissionapplications, although it has the lowest priority in transmission.

It is possible to set the priority in ATM switches by assigning priority tothe different services. This is accomplished by guaranteeing a quality ofservice (QoS) for each accepted call setup. A traffic descriptor isspecified by the user of the service; and the system ensures that theservice requested could be met by the virtual circuit that is set up.

There are four main classes of traffic defined to implement quality ofservice. They are the constant bit rate (CBR), the real-time variable bitrate (VBR-rt), the non-real-time variable bit rate (VBR-nrt), and theavailable bit rate (ABR). Voice is assigned CBR. Streaming video such asreal-time video on the Internet is assigned VBR-rt. The VBR-nrt isapplicable to transmission of still images. The IP data traffic gets thelowest bandwidth priority with ABR.

There are two markets for ATM switches using ATM technology, publicand private. A public network is the network that is established by theservice providers. A private network is primarily a campus network.Network management clearly distinguishes between these two markets,as we shall see in Section 12.3.

Although analog high-frequency multiplexing is still in vogue for WANtransportation in legacy systems, digital transmission is thepredominant mode of transportation. The basic voice band, 04 kHz, isconverted to 64 Kbps digital signal universally. However, multiplexinghierarchy of the basic signal has evolved differently in North America,Europe, and Japan. For example, T1 transmission carrier shown inFigure 2.28(a) has a data rate of 1.544 Mbps carrying 24 voicechannels. Equivalent of this in Europe is the E1 transmission at a datarate of 2.048 Mbps carrying 30 channels. Thus, whenever digitaltransmission happens across the pond between Europe and NorthAmerica, there is an expensive conversion involved between the two


5 of 6 23-Sep-15 7:58 PM

Related Content

types of systems.

The digital hierarchy has been brought into synchronization across theworld using 155.52 Mbps as the basic data rate in carrier technologyusing fiber optics. However, the names are different in Europe and NorthAmerica. In Europe, it is called SDH and in North America SONET. Theunits of SDH are Synchronous Transport Signal (STS-n) where n is thehierarchical level. The optical carrier starts with the unit of OC-1 (Opticalcarrier-level 1), which is 51.84 Mbps and thus the basic SONET is levelOC-3.

It was once considered possible that ATM would be at all desktopworkstations. However, this has not been the case and IP over Ethernethas become the most used LAN.

The services provided by ATM differ from conventional LAN in threeways. First, ATM is connection oriented. Second, ATM makes one-to-oneconnection between pairs of workstations in contrast to the broadcastand multicast mode in conventional LAN. Third, a LAN MAC address isdedicated to the physical network interface card and is independent ofnetwork topology. The 20-byte ATM address is not.

In order to use ATM in the current LAN environment, it has to fit into thecurrent TCP/IP LAN environment. Because of the basic differencesmentioned in the previous paragraph, although the ATM Forum hasdeveloped ATM specifications for LAN emulation (LE or LANE) thatemulate services of the current LAN network across an ATM network, ithas been discontinued. Hence, we will not discuss this any further.


6 of 6 23-Sep-15 7:58 PM

Chapter 12. Broadband Network Management: WAN > ATM Network Management

Broadband network management consists of managing the WAN usingATM technology, as well as access networks from the central office to thehome. We will discuss the former in this section. We will discuss accesstechnology management in the next chapter.

WAN facilities are provided by public service providers, who perform thefollowing management functions: operation, administration, maintenance,and provisioning (OAMP). Typically, a large enterprise or corporationservices its private network. However, they too use the public serviceproviders facilities to transport information over a long distance. This isreferred to as public network. ATM networks are classified as private andpublic networks, as shown in Figure 12.3. The standards for themanagement of each and the interactions between them have beenaddressed by the ATM Forum, which is an international organizationaccelerating cooperation on ATM technology. The user interface to theprivate network is the private user-network interface (UNI), and theinterface to the public network is the Public UNI.

Figure 12.3. Private and Public ATM Network User NetworkInterfaces

The ATM Forum has defined a management interface architecture, ATMnetwork reference model, as shown in Figure 12.4. Private networks aremanaged by private network managers or private network managementsystems (NMSs). Public network managers or public network management


1 of 28 23-Sep-15 7:58 PM

systems manage the public networks. To distinguish between a humanmanager and the management system, we will refer to the networkmanager in the context of a system as the NMS, unless explicitly stated.There are five interfaces between systems and networks. M1 and M2 are,respectively, the interfaces between private NMS and either end user orprivate network. The end user can be a workstation, ATM switch, or anyATM device. A private ATM network is an enterprise network.

Figure 12.4. ATM Forum Management Interface ReferenceArchitecture

A private NMS can access its own network-related information in a publicnetwork via an M3 interface to the public NMS. The public NMS, whichmanages the public network, responds to the private NMS via the M3interface with the relevant information or takes the appropriate actionrequested.

M4 is the interface between a public NMS and a public network. M5 is theinterface between NMSs of two service providers. The ATM Forum has notyet specified this interface.

Beside the M-interfaces, Figure 12.4 also shows interfaces between anATM end user or device and an ATM network, as well as interfacesbetween ATM networks. These are distinct from the M-interfaces betweenNMSs and networks or end users. While the M-interfaces provide atop-down management view of network or device, the ATM Forum definesthe ATM link-specific view of configuration and fault parameters across aUNI. These are the UNI interfaces presented in Figures 12.3 and 12.4.The specifications for these are contained in the integrated localmanagement interface (ILMI), which we will discuss further in Section12.3.4.

The I in ILMI originally stood for Interim, not Integrated. See


2 of 28 23-Sep-15 7:58 PM

af.ilmi.0065.000. Its specifications were supposed to have been replacedby IETF specifications. However, it turned out that some were and otherswere not. Hence, I in ILMI now designates Integrated. The ILMI fits intothe overall model for an ATM device, as shown in Figure 12.5. The ATMmanagement information is communicated across the UNI or the network-to-network interface (NNI). These interfaces are with ATM devices(end-systems, switches, etc.) that belong to either a private or a publicnetwork. Any interface with the public ATM network is a public UNI or apublic NNI. Any interface with a private ATM network is a private UNI or aprivate NNI. The devices communicate across UNI and NNI via an ATMinterface management entity (IME) module in the entity. There are threeversions of IMEuser, network, and system based on where it is used.

Figure 12.5. Definitions and Context of ILMI

[View full size image]

Figure 12.5 shows the various physical connections, virtual connections,and the ILMI communication links between the devices and the networks.


3 of 28 23-Sep-15 7:58 PM

The ILMI communication occurs over both physical and virtual links usingSNMP or AAL5 protocols. We will discuss the MIB related to managementin the following sections.

Two public carrier networks interface with each other via a broadbandintercarrier interface (BICI), as shown in Figure 12.4. BICI is also knownas network-to-network interface (NNI).

The MIB and the structure of management of information that arerequired for the management of the ATM network are specified betweentwo sets of documents defined by IETF and the ATM Forum. The globalview of the Internet MIB tree associated with the ATM is presented inFigure 12.6. The two major branches are mib-2 and atmForum (underenterprises). The structure of management information is defined usingthe ASN.1 syntax. The MIB associated with the ATM is primarily concernedwith the ATM sublayer parameters. The parameters associated with higherlayers are handled using the standard MIB discussed earlier in Chapter 4.The documents that address the various groups are listed in Table 12.3.

Figure 12.6. Internet ATM MIB


Table 12.3. Internet ATM MIB Groups and Documents

Entity OID Description Document


4 of 28 23-Sep-15 7:58 PM


system mib-2 1 System RFC 1213

interfaces mib-2 2 Interfaces (modified) RFC 1213

RFC 1573

ifMIB mib-2 31 Interface types RFC 1573

transmission mib-2 10 Transmission RFC 1213

ds1 transmission18

DS1 carrier objects RFC 1406

ds3 transmission30

DS3/E#interface objects RFC 1407

sonetMIB transmission39

SONET MIB RFC 1595

atmMIB mib-2 37 ATM objects RFC 1695

atmForum enterprises353

ATM Forum MIB/M3specification

af-nm-0019.000

M4 interface af-nm-0020.000af-nm-0020.001

CMIP specification for M4interface

af-nm-0027.000

M4 network-view interface af-nm-0058.000

AAL management for theM4 NE View

af-nm-0071.000

Circuit emulation serviceinternet-workingrequirements, logical andCMIP MIB

af-nm-0072.000

M4 network-view CMIPMIB Spec v1.0

af-nm-0073.000

M4 network-viewrequirements and logicalMIB addendum

af-nm-0774-000

atmForumAdmin atmForum 1 ATM administrative af-ilmi-0065.000

atmForumUni atmForum 2 ATM user networkinterface

af-ilmi-0065.000


5 of 28 23-Sep-15 7:58 PM


atmUniDxi atmForum 3 Data exchange interface(DXI) specification

af-dxi-0014.000

Multiprotocol over ATM af-mpoa-0087.000

Multiprotocol over ATMVersion 1.0 MIB

af-mpoa-0092.000

There are five nodes shown under mib-2 in Figure 12.6. We describedsystem and interfaces groups in Chapter 4. the interfaces group hasevolved to handle the sublayers, such as ATM; the details are described inRFC 2863. The transmission group contains subgroups for each medium oftransmission. The ATM objects, as defined in the atmMIBObjects groupunder atmMIB, are specified in RFC 1695.

The atmForum group is subnode 353 under the enterprises node. TheatmForum group contains five subgroups, as shown in Figure 12.6:atmForumAdmin, atmForumUni, atmUniDxi, atmLanEmulation andatmForumNetworkManagement. The ATM administrative(atmForumAdmin) and ATM UNI (atmForumUni) groups are defined in theintegrated local management interface (ILMI) specification, af-ilmi-0065.000. The ATM DXI (atmUniDxi) group is the Data ExchangeInterface and is discussed in Section 12.3.9. It is the ATM interfacebetween DTE and DCE and is described in af-dxi-0014.000. The MIB forM4 interface (atmForumNetworkManagement) is covered in af-nm-0095.001.

Although ILMI was conceived as interim specifications, it has becomepermanent. The ATM network management uses both SNMP MIB and ATMForum MIB. Figures 12.7 and 12.8 (af-ilmi-0065.000 and Section 4 ofUNI) conceptually present the role of the two network managementprotocols. Figure 12.7 presents the M1 interface. An SNMP agent isshown embedded in an ATM device, and the NMS communicates with itusing SNMP protocol and IETF MIB modules. RFC 2863 specifies theinterface parameters and types, including the additional tables to managethe ATM sublayer. RFC 1695 specifies the ATM objects. The transport MIBmodule is dependent on the transmission medium.

Figure 12.7. SNMP ATM Management (M1 Interface)


6 of 28 23-Sep-15 7:58 PM

Figure 12.8. Role of SNMP and ILMI in ATM Management (M2Interface)


Figure 12.8, which shows the M2 interface, comprises the network of twoATM devices. The NMS manages the network with an interface to deviceA. The ILMI protocol is used for communication between the agentmanagement entity (AME) in device A and the AME in device B. A proxyagent that resides in device A does the translation between ILMI MIB andSNMP MIB.

The M1 interface, as mentioned earlier, is between an SNMP managementsystem and an SNMP agent in an ATM device, as shown in Figure 12.7.


7 of 28 23-Sep-15 7:58 PM

Four entities, ifInNUcastPkts, ifOutNUcastPkts, ifOutQLen, and ifSpecifichave been deprecated. The interfaces (interfaces) and ifMIB (IF MIB)groups under the mgmt node are shown in Figure 12.9. Four tables havebeen added to handle sublayers. They are shown in Figure 12.9 underifMIBObjects. Table 12.4 gives a brief description of the functions thateach table performs.

Figure 12.9. Interfaces Group Tables for Sublayers

Table 12.4. Interfaces Group Tables for Sublayers

Entity OID Description (Brief)

ifXTable ifMIBObjects 1 Additional objects for theinterface table

ifStackTable ifMIBObjects 2 Information on relationshipbetween sublayers

ifTestTable ifMIBObjects 3 Tests that NMS instructs theagent to perform

ifRecvAddressTable ifMIBObjects 4 Information on type ofpackets/frames accepted onan interface

Figure 12.10 shows the three transmission modes that are used for theATM. They are DS1 (1.544-Mbps twisted-pair cable), DS3 (44.736-Mbpscoaxial cable), and SONET (nx155.52-Mbps optical fiber). DS1 and DS3are transmitted over T1 and T3 carriers, respectively. Only one of theseMIBs needs to be implemented in the agent based on which transmission


8 of 28 23-Sep-15 7:58 PM

medium is used. RFCs dealing with transmission group MIB modules arelisted in Table 12.3.

Figure 12.10. Transmission Groups for ATM

Figure 12.11 and Table 12.5 show the ATM MIB objects group. Thisgroup contains information to manage the ATM sublayer entities: trafficdescriptors, DS3 physical-layer convergence parameters (PLCP),transmission convergence (TC) sublayer parameters, virtual pathlink/virtual channel link and their associated cross-connect tables, andperformance parameters for AAL5 (ATM adaptation layer).

Figure 12.11. ATM Managed Objects Group


Table 12.5. ATM Managed Objects Group


atmNoTrafficDescriptor atmMIBObjects 1 ATM traffic descriptortype

atmInterfaceConfTable atmMIBObjects 2 ATM local interfaceconfiguration parametertable


9 of 28 23-Sep-15 7:58 PM


atmInterfaceDs3PlcpEntry atmMIBObjects 3 ATM interface DS3 PLCPparameters and statevariables table

atmInterfaceTCTable atmMIBObjects 4 ATM TC sublayerconfiguration and stateparameters table

atmTrafficDescrParamTable atmMIBObjects 5 ATM traffic descriptortype and associatedparameters

atmVplTable atmMIBObjects 6 Virtual path link table

atmVclTable atmMIBObjects 7 Virtual channel linktable

atmVpCrossConnectNext atmMIBObjects 8 Index for virtual pathcross-connect table

atmVpCrossConnectTable atmMIBObjects 9 Virtual path cross-connect table

atmVcCrossConnectNext atmMIBObjects10

Index for virtualchannel cross-connecttable

atmVcCrossConnectTable atmMIBObjects11

Virtual cross-connecttable

aal5VccTable atmMIBObjects12

AAL VCC performanceparameters table

The M2 interface for ATM management is shown in Figure 12.8. Themanagement information on ATM links between devices is gathered fromILMI MIB. The relative roles of each are shown in Figure 12.8. DetailedUNI and NNI for both private and public interfaces, specified in af-ilmi-0065.000, are shown in Figure 12.5 and were discussed in Section12.3.2.

The ILMI specifications define the administrative and UNI groups of theATM Forum MIB. The administrative group defines a general-purposeregistry for locating ATM network services such as the ATM name answerserver (ANS). Other subgroups under the administrative group have beendeprecated and handled by IETF specifications.

Figure 12.12 and Table 12.6 show the ATM UNI MIB object group. They


10 of 28 23-Sep-15 7:58 PM

define the management objects associated with the ATM layer and thephysical layer. The statistics group is deprecated. The parametersassociated with virtual path/virtual connections as well as the adjustablebandwidth rate (ABR) and QoS are covered.

Figure 12.12. ATM UNI MIB Object Group


Table 12.6. ATM UNI MIB Object Group

Entity OID Description(Brief)

atmfPhysicalGroup atmForumUni1

Defines a table ofphysical-layer statusand parameterinformation

atmfAtmLayerGroup atmForumUni2

Defines a table ofATM-layer statusand parameterinformation

atmfAtmStatsGroup atmForumUni3

Deprecated

atmfVpcGroup atmForumUni4

Defines a table ofstatus andparameterinformation on thevirtual pathconnections


11 of 28 23-Sep-15 7:58 PM

Entity OID Description(Brief)

atmfVccGroup atmForumUni5

Defines a table ofstatus andparameterinformation on thevirtual channelconnections

atmfAddressGroup atmForumUni6

Defines thenetwork-side IMEtable containing theuser-side ATM-layeraddresses

atmfNetPrefixGroup atmForumUni7

Defines a user-sideIME table ofnetwork prefixes

atmfSrvcRegistryGroup atmForumUni8

Defines thenetwork-side IMEtable containing allservices available tothe user-side IME

atmfVpcAbrGroup atmForumUni9

Defines a table ofoperationalparameters relatedto ABR virtual pathconnections

atmfVccAbrGroup atmForumUni10

Defines a table ofoperationalparameters relatedto ABR virtualchannel connections

AtmfAddressRegistrationAdminGroup atmForumUni11

M3 is the management interface between the private NMS and the publicservice provider NMS. It allows the customer to monitor and configuretheir portion of the public ATM network. The M3 interface specificationsare defined in the ATM Forum document af-nm-0019.000. Networks showthe typical configuration, how a customer would interact with the publicservice provider network via the carrier management system. There aretwo classes of M3 requirements in the figurestatus and configuration


12 of 28 23-Sep-15 7:58 PM

monitoring (Class I) and virtual configuration control (Class II).

Class I requirements are those which a public network service provideroffers to the customer. These include the customer performing monitoringand management of configuration, fault, and performance management ofa specific customers portion of a public ATM network. This service isoffered only for PVC configuration. Examples of this service are (a)retrieving performance and configuration information for a UNI link and(b) public service NMS reporting an alarm or trap message to the userNMS on a UNI-link failure.

Class II service provides greater capability to the user. The user canrequest the service provider to add, delete, or change virtual connectionsbetween a pair of customers UNIs. An example of this would be thecustomer wanting to establish a new virtual path or increase the numberof virtual circuits in a given virtual path.

A customer network management (CNM) agent residing in the privateservice providers NMS provides the M3 service. As mentioned, the serviceis limited to the portion of the public service providers network that theusers circuit traverses. If the users circuit traverses multiple serviceproviders, a separate interface with each provider is needed. The CNMsends requests to the carrier management system (see Figure 12.13),which acts as an agent to the CNM. The carrier management system theninvokes the request on the network elements (NE) or other NMS andreturns the responses to CNM.

Figure 12.13. Customer Management of Private and PublicNetworks


13 of 28 23-Sep-15 7:58 PM

The requirements for M3 and M4 are specified as mandatory or required,conditionally required, and optional. Class I requirements are mandatory,and Class II requirements are optional.

Class I Interface Management Functions. Table 12.7 presents M3Class I requirements and the MIB groups that are used to obtain theinformation. The request has SNMP read-only capability. The publicnetwork service provider should give the CNM customer the ability toretrieve all the information listed in Table 12.7.

Table 12.7. M3 Class I Interface Requirements and MIB

General UNI protocolstack information

System group [RFC 1213], interfaces group,including ifTable and ifStackTable [RFC 1213,2863], SNMP group [RFC 1213]

ATM performanceinformation oncustomers UNI

ifTable [RFC 2863]

Physical-layerperformance and statusinformation

All tables except dsx3ConfigTable [RFC 1407],all tables except dsx1ConfigTable [RFC 1406],all tables except the configuration tables andVT tables of SONET MIB [RFC 1595],atmInterfaceDs3PlcpTable/atmInterfaceTCTable


14 of 28 23-Sep-15 7:58 PM

of ATM MIB [RFC 1695]

ATM-level informationconfiguration information

atmInterfaceConfTable of ATM MIB [RFC1695]

Physical-layerconfiguration information

dsx3ConfigTable [RFC 1407] dsx1ConfigTable[RFC 1406] all configuration tables except thesonetVtConfigTable of SONET MIB [RFC 1595]

ATM-layer virtual pathlink configuration andstatus information

atmVplTable of ATM MIB [RFC 1695]

ATM-layer virtual channellink configuration andstatus information

atmVclTable of ATM MIB [RFC 1695]

ATM-layer virtual pathconnection configurationand status information

atmVpCrossConnectTable andatmVpCrossConnectIndexNext of ATM MIB[RFC 1695]

ATM-layer virtual channelconnection configurationand status information

atmVcCrossConnectTable andatmVcCrossConnectIndexNext of ATM MIB[RFC 1695]

ATM-layer trafficcharacterization (trafficdescriptors for customersUNIs) information

atmTrafficDescrParamTable of ATM MIB [RFC1695]

Event notifications onATM link going up ordown

warmStart, coldStart, linkUp, linkDown ofSNMP group [RFC 1695]

Class II Interface Management Functions. M3 Class II functionality isdivided into three groups so that the provider can implement one or moresubgroups. They are (1) ATM-level subgroup, (2) VPC/VCC-levelsubgroup, and (3) traffic subgroup.

The ATM-level subgroup should provide the CNM the ability to modify theATM Level Information Configuration Information.

The VPC/VCC-level subgroup provides the CNM the ability to modify:

Virtual path link configuration and status information.1.

Virtual channel link configuration and status information.2.

Virtual path connection configuration and status information.3.

Virtual channel connection configuration and status information.4.

The traffic subgroup shall provide the CNM the ability to modify:


15 of 28 23-Sep-15 7:58 PM

Traffic descriptors and information objects for virtual channelconnections.

1.

Traffic descriptors and information objects for virtual pathconnections.

2.

Table 12.8 presents the M3 Class II requirements and the MIB objects inthe ATM MIB group.

Table 12.8. M3 Class II Interface Requirements and MIB

ATM-level information configurationinformation

atmInterfaceConfTable in ATM MIB[RFC 1695]

Virtual path link configuration andstatus information

atmVplTable in ATM MIB [RFC1695]

Virtual channel link configurationand status information

atmVclTable in ATM MIB [RFC1695]

Virtual path connectionconfiguration and status information

atmVpCrossConnectTable andatmVpCrossConnectIndexNext ofATM MIB [RFC 1695]

Virtual channel connectionconfiguration and status information

atmVcCrossConnectTable andatmVcCrossConnectIndexNext ofATM MIB [RFC 1695]

Traffic descriptors and informationobjects for virtual path and channelconnections

atmTrafficDescrParamTable in ATMMIB [RFC 1695]

The management of public ATM network is primarily the responsibility ofnetwork service providerscarriers and Postal Telephone and Telegraph(PTT) companies. They have the challenge of not only managing thepublic network, but also keeping up with new technology. To help thisprocess, ITU-T has defined M.3010, a five-layer model of operations,telecommunications management network (TMN), which we discussed indetail in Chapter 10. The relationship of ATM to TMN is shown in Figure12.14. The top two layers, the business management layer and theservice management layer, deal with the business and service aspects ofTMN and are not addressed by the ATM Forum.

Figure 12.14. ATM Relationship to TMN-Layered Architecture


16 of 28 23-Sep-15 7:58 PM

The element layer (EL) contains NEs. The NEs specific to ATM technologyare components, such as ATM workstation, ATM switches, ATM transportdevices (cross-connect systems and concentrators), etc. The elementmanagement layer (EML) manages NEs. The network management layer(NML) has the responsibility to manage the network either directly or viathe EML.

Figure 12.15 shows the dual view of M4 interface (af-nm-0058.000).Both views are present in the architecture across the M4 interface plane.It should be noted that this is a conceptual view, and the physicalconnections can be the same for both views.

Figure 12.15. Dual Views of the M4 Interface


17 of 28 23-Sep-15 7:58 PM

In the NE-level management architecture, shown in Figure 12.16, theNMS environment, consisting of one or more NMSs, directly interfaceswith the ATM NE and manages them. There is a single M4 interfacebetween ATM NE and the NMS environment. The figure shows linksbetween NE, but the NMS directly communicates with each ATM networkelement. In an actual implementation, it is likely that the NMS isinterfacing with another NMS, which is managing the NE, but can stillpresent a network view to the higher-level management system.

Figure 12.16. NE-View Management Architecture

Figure 12.17 presents an example of network-view managementphysical configuration. It consists of two ATM networks, one a single-supplier subnetwork and the other a multi-supplier subnetwork. Eachsubnetwork has its own subNMS managing its NE. In the single-supplier


18 of 28 23-Sep-15 7:58 PM

subnetwork shown on the right side, the subNMS has only an ATM NEview. In the multi-supplier subnetwork environment shown on the leftside, the subNMS is presented only an ATM NE view, although it mayactually be communicating to lower-level NMSs of each supplier. This issimilar to the manager-of-manager architecture we discussed in Chapter4. This is explicitly shown in the figure. Both subNMSs in Figure 12.16present an ATM network view only to the NMS environment shown at thetop. Thus, the top-level NMS sees only the network view of thesubnetworks and not the NE view.

Figure 12.17. Example of Network-View Management PhysicalConfiguration


The NMS environment can manage both NE and networks, as shown inthe architectural view in Figure 12.18. We can visualize such a hybridsituation in some remote ATM devices that do not need to be under agroups NMS, but are managed by an enterprise NMS.

Figure 12.18. Example of NE + Network-View Management



19 of 28 23-Sep-15 7:58 PM

M4 Network Element-View Requirements and Logical MIB. ATMForum M4 network-element-view specifications support PVCs. Based onthe OSI application model, the specifications are confined to configurationmanagement, fault management, and performance management. Basicsecurity features are also included to ensure the authorization andauthentication of the user and the protection and privacy of data, whileM3 interface responds to queries. The ATM security framework is similarto the security framework discussed for management in Chapter 7. It iscovered in af-sec-0096-000. The network management considerations arepresented in af-nm-0103.000, which includes security requirements andlogical MIB.

The MIB specifications are specified in a logical format in af-nm-0020.000and updated in af-nm-0020.001. They could be implemented using eitherCMIP or SNMP protocols. The CMIP specifications are presented on af-nm-0027.000, af-nm-0058.000, af-nm-0071.000, af-nm-0072.000, af-nm-0073.000, and af-nm-0074.000. The SNMP specifications are detailed inaf-nm-test-0080.000 and af-nm-0095.001. We will now summarize theprotocol-independent specifications of M4 interface and the logical MIB.

Configuration Management. Configuration management provides thefollowing list of functions to manage NEs:

ATM NE configuration identification and change reporting, whichinvolves:

Operations performed over the craft interfacea.

Human intervention (removal/insertion of equipmentmodules)

b.

Customer control channels (e.g., ILMI)c.

1.


20 of 28 23-Sep-15 7:58 PM

Network failuresd.

Protection switching eventse.

Sub-ATM NE component initializationf.

Secondary effects of atomic operations performed by themanagement system

g.

Configuration of UNIs, BICIs, and BISSIs.2.

Configuration of VPL/VCL termination points and cross-connections.

3.

Configuration of VPC and VCC OAM segment end-points.4.

Event flow controlevent forwarding discriminator function.5.

Fault Management. The following set of functions is specified to detect,isolate, and correct abnormal operation:

Notifying the NMS of a detected failure.1.

Logging failure reports.2.

Isolating faults via demand testing.3.

The specific functions are:

Failure reporting of the various alarms listed in Table 12.9. Thegeneric troubles that cause the alarm are also listed in the table.

Table 12.9. Generic Troubles in ATM NEs

Alarm Category Generic Trouble

Communication alarms Alarm indication signal (AIS)

Loss of cell delineation (LCD)

Loss of frame (LOF)

Loss of pointer (LOP)

Loss of signal (LOS)

Payload type mismatch

Transmission error

Path trace mismatch

1.


21 of 28 23-Sep-15 7:58 PM

Alarm Category Generic Trouble

Remote defect indication (RDI)

Signal label mismatch

Equipment alarms Back-plane failure

Cell establishment error

Congestion

External interface device problem

Line card problem

Multiplexer problem

Power problem

Processor problem

Protection path failure

Receiver failure

Replaceable unit missing

Replaceable unit problem

Replaceable unit type mismatch

Timing problem

Transmitter failure

Trunk card problem

Processing erroralarms

Storage capacity problem

Memory mismatch

Corrupt data

Software environment problem

Software download failure

Version mismatch

Environmental alarms Cooling fan failure

Enclosure door open

Fuse failure

High temperature

General Vendor specific


22 of 28 23-Sep-15 7:58 PM

Operations, administration, and maintenance (OAM) cell loopbacktesting.

2.

Performance Management. The functions of performance monitoringfor an ATM network are:

Performance monitoring.1.

Traffic management.2.

UPC (user parameter control)/NPC (network parameter control)disagreement monitoring.

3.

Performance management control.4.

Network data collection.5.

To accomplish these general functions, the following specific functions arespecified:

Physical-layer performance monitoring.1.

ATM cell-level protocol monitoring.2.

UPC/NPC disagreement monitoring.3.

Network-View Requirements and Logical MIB. The M4 network viewfor the management of an ATM public network is concerned with NMLinformation. It addresses the different perspectives of the serviceproviders, each of whom need both network management and servicemanagement capabilities.

The ATM Forum document af-nm-0058.000 details the requirements forthe ATM network management across the M4 network view interface. Theassociated MIB is specified in logical form and is not management protocolspecific. The functional areas addressed in the specifications are:

Transport network configuration provisioning (includingsubnetwork provisioning and link provisioning).

1.

Transport network connection management (including setup/reservation/ modification for subnetwork connection, linkconnection, trails, and segments).

2.

Network fault management (including congestion monitoring andconnection and segment monitoring).

3.

Network security management.4.

Managed entities have not been defined in the MIB for meeting all the


23 of 28 23-Sep-15 7:58 PM

above requirements.

Transport Network Provisioning/Layered Network Provisioning.The transport network provisioning includes subnetwork provisioning ofnetwork nodes and links. Specifically:

Subnetwork provisioning: addition and monitoring information onaddition, deletions, and changes in NEs and their configuration.

1.

Link provisioning: set up, modify, and release subnetwork links.2.

Subnetwork Connection Management. The M4 network-view managedentities support subnetwork management of reservation and modificationof subnetwork connections, link connections, trails, and segments.Specifically, this involves:

Point-to-point subnetwork connection: VPVC subnetworkconnection between pair of end points.

1.

Multipoint subnetwork connection: Multipoint VPVC connectionsbetween pair-wise end points.

2.

Link connection set-up: VPVC connections between subnetworks.3.

Segment setup: Set up and support VPVC segment terminationend points.

4.

Trail setup: Support and set up trails containing information onsubnetwork connections and links.

5.

When a part of the ATM trail spans multiple administrative domains, eachNMS is responsible for its domain setup and maintains its trails.

Connection Release. The connections release across the M4 interfaceinvolves the management of resources to be made available after use.Specifically, the network management should support:

Release subnetwork connections and release resources of bothpoint-to-point and multipoint connections.

1.

Release link connections between subnetworks.2.

Subnetwork State Management. The NMS needs to be aware of theoperational status of the subnetworks with regard to the network beingready to perform its intended functions, including link connections, trailoperational changes, and network components.

Transport Network Fault Management. The M4 interface managementis required to report network-view alarms and provide testing capability to


24 of 28 23-Sep-15 7:58 PM

isolate the problem. Specifically, this includes:

Log network alarms within a subnetwork to be retrieved by theNMS.

1.

Autonomously notify failures, such as termination point failures.2.

Provide loopback-testing capability that supports OAM cellloopback along a subnetwork connection or a segment of it.

3.

Network Security Management. The security framework for ATMnetworks is described in the ATM Forum document af-sec-0096.000. Itaddresses the security concerns from the perspectives of customers,public communities, and network operators. The main security objectivesare (1) confidentiality (confidentiality of stored and transferredinformation), (2) data integrity (protection of stored and transferredinformation), (3) accountability (for all ATM network service invocationsand for ATM network management activities), and (4) availability (correctaccess to ATM facilities).

Seven generic threats are considered in the threat analysis of an ATMnetwork. They are (1) masquerade or spoofing (pretence by an entity tobe a different entity), (2) eavesdropping (breach of confidentiality bymonitoring communication), (3) unauthorized access, (4) loss orcorruption of information, (5) repudiation (an entity involved in acommunication exchange subsequently denies the fact), (6) forgery, and(7) denial of service (failure to fulfill its functions by an entity orpreventing others from fulfilling).

Table 12.10 maps the threats to the objectives. Not all threats affect allsecurity objectives. For example, masquerade and unauthorized accessthreatens all security objectives, whereas eavesdropping only affectsconfidentiality of information.

Table 12.10. Mapping of Threats and Objectives

Threat Confidentiality DataIntegrity

Accountability Availability

Masquerade x x x x

Eavesdropping x

Unauthorizedaccess

x x x x

Loss orcorruption of

x x


25 of 28 23-Sep-15 7:58 PM

Threat Confidentiality DataIntegrity

Accountability Availability

information

Repudiation x

Forgery x x

Denial ofservice

x

A set of functional security requirements is identified to deal with thegeneric threats, and security services should be built to address theserequirements. Table 12.11 maps the security requirements to theservices needed to fulfill them. You may refer to af-sec-0096.000 forspecification of each of the services. Security recovery and managementof security do not have associated services but are part of therequirements.

Table 12.11. Mapping of Security Requirements and Services

Functional SecurityRequirements

Security Services

Verification ofidentities

User authentication

Peer entityauthentication

Data originauthentication

Controlled access andauthorization

Access control

Protection ofconfidentiality

Stored data Access control

----------------------- ------------------------------------------

Transferred data Confidentiality

Protection of dataintegrity

Stored data Access control

----------------------- ------------------------------------------

Transferred data Integrity

Strong accountability Non-repudiation


26 of 28 23-Sep-15 7:58 PM

Functional SecurityRequirements

Security Services

Activity logging Security alarm, audit trail, andrecovery

Alarm reporting Security alarm, audit trail, andrecovery

Audit Security alarm, audit trail, andrecovery

Securityrecovery/Managementof Security

The security framework is specifically applied across the M4 interface. Thenetwork management M4 security requirements and the logical MIB aredocumented in af-nm-0103.000. The MIB is defined independent ofprotocol implementation and is used for the transfer of information acrossATM management interfaces. Both resources and services are defined inthe MIB as managed entities in the ATM network element.

The Digital Exchange Interface (DXI) is an interface between a DigitalTerminal Equipment (DTE) and a Digital Circuit Equipment (DCE) thatconnects to a public data network. In the ATM network, the public datanetwork is a network of ATM switches. Figure 12.19 shows a high-levelview of the DXI interface. Typically, a DTE will be a hub or the router andthe DCE is a Digital Service Unit (DSU), which interfaces to an ATMswitch. More details on this interface and management of it can be foundin the ATM Forum document on Data Exchange Interface Specification,af-dxi-0014.000.

Figure 12.19. The ATM DXI Interface

Figure 12.20 shows the ATM DXI Local Management Interface (LMI). TheATM LMI defines the protocol for exchange of information across the DXIinterface, and supports DXI, AAL, and UNI-specific managementinformation. The LMI protocol supports the SNMP management systemand the ILMI Management Entity (IME) running on an ATM switch. TheATM DXI LMI MIB (af-dxi-0014.000) supports IETF ATM MIB and the ATM


27 of 28 23-Sep-15 7:58 PM

Related Content

Forum UNI MIB.

Figure 12.20. ATM DXI Local Management Interface


28 of 28 23-Sep-15 7:58 PM

Chapter 12. Broadband Network Management: WAN > MPLS Network Technology

Multiprotocol Label Switching (MPLS) is a fast-emerging WAN technologythat replaces pure IP and ATM networks [RFC 2702]. It combines therichness of IP and the performance of ATM networks.

One of the important developments in WAN transport technologies is theevolution of the MPLS protocol from IP and ATM. An IP-based network isfeature rich because of its extensive implementation and compatibilitywith Ethernet LAN. It routes packets intelligently. However, it is slow inperformance as it has to open each packet at the layer 3 level todetermine its next hop and its output port. The simultaneous transportof real-time and non-real-time traffic is difficult.

In contrast to IP, the ATM protocol is a high-performance cell-basedprotocol switching cells at the layer 2 level. It is capable of handlingreal-time and non-real-time traffic simultaneously and thus is superiorto the IP-based network. It has been deployed extensively in the WAN.However, its address incompatibility with the popular Ethernet LAN,along with difficult end-to-end circuit provisioning, has limited its usageat the customer premises network and hence related applications.

The MPLS protocol evolved to combine the richness of IP and theperformance of ATM networks. It is being deployed in convergentnetwork for broadband services handling real-time and non-real-timetraffic simultaneously, thus achieving high quality of service transport. Itcan be deployed in the legacy network of either IP- or ATM-base. Asimplified representation of an MPLS domain network with an ingressMPLS router and an egress MPLS router is shown in Figure 12.21. Thisshows the MPLS header, named label, added between layers 2 and 3 tothe packet or cell delivered by the external network to the ingressrouter and removed by the egress router before delivering to itsexternal router. As we will soon learn, the MPLS header at the output ofthe MPLS domain may not be the same as the one at the input to thedomain.

Figure 12.21. Simplified MPLS Network


1 of 12 23-Sep-15 7:56 PM

In general, the packet header contains information to choose the nexthop, which comprises information to perform two functions. The firstfunction partitions the entire set of possible packets into a set of forwardequivalent classes (FECs). The second function maps each FEC to a nexthop. In so far as the forwarding decision is concerned, different packetsthat are mapped into the same FEC are indistinguishable.

In conventional IP, forwarding is done based on the packets with thesame FEC being sent to the next hop on the same port. The FEC isdetermined by comparing the longest match for each packetsdestination address, called the address prefix. This is done for eachpacket at each router. This is shown in Figure 12.22(a). The incomingPackets 1 and 3 generate the same FEC 1 and are transmitted over Port3. Packets 2 and 4 generate FEC 2 and exit out of Port 4.

Figure 12.22. MPLS Packet Forwarding Based on the FECIndex Table



2 of 12 23-Sep-15 7:56 PM

In MPLS, the assignment of FEC to packets is done once at the ingressrouter. All packets that belong to a particular FEC are assigned the samepath or paths in the case of multipath routing, leaving the node. This isshown in Figure 12.22(b).

In MPLS, the assignment of FEC to a particular packet is done just onceat the ingress router and is encoded in a short fixed value length in thelabel. Once a packet is assigned to an FEC, no further header analysis isdone by subsequent routers; all forwarding is driven by the labels. Thelabel is sent along with the packet. At subsequent hops the label is usedas an index into a table, which specifies the next hop and a new label.The old label is replaced with the new label and the packet is forwardedto its next hop, as shown in Figure 12.22.

The MPLS forwarding paradigm has a number of advantages overconventional network layer forwarding. For example, MPLS forwardingcan be done by switches that are capable of doing label lookup andreplacement, but are either not capable of analyzing the network layer


3 of 12 23-Sep-15 7:56 PM

headers or are not capable of analyzing the network layer headers atadequate speed. The reader is referred to RFC 3031 to learn about thevarious benefits of MPLS that can be applied to broadband services.

Some MPLS routers analyze a packets network layer header not merelyto choose the packets next hop, but also to determine a packetsprecedence or class of service. They may then apply different discardthresholds or scheduling disciplines to different packets. MPLS allows(but does not require) the precedence or class of service to be fully orpartially inferred from the label. In this case, one may say that the labelrepresents the combination of an FEC and a precedence or class ofservice.

MPLS differs from an IP network, where each packet is looked at byeach router and is assigned an FEC using the longest match for eachpackets destination address. This obviously slows down the networkdata flow process. As IP packets are used by MPLS, the rich features ofIP are preserved as the egress MPLS router strips the label and outputsthe original IP packet.

Let us now ask how does MPLS preserves the efficiency of an ATM. In anATM, switching is done in layer 2. End-to-end path, virtual path,containing virtual circuit is determined by the VPIVCI. Thisestablishes a virtual circuit-switched path using soft switches. Asmentioned earlier, the MPLS router selects the next hop path using atable just as in the ATM switch and hence is efficient.

In the ATM network, a virtual circuit that has been established betweentwo stations, A and Z, is shown in Figure 12.2 The routing tables in theATM switches, A, B, and D, associated with this virtual circuit are shownin Table 12.2. The virtual circuit is first established prior to sending thedata. In our example, the virtual circuit is the combination of virtualcircuit links VCI-1, VCI-2, VCI-3, and VCI-4. Once the virtual circuit isestablished, all packets are transported in the sequence in which theywere transmitted by the source along the same path for a given session.Thus, Packets 1, 2, and 3 transmitted by Station A arrive at Station Z inthe correct sequence traversing the same links. Since the path is fixedfor the entire session, the transmission rate is considerably increasedjust as in the circuit-switched transmission mode.

In MPLS, packets with input labels are mapped to next hop with outputlabels. The header in layer 3 is ignored. Once a packet is assigned anFEC, no further analysis is done by subsequent routers. Figure 12.23illustrates IP routing with Interior Gateway Protocol (IGP) without trafficengineering (TE) and tunnels, and Figure 12.24 illustrates the Interior


4 of 12 23-Sep-15 7:56 PM

Gateway Protocol network topology with TE and tunnels implementation[Cisco, MPLS-TE]. The respective routing tables for router R1 areshown in Tables 12.12 and 12.13.

Figure 12.23. Topology without MPLS Tunnels

Figure 12.24. MPLS Topology with Tunnels

Table 12.12. R1 Routing Table withoutTunnels

Dest OutputInterface

Next Hop Metric

2.2.2.2 I1 2.2.2.2 1

3.3.3.3 I1 2.2.2.2 2

4.4.4.4 I1 2.2.2.2 3


5 of 12 23-Sep-15 7:56 PM

Dest OutputInterface

Next Hop Metric

I2 6.6.6.6 3

5.5.5.5 I1 2.2.2.2 4

I2 6.6.6.6 4

6.6.6.6 I2 6.6.6.6 1

7.7.7.7 I2 6.6.6.6 2

8.8.8.8 I1 2.2.2.2 4

I2 6.6.6.6 4

For simplicity, the IP addresses of the routers are designated as i.i.i.i foreach router Ri. Table 12.12, representing Figure 12.23, shows theoutput interface (logical port) and next hop for packets emanating fromR1 to Ri. The last column in the table shows the metric of the number ofhops from R1 to the destination router. The paths are chosen using IGP.For example, there are two choices for the label-switched path (LSP)from R1 to R4, both of which are shown in Table 12.12. The two pathsare R1R2R3R4 and R1R6R7R4. They both have the same metricof 3 hops. The former is transmitted via logical port I1 of R1 and thelatter I2 of R2. As can be observed, the router knows only its neighborrouter in its routing table.

LSPs for the same topology used in the MPLS protocol with MPLS-TE andtunnels are shown in Figure 12.24, and the routing table in the label-switching router (LSR) R1 is presented in Table 12.13. The LSRs R4and R5 are directly reached from R1 through tunneling. The transit timedelay is low and the throughput is higher as the intermediate routers inthe tunnel behave as pass through. The path from R1 to R4 now exitsout of the logical port T1 and has the next hop as R4. There are twometrics shown for educational purpose. It is three if no absolute valuefor the tunnel path metric. If an absolute value, in our case chosen as 1,the value of the metric is 1. For the case of an LSP from R1 to R5, thecorresponding metrics are 4 and 1.

Table 12.13. R1 Routing Table withTunnels

Dest 0 Intf Next Hop Metric

2.2.2.2 I1 2.2.2.2 1


6 of 12 23-Sep-15 7:56 PM

Dest 0 Intf Next Hop Metric

3.3.3.3 I1 2.2.2.2 2

4.4.4.4 T1 4.4.4.4 3 /1

5.5.5.5 T2 5.5.5.5 4/1

6.6.6.6 I2 6.6.6.6 1

7.7.7.7 I2 6.6.6.6 2

8.8.8.8 T1 4.4.4.4 4/2

Traffic engineering (TE) is concerned with the performance optimizationof operational networks. It includes measurement, modeling,characterization, and control of Internet traffic to achieve specificperformance objectives. A major goal of Internet TE is to facilitateefficient and reliable network operations, while simultaneouslyoptimizing network resource utilization and traffic performance. Theaspects of TE that are of interest concerning MPLS are measurementand control.

The control capabilities offered by existing Internet interior gatewayprotocols are not adequate for TE. A popular approach to circumvent theinadequacies of current IGPs is through the use of an overlay model,such as IP over ATM or IP over frame relay. The overlay model extendsthe design space by enabling arbitrary virtual topologies to beprovisioned atop the networks physical topology. The virtual topology isconstructed from virtual circuits that appear as physical links to theIGP-routing protocols. The overlay model provides additional importantservices to support traffic and resource control, including: (1)constraint-based routing at the VC level, (2) support for administrativelyconfigurable explicit VC paths, (3) path compression, (4) call admissioncontrol functions, (5) traffic shaping and traffic-policing functions, and(6) survivability of VCs. These capabilities enable the actualization of avariety of TE policies. For example, virtual circuits can easily be reroutedto move traffic from over-utilized resources onto relatively under-utilizedones.

For TE in large dense networks, it is desirable to equip MPLS with a levelof functionality at least commensurate with current overlay models.Fortunately, this can be done in a fairly straightforward manner. MPLS isstrategically significant for TE because it can potentially provide most ofthe functionality available from the overlay model in an integrated


7 of 12 23-Sep-15 7:56 PM

manner, and at a lower cost than the currently competing alternatives.Equally important, MPLS offers the possibility to automate aspects of theTE function.

As defined earlier, a router capable of supporting MPLS is called a label-switching router (LSR) and the end-to-end path of an MPLS circuit iscalled a label-switching path (LSP). The LSP starts at the ingress routerat the head end and terminates at the egress router at the terminatingend. As discussed above, the IGP has been extended to engineerintra-area and inter-area traffic [RFC 4105]. Figure 12.25 shows arepresentation of MPLS-TE implementation [Cisco]. LSRs use IGPextensions to create and maintain a TE link state database that issimilar to the one used by the open shortest path first(OSPF)/intermediate system to intermediate system (ISIS). It containsTE network topology that is updated by IGP flooding whenever a changeoccurs. Route is set up by Resource Reservation Protocol-TrafficEngineering (RSVP-TE), and link admission control establishes tunnelsthat have resources. TE control establishes and maintains links, and theforwarding engine manages the data flow across the LSP.

Figure 12.25. MPLS-TE System Block Diagram (Head-EndRouter)

The control and data planes are separated in MPLS. This permits a


8 of 12 23-Sep-15 7:56 PM

variety of signaling protocols to be used for the signal path. This offersflexibility for the implementation of VoIP using the legacy H.323signaling protocol or Session Initiation Protocol (SIP).

A label is a short, fixed length, locally significant identifier, which is usedto identify an FEC. The label, which is put on a particular packet,represents the forwarding equivalence class (FEC) to which that packetis assigned. The MPLS label is a 4-octet or 32-bit long header and isshown in Figure 12.26. It comprises 20-bit label, 3-bit experimental,1-bit stack indicator, and 8-bit time-to-live (TTL) information in thatsequence. The label is assigned by a downstream router, as we will seelater. The experimental bits are used to describe the class of servicesuch as QoS. The stack indicator bit is 0 if it is the bottom of the stackof labels or else it is 1. TTL defines the maximum number of hops.

Figure 12.26. MPLS Shim Header

The MPLS label (or stack of labels) is inserted or shimed between layer2 and layer 3, as shown in Figure 12.27. Thus, for an IP-basedprotocol, the MPLS label is inserted between the 802.3 MAC header andthe network layer IP header. In the case of the ATM protocol, the MPLSshim without the TTL replaces VPI and VCI fields.

Figure 12.27. Encapsulation of an MPLS-Labeled Packet


9 of 12 23-Sep-15 7:56 PM

Label switching is the process of extracting information from the label-routing look-up table and creating an outbound label, port, andoperations to be performed. The operations are push a new header,pop a header off the stack, or swap a header.

If Ru is the upstream LSR and Rd is the downstream in the LSRs, theymay agree that when Ru transmits a packet to Rd, Ru will label thepacket with label value L if and only if the packet is a member of aparticular FEC F. That is, they can agree to a binding between label Land FEC F for packets moving from Ru to Rd. As a result of such anagreement, L becomes Rus outgoing label representing FEC F, and Lbecomes Rds incoming label representing FEC F. The upstream routerpacket is assigned the label by the downstream router. Note that L doesnot necessarily represent FEC F for any packets other than those thatare being sent from Ru to Rd. L is an arbitrary value whose binding to Fis local to Ru and Rd. Ru and Rd with label L are called a label-bindingpair.

The packets being sent from Ru to Rd do not imply either that the


10 of 12 23-Sep-15 7:56 PM

packet originated at Ru or that its destination is Rd. They includepackets that are transit packets at one or both of the LSRs. Figure12.28 shows two upstream LSRs, Ru1 and Ru2, communicating withthe downstream LSR Rd. Packet L1 with FEC F and Packet L2 from Ru2belonging to the same FEC F arrive at downstream LSR Rd. Labels L1and L2 are outgoing labels of Ru1 and Ru2, respectively. They are Rdsincoming labels. Since they belong to the same FEC, they exit thesame port of Rd with the newly assigned Labels L3 and L4 to the nexthop from Rd. L3 is assigned by the LSR that is downstream from the Rdshown in Figure 12.28.

Figure 12.28. Label Generation

In conclusion, MPLS offers the best of IP and ATM network and is wellsuited for multimedia broadband network. It is currently being deployedby the broadband service providers at a rapid rate.

Let us now try to put together an integrated picture of LSP, LSR, LDP(label distribution protocol), and label in an MPLS network. An LSPcomprises the path from ingress to egress router of an MPLS circuit.This is shown in Figure 12.29. It is made up of one or more entries inMPLS-switching tables in each of the LSRs along the LSP path. Each LSRhas one input label assigned by the upstream LSR and one or moreoutput labels based on whether a packet is unicast or multicast. Anoutput label is interpreted by the downstream LSR as an input labellocally. An input label is used in an LSR to determine the output label(s)and the output interface(s) based on LSP. Thus, the output label 23tagged by LSR R2 is interpreted by LSR R3 as the input label and isused to tag the output packet with output label 34 and transmit it onoutput interface 3. The input to output relationship is derived from thecross-connect table of the LSR.


11 of 12 23-Sep-15 7:56 PM

Related Content

Figure 12.29. LSP, LSRs, and Labels

LDP is used by the egress LSR to notify the LSP to all the LSRs that areaffected along the LSPs path. An LDP is a set of procedures by whichone LSR informs another of the meaning of labels used to forward trafficbetween and through them. [RFC 3036, 3037] LDP associates an FEC[RFC 3031] with each label it distributes. Two LSRs that use LDP toexchange FEC-label binding information are known as LDP Peers, andwe speak of there being an LDP Session between them.

An alternative to LDP, which is distributed architecture, is to useRSVP-TE that has a signaling element built in to distribute the LSPinformation.

Figure 12.29(a) presents the case of an LSP for the non-tunnel case oftransmission from R1 to R4 in Figure 12.23. Figure 12.29(b) showsthe case for a tunneled path T1 from R1 to R4 shown in Figure 12.24.


12 of 12 23-Sep-15 7:56 PM

Chapter 12. Broadband Network Management: WAN > MPLS OAM Management

In an MPLS network, the data planes and control planes are different. Adata plane that carries traffic is used for MPLS OAM (operations,administration, and maintenance). In other words, OAM packets followthe data path. Sections 12.5.212.5.6 enumerate the OAMrequirements extracted from [Cavendish et al., 2004; Dini et al.,2004; ITU Y1710; Nadeau et al., 2005; RFC 3429, 4377]. TheMPLS layer can be visualized as the layer positioned between layers 2and 3 and hence OAM of MPLS needs to address both networks, namelyATM/MPLS and IP/MPLS. OAM in ATM/MPLS network is briefly addressedand compared with the IP/MPLS network in Section 12.5.1; and theOAM IN IP/MPLS network is covered in greater detail in Sections12.5.212.5.4.

As we have seen, there is commonality between ATM and MPLS in thelower layers. The LSP in MPLS serves the functions of VCC and VCP inthe ATM. However, there is a paradigm shift in migrating from ATM OAMto MPLS OAM as illustrated in Table 12.14 [Dini et al., 2004].

Table 12.14. ATM OAM vs. MPLS OAM

ATM OAM MPLS OAM

Paradigm: Virtual circuits Paradigm: Label-switched paths

Bi-directional Usually uni-directional

Established via ATM signaling andmanagement

Established closely tied to controlplanes

Fixed hierarchy VPVC Variable label stack

Connection-oriented Can be Connectionless

Single route May use ECMP (equal-costmultipath)

No penultimate popping Penultimate hop popping


1 of 20 23-Sep-15 7:56 PM

The ATM layer is constantly supervised using OAM cells. OAM functionsof the ATM layer are accomplished by OAM cell flows being performedfor both VC and VP connections. The fault management function isperformed by continuity checking, verifying whether a VCC that hasbeen idle for a period of time is still functioning. This is done by sendingcontinuity-check cells along that VCC. Performance management is doneby transmitting OAM cells at fixed intervals in both directions fornotifying other NEs of detected errors.

MPLS is usually uni-directional and has a tightly coupled relationshipbetween data and control planes. In contrast to ATM, the OAM packetsare variable label stack and can be connectionless.

When an LSP fails to deliver a packet to the egress LSR, the failure canbe due to several reasons and cannot always be detected by the MPLScontrol plane. LSP failures require the testing of specific packet flows,besides failure of the network infrastructure, because in many instancespacket flows may get interrupted without network (link/node) failure.Figure 12.30 shows various kinds of plausible fault scenarios. A and Bare the intended receivers for A and B, respectively. Fault (defect)scenarios are: (a) Simple loss of connection, (b) Misconnection, (c)Swapped connection, (d) Mismerging, and (e) Loop/unintendedreplication [Cavendish et al., 2004]. In order to ensure that LSPfailure is always detected, it is important that the OAM packets usedtravel the same path as regular data packets as the data plane alwaysinteracts with the control plane.

Figure 12.30. MPLS Fault Scenarios


2 of 20 23-Sep-15 7:56 PM

There are two distinct methods specified in each of the standards, ITU-Tand IETF, for detecting LSP fault detection and recovery. ITU-TRecommendation Y 1711 has standardized connectivity verification (CV)function, whereas bi-directional forwarding detection (BFD) function isspecified within IETF.

Connectivity Verification (ITU-T Y.1711). The basic idea of the CVfunction is to send test packets (CV packets) periodically (one persecond) from the ingress LSR to the egress LSR with the identity of theingress LSR and originating LSP. The egress LSR analyzes theidentification information of the received CV packets to detect defects.Figure 12.31 shows the CV packet format. It has two label stackentries. The top label is the same as that in user packets. The secondlabel identifies the OAM packet via a reserved value of 14. Theexperimental bits, EXP, in the top label, and the OAM label are both setto the highest priority value used across the entire network. The stackbit, S, of the OAM label is set to 1 because it is the bottom label.Moreover, to ensure that CV packets will not go beyond the egress LSR,the MPLS network is treated as one-hop. The function type fieldindicates the OAM type: a code point 01H is assigned for CV packets.The next two fields, LSR ID and LSP ID, form the LSP trail terminationsource identifier (TTSI). The TTSI is used to identify the ingress LSR (viathe IP address of its output port) and originating LSP (via the LSP ID).


3 of 20 23-Sep-15 7:56 PM

The BIP 16 (Byte Interleaved Parity) field is used for errordetection/correction. The ingress LSR sends CV packets with its TTSIvalue every second. The egress LSR receives CV packets and analyzestheir TTSI values to determine whether they are expected values or not.If the egress LSR receives an expected TTSI, it declares no defect. If itdoes not receive any CV packet for more than 3 seconds, it declares aloss of connectivity defect (dLOCV), Figure 12.30(a). The egress LSRwaits before declaring dLOCV to avoid misdetection of dLOCV due to anOAM packet loss. If there are two consecutive OAM packet losses, theegress LSR may mistakenly declare dLOCV, but this happens very rarelyin normal operating conditions. In addition, as an egress LSR waits for 3seconds, the expected delay variation, which is normally much less thana second, is acceptable. If an egress LRS receives an unexpected TTSI,it declares a TTSI mismatch defect (dTTSI-mismatch), Figure 12.30(b)and (c). If it receives both unexpected and expected TTSI values, itdeclares a TTSI mismerge defect (dTTSI-mismerge), Figure 12.30(d).If it receives more than five CV packets with expected TTSI valueswithin 3 seconds, it declares excessive reception defect (dExcess),Figure 12.30(e). Before we discuss BFD, we will address LSP ping andLSP traceroute, which are used in BFD.

Figure 12.31. Connectivity Verification Packet

LSP Ping. LSP ping, which is described in detail in RFC 4379, ismodeled after the Internet ping paradigm. The basic idea is to verifythat packets that belong to a particular FEC actually end their LSP on anLSR that is an egress for that FEC. This test is carried out by sending apacket (called an MPLS echo request) along the same data path asother packets belonging to this FEC. An MPLS echo request also carriesinformation about the FEC whose MPLS path is being verified. This echorequest is forwarded just like any other packet belonging to that FEC. In


4 of 20 23-Sep-15 7:56 PM

the ping mode, the packet should reach the end of the path, at whichpoint it is sent to the control plane of the egress LSR, which thenverifies whether it is indeed an egress for that FEC.

An MPLS echo request is a (possibly labeled) IPv4 or IPv6 UDP packet;the contents of the UDP packet have the format shown in Table 12.15.

Table 12.15. MPLS Echo Request Packet

Byte 1 Byte 2 Byte 3 Byte 4

Version Number Global Flags

Message Type Reply Mode Return Code ReturnSubcode

Senders Handle

Sequence Number

TimeStamp Sent (seconds)

TimeStamp Sent (microseconds)

TimeStamp Received (seconds)

TimeStamp Received (microseconds)

TLV

The Version Number is currently 1. The Global Flags field is a bit vectorthat includes a V flag prescribing the FEC Stack Validation. The MessageType is either echo request (1) or echo reply (2). The Reply Mode cantake one of the following values: Do not reply (1), Reply via anIPv4/IPv6 UDP packet (2), Reply via an IPv4/IPv6 UDP packet withRouter Alert (3), and Reply via application-level control channel (4).Return Codes and Subcodes are set to zero an

Network Management Principles and Practice Mani Subramanian 2nd Edition Ch12

Documents

Transcript of Network Management Principles and Practice Mani Subramanian 2nd Edition Ch12