Glossary

The term quality of service (QoS) broadly refers to a network's capacity to provide better service to selected network traffic for various technologies, including IP routed networks, Frame Relay, ATM, Ethernet, 802.1 networks, SONET, and so on. The majority of terms in this glossary refer to IP QoS. For further listings, refer to the following website:

http://www.cisco.com/univercd/cc/td/doc/cisintwk/ita/index.htm

1+1: A method of protecting traffic in which a protection channel exists for each working traffic channel. For optical systems, the protection channel fibers can be routed over a path separate from the working fibers. The traffic signal is bridged to both the working and protection transmitters so the protection signal can be selected quickly if the working channel fails.

1:n: A method of protecting traffic in which one protection channel exists for n traffic channels. Only one traffic channel can be switched to the protection channel at any given time.

1G mobile network: First-generation mobile network. Refers to the initial category of mobile wireless networks that use analog technology only. Advanced Mobile Phone Service (AMPS) is an example of a 1G mobile network standard.

100BASE-FX: 100-Mbps baseband Fast Ethernet specification using two strands of multimode fiber-optic cable per link. To guarantee proper signal timing, a 100BASE-FX link cannot exceed 1312 feet (400 meters) in length. Based on the IEEE 802.3 standard.

100BASE-T: 100-Mbps baseband Fast Ethernet specification using UTP wiring. Like the 10BASE-T technology on which it is based, 100BASE-T sends link pulses over the network segment when no traffic is present. However, these link pulses contain more information than those used in 10BASE-T. Based on the IEEE 802.3 standard.

100BASE-T4: 100-Mbps baseband Fast Ethernet specification using four pairs of Categories 3, 4, or 5 UTP wiring. To guarantee the proper signal timing, a 100BASE-T4segment cannot exceed 328 feet (100 meters) in length. Based on the IEEE 802.3 standard.

100BASE-TX: 100-Mbps baseband Fast Ethernet specification using two pairs of either UTP or STP wiring. The first pair of wires receives data; the second transmits data. To guarantee the proper signal timing, a 100BASE-TX segment cannot exceed 328 feet (100 meters) in length. Based on the IEEE 802.3 standard.

100BASE-X: 100-Mbps baseband Fast Ethernet specification that refers to the 100BASE-FX and 100BASE-TX standards for Fast Ethernet over fiber-optic cabling. Based on the IEEE 802.3 standard.

1000BASE-F: A 1-Gbps IEEE standard for Ethernet LANs.

2B1Q: 2 binary 1 quaternary. An encoding scheme that provides a 2-bits per baud, 80-kbaud per second, 160-kbps transfer rate. The most common signaling method on ISDN U interfaces. The 1988 ANSI spec T1.601 defines this protocol in detail.

2G mobile network: Second-generation mobile network. Refers generically to a category of mobile wireless networks and services that implement digital technology. GSM is an example of a 2G mobile network standard.

2G+ mobile network: Second-generation plus mobile network. Refers generically to a category of mobile wireless networks that support higher data rates than 2G mobile networks. GPRS is an example of a 2G+ mobile network standard.

24th channel signaling: See 2G mobile network

3G mobile network: Third-generation mobile network. Refers generically to a category of next-generation mobile networks, such as UMTS and IMT-2000.

4B/5B local fiber: 4-byte/5-byte local fiber. Fiber channel physical media used for FDDI and ATM. Supports speeds up to 100 Mbps over multimode fiber.

6BONE: The Internet's experimental IPv6 network.

8B/10B local fiber: 8-byte/10-byte local fiber. Fiber channel physical media that supports speeds up to 149.76 Mbps over multimode fiber.

802.x: A set of IEEE standards for the definition of LAN protocols.

A

access router: A router, at a customer site, that connects to the network service provider WAN. Sometimes termed a customer premises equipment (CPE) router.

ACL: access control list. ACLs are used for security inside of programs and operating systems. For example, Cisco uses ACL to control access across routers, and Windows NT uses ACLs for directory and file access.

administrative domain: A collection of network elements under the same administrative control and grouped together for administrative purposes. It is usually managed by a single corporate entity. For QoS enforcement purposes, a network domain refers to any domain that shares a common QoS policy. It may or may not overlap other kinds of domains such as IP or NT domains.

admission control: A policy decision applied initially to QoS requests for controlling the admission of network traffic from outside a given administrative domain. Admission control is closely tied to accounting and relies on source authentication. Contrast with policing, which occurs after a request is accepted and data is flowing.

AF: assured forwarding. A specific DiffServ behavior that divides IP packets into four separate per-hop behavior (PHB) classes. Using these classes, a provider may offer different levels of service for IP packets received from a customer domain. Each AF class is allocated a specified amount of buffer space and bandwidth.

ARP: Address Resolution Protocol. Internet protocol, defined in RFC 826, used to map an IP address to a MAC address. Allows host computers and routers to determine the data link layer address corresponding to the IP address in a packet routed through the LAN. Although the packet is addressed to an IP address, the LAN hardware responds only to data link layer addresses. The host or router with the destination IP address replies with its own data link layer address in an ARP response, which the forwarding host or router will use to construct a data link layer frame. The result is stored in cache memory so subsequent packets addressed to the same destination can be routed without an explicit ARP process.

ARQ: automatic repeat request. A method of checking transmitted data on high-speed data communications systems in which the sender encodes an error-detection field based on the contents of the message. The receiver recalculates the check field and compares it with the received field. If the fields match, a positive acknowledgment (ACK or PAK) is returned to the sender. If the fields do not match, a negative acknowledgment (NAK) is returned to the sender.

AS: autonomous system. A self-connected set of networks that are generally operated within the same administrative domain. See also administrative domain

ATM: Asynchronous Transfer Mode. A data framing and transmission architecture with built-in QoS capabilities, designed to carry voice, video, and data. ATM operates at Layer 2 of the OSI model and is a high-speed, connection-oriented, packet-switching, multiplexing architecture. Bandwidth is divided into fixed-size cells of 53 bytes each, including headers, which are allocated to services on demand. Bandwidth can be dynamically allocated. ATM can offer bandwidth rates of up to multigigabit bandwidth. Although relatively few native ATM applications exist, TCP/IP traffic can be sent over an underlying ATM layer. In principle, ATM could be used over the LAN, MAN, or WAN; in practice, ATM is prevalent in WANs and Internet backbones.

AVVID: See Cisco AVVID

B

backbone: A network linking together multiple domains, either enterprises or service provider.

bandwidth: Transmission capacity of a computer channel or communications line or bus, usually stated in bits per second (bps). Bandwidth indicates the theoretical maximum capacity of a connection, but as the theoretical bandwidth is approached, negative factors such as transmission delay can cause deterioration in quality.

bandwidth manager: A traffic manager deployed at congestion points that limits access to network resources. It often requires locating a proprietary hardware device directly on the network and may be an additional point of failure. It cannot coordinate multiple traffic flows or resolve conflicting QoS requests made by multiple clients, and is therefore not an end-to-end QoS solution.

bandwidth throughput: Average traffic throughput over a given time interval, expressed in bits per second (bps).

best-effort service: The default behavior of TCP/IP networks in the absence of QoS measures. TCP/IP nodes will make their best effort to deliver a transmission, but will drop packets indiscriminately in the event of congestion, managing the bandwidth, or assigning priority to delay sensitive packets. The Internet today is a good example of best-effort service. Best effort is suitable for a wide range of networked applications such as general file transfers or e-mail.

BGP: Border Gateway Protocol. An Internet routing protocol used to pass routing information between different administrative routing domains or autonomous systems. BGP does not pass explicit topology information. BGP is often used between Internet service providers (ISPs).

border router: Generally describes routers on the edge of an autonomous system (AS). Uses BGP to exchange routing information with another administrative routing domain. Can also describe any router that sits on the edge of a routing sub area, such as an Open Shortest Path First (OSPF) area border router.

broadcast storm: An overload condition in a network created by an incorrect packet broadcast onto the network that causes multiple hosts to respond all at once. Typically, the response contains equally incorrect packets, which causes the storm to grow exponentially in severity.

business-critical application: A networked application that is deemed critical to the success of the business organization.

C

CBQ: class-based queuing. A public domain QoS methodology for classifying packets and queuing them according to criteria defined by an administrator to provide differential forwarding behavior for each traffic class.

CBR: constant bit rate. Multimedia streams, audio, and video are examples of CBR applications because they send at a relatively steady data rate with constant bandwidth allocations. A class of service defined by ATM.

CBWFQ: class-based weighted fair queuing. Enables you to define traffic classes that are based on certain match criteria, such as access control lists, input interface names, protocols, and QoS labels.

CDP: Cisco Discovery Protocol. A device-discovery protocol that runs on all Cisco-manufactured equipment. Enables a device to advertise its existence to other devices and receive information about other devices in the network.

CDV: cell delay variation. In an ATM network, the variation in cell delay through the network.

CER: cell error ratio. In an ATM network, ratio of errored cells to transmitted cells. Measures the accuracy of cell transmission.

CFQ: class-based fair queuing. Per-class packet scheduling.

CIR: committed information rate. A minimum access rate the service provider commits to providing the customer.

Cisco AVVID: Cisco Architecture for Voice, Video, and Integrated Data.

Cisco IP phone: Telephone that provides voice communication over an IP network.

Cisco IP Telephony Solutions: A software and hardware product suite offering an IP alternative to traditional private branch exchanges (PBXs).

class: An abstraction that can be determined by different policy criteria such as IP packet header content (such as source or destination addresses, port numbers, or transport protocol), time of day, ingress point, and so forth. The definition of a class can differ at different locations on the network.

classifier: An entity that selects packets based on the content of packet headers according to defined rules.

CLEC: competitive local exchange carrier. A telephone company that competes with other local phone companies within an exchange, or calling area, to provide local telephone service.

CLI: command-line interface. A method of interacting with a device (for example, router or switch) by giving it lines of textual commands (such as a sequence of characters) either from keyboard input or from a script.

client: Node or software program (front-end device) that requests services from a server.

client/server computing: Term used to describe distributed computing (processing) network systems in which transaction responsibilities are divided into two parts: client (front end) and server (back end).

client/server model: The process of workload sharing between the client, the server, and the network. Examples include the name server/name resolver paradigm of the Domain Name System (DNS), as well as file server/file-client relationships, such as Network File System (NFS) and diskless hosts. See also DNS and NFS

CLP: cell loss priority. In ATM cells, a bit that specifies whether a cell can be discarded, if necessary; for example, if network congestion or policing occurs.

CLR: cell loss ratio. In an ATM network, the ratio of lost ATM cells to transmitted cells; measures the percentage of cells lost between two points in the network. Cell loss is typically caused by severe network congestion; it can also result from bit errors in the cell header.

CMR: cell misinsertion rate. In an ATM network, the total number of misinserted cells observed during a specified time interval divided by the time interval duration.

codec: coder-decoder. 1. A device that typically uses pulse code modulation to transform analog signals into a digital bit stream, and digital signals back to analog. 2. In Voice over IP, Voice over Frame Relay, and Voice over ATM, a software algorithm used to compress/decompress speech or audio signals.

codepoint: Codepoint markings are made in a new implementation of the IP version 4 type of service (IPv4 ToS) header called the DiffServ field (6 bits, reserving 2 for congestion notification), and are used to select a per-hop behavior (PHB). This marking takes place on the host, on a boundary, or on an edge device.

congestion avoidance: Congestion avoidance is the action a network takes to head off congestion before it can occur, anticipating circumstances in which flows or aggregated flows might no longer receive designated service levels due to excessive traffic loads at points in the network. An example is the constructive application of a drop policy such as RED to provide implicit feedback to host systems to reduce network traffic during congestion.

congestion control: Mechanisms that control traffic flow so that intermediate network devices and end stations are not overwhelmed.

congestion management: A mechanism at multiplexing points that imposes order when traffic exceeds network capacity for a flow or set of aggregated flows. It determines whether some packets must be discarded, and, if so, it preserves the more important packets. Queuing, scheduling, and traffic shaping are among the most popular techniques.

controlled load: Tightly approximates best-effort service under unloaded conditions, a high-level but nonguaranteed service. In the proposed IETF integrated services model, this level of service is designed for multimedia applications where time delay is not critical but quality of the delivery is important. This service is appropriate for applications such as one-way voice or video, but not for real-time applications.

converged network: A network that combines varied traffic types such as data, voice, and multimedia. Most analysts expect the converged network of the future to be based on Internet protocols.

conversation: The exchange of traffic that takes place between a client and server during the execution of an application transaction. Similar in nature to that of a human conversation, involving requests and responses.

COPS: Common Open Policy Service. A standard defining a simple protocol for provisioning QoS by outsourcing policy-based admission control over requests for network resources.

core router: A router on the network service provider WAN that has no direct connections to any routers at customer sites.

CoS: class of service. A category based on type of user, type of application, or some other criteria that QoS systems can use to provide differentiated classes of service. The characteristics of the CoS may be appropriate for high-throughput traffic, for traffic with a requirement for low latency, or just for best effort.

cRTP: compressed RTP. IP RTP header compression reduces the 40-byte IP+UDP+RTP header to 2 to 4 bytes, thereby reducing the bandwidth required per voice call on point-to-point links. The header is compressed at one end of the link and decompressed at the other end. See also RTP

CTD: cell transfer delay. In an ATM network, measures the peak and mean delay that cells experience while traveling from one point in the network to another.

customer router: A router at a customer site that is not directly connected to the network service provider WAN.

D

DCOM: Distributed Component Object Model. Protocol that enables software components to communicate directly over a network. Developed by Microsoft and previously called Network OLE, DCOM is designed for use across multiple network transports, including Internet protocols such as HTTP.

demarcation point: The point where the service provider's equipment meets a customer's equipment.

differentiated services boundary: The edge of a DiffServ domain, where classifiers and traffic conditioners are likely to be deployed. A DiffServ boundary can be further subdivided into ingress and egress nodes, where the ingress/egress nodes are the downstream/upstream nodes of a boundary link in a given traffic direction. A DiffServ boundary may be co-located with a host, subject to local policy.

differentiated services domain: A contiguous portion of the Internet over which a consistent set of DiffServ policies are administered in a coordinated fashion. A DiffServ domain can represent different administrative domains or autonomous systems, different trust regions, different network technologies (such as cell/frame), hosts and routers, and so forth.

differentiated services field: The IPv4 header ToS octet or the IPv6 Traffic Class octet when interpreted in conformance with the definition given in RFC 2474.

DiffServ: differentiated services. An IETF standard for a small, well-defined set of per-packet building blocks from which a variety of services may be built, thereby providing a framework for delivering QoS in networks.

DNS: Domain Name System. The system used to translate names of network devices into IP addresses.

DSBM: designated subnet bandwidth manager. A device on a managed subnetwork that acts as the subnet bandwidth manager for the subnetwork to which it is attached. This is done through an election process specified in the IETF SBM protocol specification.

E

e2e QoS: end-to-end QoS. The capability of a network to deliver service needed by a specific network application from end to end, with the capability to provide both class of service and reserved bandwidth for different types of network traffic.

edge device: A device, such as a router or a gateway, that is deployed at the border of an administrative domain. Such devices control traffic through one point only.

edge-to-edge QoS: Applies to QoS within a network that connects to other networks rather than hosts or end systems (the typical service provider network, for example), with some level of control over bandwidth, jitter, delay, and loss provided by the network.

EF: expedited forwarding. A per-hop behavior (PHB) in the DiffServ standard used to create a virtual leased-line service.

EPD: early packet discard. A congestion-avoidance mechanism generally found in ATM networks.

EtherChannel: Cisco EtherChannel technology is a trunking technology based on grouping several full-duplex 802.3 Ethernet links to provide fault-tolerant, high-speed links between switches, routers, and servers.

ETSI: European Telecommunications Standards Institute. An independent, nonprofit organization that produces telecommunications standards.

F

FEC: forward error correction. Sending redundant data so that if data loss occurs, data recovery is possible without retransmission.

Filterspec: Part of the Resource Reservation Protocol (RSVP) that is used by senders in a PATH message to identify themselves, and by recipients in RESV messages to select which senders they are interested in receiving messages from.

FIFO: first in, first out. Simple tail-drop FIFO queue used in (best-effort) IP service. The simplest queuing algorithm used in switches and routers because the first packets to arrive in a buffer are the first to leave. FIFO offers no configuration options; therefore, it is sometimes accompanied by other schemes such as random early detection (RED) and weighted fair queuing (WFD).

flow: A set of packets traversing a network element, all of which are covered by the same request for control of QoS.

flowspec: Information about the QoS that is required by the flow and contained in an RSVP message and that tells each router along a data path which resources it should reserve for that particular flow.

G

G.711: An audio-compression standard used for digital telephones on a digital PBX/ISDN. In G.711, encoded voice is already in the correct format for digital voice delivery in the PSTN or through PBXs. G.711 uses a bandwidth of 64 kbps. G.711-compliant devices can communicate with other G.711 devices, but not with G.723 devices. Described in the ITU-T standard in its G-series recommendations.

G.723.1: Describes a compression technique that can be used for compressing speech or audio signal components at a very low bit rate as part of the H.324 family of standards. This codec allows dissimilar communication devices to communicate with each other using a standardized communications protocol. Used for digital telephones on a digital PBX/ISDN, producing digital audio at either 6.4 or 5.3 kbps. Described in the ITU-T standard in its G-series recommendations.

G.729: ITU-T's standard voice algorithm. Describes the encoding/decoding of speech.

gatekeeper

1. Component of an H.323 conferencing system that performs call address resolution, admission control, and subnet bandwidth management. 2. Telecommunications: H.323 entity on a LAN that provides address translation and control access to the LAN for H.323 terminals and gateways.

The gatekeeper can provide other services to the H.323 terminals and gateways, such as bandwidth management and locating gateways. A gatekeeper maintains a registry of devices in the multimedia network. The devices register with the gatekeeper at startup and request admission to a call from the gatekeeper. In the Cisco CallManager, for example, the gatekeeper is a device that supports the H.225 RAS message set used for call admission control (CAC), bandwidth allocation, and dial-pattern resolution. There is one gatekeeper device per Cisco CallManager cluster.

gateway: The point at which a circuit-switched call is encoded and repackaged into IP packets. A gateway is an optional element in an H.323 conference and bridges H.323 conferences to other networks, communications protocols, and multimedia formats.

guaranteed service (hard QoS, quantitative QoS, reserved bandwidth): A service level that attempts to guarantee a minimal delay for traffic delivery. In the IETF IntServ model, guaranteed service is intended for real-time applications, such as teleconferencing. Guaranteed service is an absolute reservation of network resources, typically bandwidth, which implies reservation of buffer space along with the appropriate queuing disciplines and so on to ensure that specific traffic gets a specific service level. Typically, guaranteed service is associated with a fine level of traffic classification, so that particular flows (or aggregates) have network resources reserved for them so that required guarantees can be met.

GUI: graphical user interface. User environment that uses pictorial as well as textual representations of the input and output of applications and the hierarchical or other data structure in which information is stored.

HI

H.323: A standard approved by the International Telecommunication Union (ITU) that defines how audio-visual conferencing data is transmitted across networks.

IEEE: Institute of Electrical and Electronics Engineers. Professional organization whose activities include the development of communications and network standards.

IEEE 802.1D: 802.1p was incorporated into and superceded by the IEEE 802.1D standard.

IEEE 802.1p: An IEEE standard for improving support of time-critical and multicast-intensive applications across bridged LANs. A Layer 2 scheme that lets end stations request priority and network devices enforce it, using a tag in the packet header. Incorporated into and superceded by IEEE 802.1D.

IEEE 802.1Q: An IEEE standard for providing a virtual LAN capability within a campus network. It establishes a standard format for frame tagging (Layer 2 VLAN markings).

IETF: Internet Engineering Task Force. An engineering and protocol standards body that develops and specifies protocols and Internet standards, generally in the network layer and above. See www.ietf.org.

IntServ: integrated services. One of two architectures (the other being DiffServ) defined by the IETF. IntServ uses the Resource Reservation Protocol (RSVP) to explicitly signal the QoS needs of an application's traffic along the devices in the end-to-end path through the network. If every network device along the path can reserve the necessary bandwidth, the originating application can begin transmitting. Request For Comments (RFC) 2205 defines RSVP, and RFC 1633 defines IntServ.

IP Precedence/CBQ: A 3-bit value in the IP packet header meant to designate the relative priority of a packet, applied on a host, access router, or gateway, and then used by core routers. Values range from 0 to 7, but typically 6 and 7 are not used by applications because network control messages use these. For example, a brokerage firm might assign a higher IP Precedence/CBQ value to real-time stock trades than to e-mail to ensure that the trading gets expedited delivery. Same as ToS bits. The DiffServ codepoint has been designed to be backward compatible with IP Precedence.

IPSec: IP Security Protocol. Applicable to both IPv4 and IPv6, IPSec defines encryption, authentication, and key management. An extension header is added to the packet to identify those that are authenticated and encrypted.

IPT: IP telephony. Transport of telephone calls over the Internet, no matter whether traditional telephony devices, multimedia PCs, or dedicated terminals take part in the calls, and no matter whether the calls are entirely or only partially transmitted over the Internet.

IPv4: Internet Protocol version 4. The most widely deployed version of the Internet Protocol, IPv4 provides some basic traffic-classification mechanisms with its IP Precedence/CBQ and ToS header fields. However, network hardware and software traditionally have not been configured to use them.

IPv6: Internet Protocol version 6. An update to the Internet Protocol. Most of the refinements concentrate on basics such as expanding the IP address-numbering scheme to accommodate the growth of the Internet. However, IPv6 does include a class header field that is explicitly intended to designate a class of service (an extension of IPv4's IP Precedence/CBQ field).

ISO: International Organization for Standardization. Responsible for a wide range of standards, including those relevant to networking. ISO developed the popular OSI networking reference model. See also OSI

ISP: Internet service provider. Communications service company that provides Internet access and services to its customers.

ITU: International Telecommunications Union. Under the auspices of the United Nations. Develops radio (ITU-R) and telecommunications (ITU-T) standards. Prior to 1993, the ITU-T Standardization Sector was known as the Consultative Committee for International Telegraphy and Telephony (CCITT).

JKL

jitter: The distortion of a signal as it is propagated through the network, where the signal varies from its original reference timing, and packets do not arrive at its destination in consecutive order, or on a timely basisthey vary in latency. In packet-switched networks, jitter is a distortion of the interpacket arrival times compared to the interpacket times of the original transmission. Also referred to as delay variance. This distortion is particularly damaging to multimedia traffic.

keepalive message: A message sent by one network device to another indicating that the circuit between the two is still active.

LAN: local-area network. Data communications network connecting computers and related equipment, usually over an area not greater than 10 km.

LANE: LAN emulation. A set of services and protocols that provide for the emulation of LANs, using ATM as a backbone to allow connectivity between LAN and ATM attached end stations.

latency: Delay in a transmission path or in a device within a transmission path. In a router, latency is the amount of time between when a data packet is received and when it is retransmitted. Also referred to as propagation delay.

Layer 1: The physical network layer in the OSI model. Applies to basic wiring. An example of a Layer 1 protocol is 10BASE-T.

Layer 2: The data link layer in the OSI model. Examples include Ethernet, Token Ring, and HDLC. Layer 2 is divided into the Media Access Control (MAC) sublayer and the Link Layer Control (LLC) sublayer. Layer 2 is where LAN protocols such as Ethernet, Token Ring, and FDDI are defined.

Layer 3: The network protocol layer in the OSI model. Examples include IP and IPX. Layer 3 is the packet, or network, layer. A routing protocol, such as Routing Information Protocol (RIP), operates at Layer 3. Other Layer 3 protocols provide network signaling and other functions. Layer 3 protocols are implemented in routers and L3 switches.

Layer 4: The transport layer in the OSI model. It provides end-to-end connections across the network. An example of a Layer 4 protocol is TCP. Below the transport layer, each protocol operates in both end stations and intermediate network nodes. Transport layer (and higher) protocols operate only in end stations. Typical Layer 4 services include guaranteed delivery of data across a network and segmentation of large messages into packets small enough to be handled by the lower-layer protocols.

Layer 5: The session layer in the OSI model. It manages communications. For example, it sets up, maintains, and terminates virtual circuits between sending and receiving stations. It also sets up boundaries for the start and end of message and establishes how messages will be sent (for example, half or full duplex, and so on).

Layer 6: The presentation layer in the OSI model. It deals with the format and representation of data that applications use; specifically, it controls the formats of screens and files. Layer 6 defines such things as syntax, control codes, special graphics, and character sets.

Layer 7: The application layer in the OSI model. It serves as the window for users and application processes to access network services. This layer applies to the actual meaning as opposed to the format of the syntax of applications (see Layer 6) and permits communication between users.

LDAP: Lightweight Directory Access Protocol. A standardized TCP/IP protocol for access to a central X.500-based directory that is shared by many different services. LDAP repositories are not capable of issuing dynamic policy authorizations and do not have policy-enforcement capabilities. LDAP can be used as a vehicle for propagating policy information from policy servers to switches. LDAP is likely to play a central role in managing dynamic networks.

LDP: Label Distribution Protocol. A fundamental concept in Multiprotocol Label Switching (MPLS) is that two label-switching routers (LSRs) must agree on the meaning of the labels used to forward traffic between and through them. This common understanding is achieved by using the LDP. LDP is the set of procedures and messages by which LSRs establish label-switched paths (LSPs) through a network by mapping network layer routing information directly to data link layer switched paths.

leaky bucket: A traffic-shaping mechanism in which only a fixed amount of traffic is admitted to the network. The traffic is leaked into the network. Excess traffic is held in a queue until it can either be accommodated or must be discarded. See also token bucket

leased line: A private, dedicated telecommunications line reserved for a single customer, often used to connect sites in a WAN. The bandwidth on a leased line depends on the service; for example, a T1 line provides 1.544 Mbps in North America, and an E1 line provides 2.048 Mbps in Europe and other countries.

LLQ: low-latency queuing. A mixture of priority queuing (PQ) and class-based weighted fair queuing (CBWFQ). It is the recommended queuing method for Voice over IP (VoIP) and IP telephony. The priority queue takes all packets that meet the definition defined by a class map and applies a policy to them.

local exchange: A local telephone company that provides telephone service within an exchange, or calling area. Also known as local carriers or local exchange carriers (LECs).

local loop: Telecommunications connection between a home or business and the local telephone exchange or central office.

M

MAC: Media Access Control. Lower of the two sublayers of the data link layer defined by the IEEE. The MAC sublayer handles access to shared media. See also MAC address

MAC address: Standardized data link layer address that is required for every port or device that connects to a LAN. Other devices in the network use these addresses to locate specific ports in the network and to create and update forwarding tables and data structures.

MAN: metropolitan-area network. The interconnection of networks in a city into a single larger network (which may then also offer efficient connection to a WAN). It is also used to mean the interconnection of several LANs by bridging them with backbone lines. The latter usage is also sometimes referred to as a campus network.

mechanism: A specific algorithm or operation (such as a queuing discipline) that is implemented in a node to realize a set of one or more per-hop behaviors.

media stream: The information content carried on a call. Refers to what is actually transmitted and received over the line, and can be read or written by a media stream application program interface (API).

MIB: Management Information Base. A database of managed objects accessed by network management protocols.

Microsoft NetMeeting: A virtual meeting application from Microsoft.

MPLS: Multiprotocol Label Switching. Architecture for integrating a mechanism for label swapping with Layer 3 routing to accelerate packet forwarding. Forwarding decisions are based on fixed-length labels inserted between the data link and network layer headers to increase forwarding performance and path-selection flexibility. The outcome is to make an IP-routed network more connection oriented in nature, where traffic is routed along a labeled path in the topology.

MPOA: Multiprotocol over ATM. An ATM forum standard for the routing of multiple protocols across an ATM substrate. It enables ATM services to be integrated with existing local-area networks (LANs) that use Ethernet, Token Ring, or TCP/IP protocols. The goal of MPOA is to allow different LANs to send packets to each other via an ATM backbone. Unlike other techniques, such as LAN emulation (LANE), which operates at Level 2 of the OSI reference model, MPOA operates at Level 3. MPOA requires a route server that manages the virtual circuits. The first version of MPOA supports only IP traffic.

MQC: modular QoS CLI. Simplifies the configuration of QoS on Cisco routers and switches by defining common command syntax and the resulting set of QoS behaviors across platforms.

N

NBAR: Network-Based Application Recognition. A classification engine that can recognize network applications that dynamically assign TCP or UDP port numbers. After the application is recognized, the network can invoke specific services for that particular application.

network element: A networking device, such as a router, a switch, or a hub, where resource-allocation decisions have to be made and the decisions have to be enforced. The element is therefore potentially capable of exercising QoS control over data flowing through it.

Network Management Forum: A nonprofit, global consortium of service providers, equipment suppliers, and software developers for improving the management of public networks and services.

NFS: Network File System. A commonly used distributed file system that allows remote file access across a network.

O

OID: object identifier. A unique designator of attributes of managed resources used in SNMP network management. An OID is a unique number that defines an object in the MIB. The numbers that comprise the OID describe a path to the object through a tree hierarchy, which is often called an OID tree or a registration tree. When the SNMP agent wants to access a specific object, it traverses the OID tree in the MIB file to find the object.

OSI: Open Systems Interconnection. In the 1980s, the European-dominated International Organization for Standardization (ISO) began to develop its (OSI) networking suite. OSI has two major components: an abstract model of networking (the basic reference model, or seven-layer model, see Layers 1 through 7) and a set of concrete protocols.

OSS: operations support system. Traditionally, a term used by telecommunications companies to manage, monitor, and control the telecommunications networks. Billing, customer-care systems, directory services, network element, and network management are all operational support systems. Now also applied to enterprise environments to refer to the interaction of multiple management systems to manage the overall delivery of applications.

overprovisioning: A way to address current limitations of best-effort networks by provisioning more bandwidth than expected network peak requirements. Overprovisioning increases the probability, but does not guarantee the quality, of transmission of time-sensitive and bandwidth-intensive applications. Overprovisioning is most costly for the WAN.

P

packet classification: A methodology for organizing packets into a group useful for QoS. Classification may be done over a range of granularities, from groups of aggregated flows to individual flows or even subflows. Typically, classification is done in a way similar to defining access lists (that is, based on some contents of the packet header). In this case, a packet may be classified by information in the L2, L3, or L4 headers (source/destination addresses, port numbers, subarea address, applications, user, as well as various Layer 2 attributes). Classification can also be done based on information within the packet payload. Classifications can be broad for aggregated flows, such as traffic destined for a subnetwork X, or as narrow as a single flow or even subflow.

patch: A small addition to the original software code, written to bypass or correct a problem, and also provided between software releases.

PDP: policy decision point. The point where policy decisions are made. Responsible for handling policy decisions on behalf of policy-enforcement points (PEPs).

peering agreement: A reciprocal agreement that lets Internet service providers (ISPs) share backbone links so traffic can reach destinations beyond the ISP's management domain. Agreements between network service providers to allow traffic to transit each other's networks.

PEP: policy-enforcement point. A port on a network device where the policy decisions are actually enforced.

Performance Monitor: A Microsoft Windows server application that displays server activities in real time.

PHB: per-hop behavior. The forwarding treatment given to a specific class of traffic, based on criteria defined in the DiffServ field. Routers and switches use PHBs to determine priorities for servicing various traffic flows.

PHB group: A set of one or more PHBs that can only be meaningfully specified and implemented simultaneously, due to a common constraint applying to all PHBs in the set, such as a queue-servicing or queue-management policy. A PHB group provides a service building block that allows a set of related forwarding behaviors to be specified together (such as four dropping priorities). A single PHB is a special case of a PHB group.

policing: Packet-by-packet monitoring function at a network border (ingress point) that ensures a host (or peer or aggregate) does not violate its promised traffic characteristics. Policing means limiting the amount of traffic flowing into or out of a particular interface to achieve a specific policy goal. Policing typically refers to actions taken by the network to monitor and control traffic to protect network resources, such as bandwidth, against unintended or malicious behavior. Traffic shaping may be used to achieve policing goals or to do congestion management.

policy: The combination of rules and services where rules define the criteria for resource access and usage in order to manage the bandwidth made available to specified traffic. A policy dictates a number of conditions that must be met before a specified action can be taken.

policy control: The application of rules to determine whether access to a particular resource should be granted.

policy domain: The part of a network subject to policy. Policy applies to one domain, and policy domains do not overlap. Note that a policy domain is not the same as a DNS domain or Windows NT domain.

policy element: Subdivision of policy objects; contains single units of information necessary for the evaluation of policy rules. A single policy element carries user or application identification, whereas another policy element may carry user credentials or credit card information. Examples of policy elements include identity of the requesting user or application, user/app credentials, and so forth. The policy elements themselves are expected to be independent of which QoS signaling protocol is used.

policy object: Contains policy-related info, such as policy elements, and is carried in a request or response related to a resource-allocation decision.

policy resolution: When a policy server or other policy decision point is attempting to apply a policy rule to address a particular situation, it may find a number of possible policies for which all the matching criteria fit. In such cases, additional criteria should be introduced, such as rule ordering or rule specificity, to provide an unambiguous answer. Policy resolution may be user assisted or automated.

policy server: A server that authorizes QoS requests received from routers or gateways and coordinates bandwidth usage on multiple network devices to ensure consistent end-to-end service throughout the data-path. A policy server ensures that packets receive the appropriate QoS, based on a set of policies defined by the network administrator. (Cisco QPM and QDM are examples of policy servers.)

premium service: In DiffServ terms, premium service is a peak-limited, extremely low-delay service, resembling a leased line. At the network edge, where a premium class is first created, it must be either shaped or policed to a rate with no more than a two-packet burst. A policer for premium service is set to drop packets that exceed the configured peak rate. For this service, the peak rate of the premium class aggregate across any boundary must be specified, and the rate must be smaller than the link capacity.

profile: The bandwidth and burst requirements for a given class of service (or application), either at the source site or between a source and destination site.

provisioned service: A service for which network resources are allocated ahead of time, in accordance with a service contract.

PVC: permanent virtual circuit. Exhibits the characteristics of a leased-line connection over a packet-switched, connection-oriented network and remains established.

Q

QoS: quality of service. A collective measure of the level of service delivered to the customer. QoS can be characterized by several basic performance criteria, including availability (low downtime), error performance, response time and throughput, lost calls or transmissions due to network congestion, connection setup time, and speed of fault detection and correction. Service providers may guarantee a particular level of QoS (defined by a service level agreement, SLA) to their subscribers. QoS-enabled hardware and software solutions sort and classify IP packet requests into different traffic classes and allocate the proper resources to direct traffic based on various criteria including application type, user or application ID, source or destination IP address, time of day, and other user-specified variables.

QoS policy: A set of actions a network takes to configure and signal for a particular QoS to be provided to a particular traffic classification.

QoS signaling: The means for transmitting QoS requests and parameters between devices or applications to deliver a QoS requirement across the network. Either in-band signaling (for example, IP Precedence or 802.1p) or out-of-band signaling (RSVP) is used to indicate that a particular QoS is desired for a particular traffic classification. IP Precedence and RSVP are the two most useful signaling mechanisms, because they take advantage of the end-to-end nature of Layer 3 protocol and the growing ubiquity of IP.

QoS strictness: Three broad categories of QoS from end to end or edge to edge, are best effort, differentiated, and guaranteed. The strictness of the QoS describes how tightly the service can be bound by specific bandwidth, delay, jitter, and loss characteristics. For example, the delay, loss, and jitter characteristics can be offered to within tight tolerances on a constant bit rate (CBR) service, whereas they are much harder to bound on a typical Internet IP connection.

QPM: QoS policy manager. Delivers differentiated services across network infrastructures for converged voice, video, and data applications. Network administrators can use QPM as a complete system for centralized QoS monitoring, policy control, and automated reliable policy deployment.

queuing: A method for metering the flow of traffic by placing packets in holding queues and retransmitting them according to a sorting algorithm, typically a simple first-in, first-out (FIFO) formula. Queues of different sizes can be used to assign levels of importance according to class of service (CoS) designations. Queues that overflow, typically discard packets to reduce network congestion.

R

RAPI: RSVP application programming interface. An RSVP-specific API that enables applications to interface explicitly with an RSVP process.

RED: random early detection. A congestion-avoidance algorithm developed in the early 1990s and built on the base-level TCP behavior of automatically slowing transmissions when packet loss is detected. RED tries to anticipate congestion by monitoring a queue on a router. When the specified threshold is reached, it randomly discards packets. This is an implicit signal that the originating applications should slow their transmissions before congestion becomes severe. Unlike CBQ, WFQ, and SFQ, RED does not require flow state in routers.

reservation: Part of a resource that has been dedicated for the use of a particular traffic type for a period of time through the application of policies.

resource: Something of value in a network infrastructure to which rules or policy criteria are first applied before access is granted. Examples of resources include the buffers in a router and bandwidth on an interface.

round-robin queuing: An algorithm that services each queue in a predefined sequence. For example, it might empty 1500 bytes apiece from queue 1 (high priority), queue 2 (medium priority), and queue 3 (low priority), servicing each in turn.

RSPAN: Remote SPAN. A feature of Cisco switches that extends the monitoring capabilities of an existing network analyzer to remotely monitor source ports and destination ports that are distributed across multiple switches, enabling you to monitor the traffic on one switch using a device on another switch. See also SPAN

RSVP: Resource Reservation Protocol. An IETF standard that allows an end device and a network to negotiate specific QoS characteristics. Using RSVP, an application signals a request to reserve resources along a route from source to destination. RSVP-enabled routers then schedule and prioritize packets. A reservation for the required bandwidth is allowed or denied depending on the current network conditions.

RTCP: Real-Time Control Protocol. Monitors the QoS of an IPv6 RTP connection and conveys information about the ongoing session. See also RTP, IPv6, QoS

RTP: Real-Time Transport Protocol. A network protocol used to carry packetized audio and video traffic over an IP network. See also cRTP

S

SAA: Service Assurance Agent. Embedded software within Cisco IOS devices that performs active monitoring.

segment: 1. Section of a network that is bounded by bridges, routers, or switches. 2. In a LAN that uses a bus topology, a continuous electrical circuit that is often connected to other such segments with repeaters. 3. Term used in the TCP specification to describe a single transport layer unit of information. The terms datagram, frame, message, and packet are also used to describe logical information groupings at various layers of the OSI reference model and in various technology circles.

server: 1. Node or software program that provides services to clients. 2. In network addressing, a concentrator, data switch, or host computer being accessed.

service: A description of the overall treatment of (or a subset of) a customer's traffic across a particular domain, path, or end-to-end network. In DiffServ, service descriptions are covered by administrative policy, and services are constructed by applying traffic conditioning to create behavior aggregates that experience a known per-hop-behavior (PHB) at each node within the DiffServ domain. Multiple services can be supported by a single PHB in concert with a range of traffic conditioners.

service provider: Controls a network infrastructure and may be responsible for the charging and accounting of services.

SFQ: stochastic fair queuing. A hash function used to map flow to one of a set of queues.

signaling: Communications between devices to set up calls and tear them down.

SLA: service level agreement. A contract between a service provider and customer defining provider responsibilities in terms of network levels (for example, throughput, loss rate, delays, and jitter).

SNMP: Simple Network Management Protocol. A UDP-based network management protocol used predominantly in TCP/IP networks. SNMP can be used to monitor, poll, and control network devices.

soft QoS (qualitative QoS): A quality of service that does not consist of a 100-percent guarantee of one of the performance parameters (bandwidth, latency, jitter, and so forth), but delivers that performance with some percentage probability and, in general, gives better than best-effort service.

SONET: Synchronous Optical Network. North American transport standard for optical networks with speeds from approximately 52 Mbps (OC-1) to 10 Gbps (OC-192).

SPAN: Switched Port Analyzer. Feature of the Cisco switches that extends the monitoring capabilities of existing network analyzers into a switched Ethernet environment. SPAN mirrors the traffic at one switched segment onto a predefined SPAN port. A network analyzer attached to the SPAN port can monitor traffic from any of the other catalyst switched ports. See also RSPAN

standalone router: A router that is considered in isolation from all other routers in the network. The standalone router is seen as the only router on the WAN, so it can be used for test or demonstration purposes.

static policy: A policy put in place as a service is provisioned and then not changed dynamically by changing application conditions.

Subnet Bandwidth Manager: An IETF standard for handling resource reservations on shared and switched IEEE 802-style local-area media.

SVC: switched virtual circuit. Virtual connection set up only for the duration of a single communications session. In contrast, a permanent virtual circuit (PVC) remains available at all times.

T

TCA: traffic-conditioning agreement. An agreement specifying classifier rules and any corresponding traffic profiles and metering, marking, discarding, and/or shaping rules that are to apply to the traffic streams selected by the classifier. A TCA encompasses all the traffic-conditioning rules explicitly specified within an SLA, along with all the rules implicit from the relevant service requirements.

TCP rate control: A technology implemented at network end points that attempts to regulate the introduction of traffic into the network.

throughput: In data transmission, throughput is the amount of data moved from one place to another in a given time period.

token bucket: A traffic-shaping mechanism in which a predetermined amount of tokens in a bucket represent the capacity allowed to each class of traffic. Packets are forwarded until they exhaust their supply of tokens. When the token supply is exhausted, packets may be discarded or delayed until the bucket is replenished. This controls the transmit rate and accommodates bursty traffic. In some systems, a customer's token supply might correspond to a service fee. See also leaky bucket

ToS: type of service. A field within an IP header that can be used by the device originating the packet, or by an intermediate networking device, to signal a request for a specific QoS level. ToS uses 3 bits to tell a router how to prioritize a packet, and 1 bit apiece to signal requirements for delay, throughput, and reliability. ToS is also known as IP precedence bit format and the IP Precedence field.

traffic conditioner: An entity that performs traffic-conditioning functions, and which may contain meters, policers, shapers, and markers. Traffic conditioners are typically deployed in boundary nodes, not in internal nodes of a domain.

traffic conditioning: Control functions that can be applied to a behavior aggregate, application flow, or other operationally useful subset of traffic, such as routing updates. These may include metering, policing, shaping, and packet marking. Traffic conditioning is used to enforce agreements between domains and to condition traffic to receive a differentiated service within a domain by marking packets and by monitoring and altering the temporal characteristics of the aggregate where necessary.

traffic profile: A description of the temporal properties of a traffic stream such as rate and burst size.

traffic shaping: A group of techniques that attempts to regulate or meter the flow of packets through the network. See also leaky bucket, token bucket

traffic stream: An administratively significant set of one or more microflows that traverse a path segment. A traffic stream may consist of the set of active microflows that are selected by a particular classifier.

V

VC: virtual circuit. A logical connection between two network nodes that acts as though it is a direct physical connection, even though it may physically be packet based. The term is used most frequently to describe connections between two hosts in a packet-switching network.

VLAN: virtual local-area network. A networking architecture that allows end systems on topologically disconnected subnetworks to appear to be connected on the same LAN. Similar in functionality to bridging.

VoIP: Voice over IP. A technology for transporting integrated digital voice, video, and data over IP networks.

VPN: virtual private network. A private data network that makes use of the public telecommunication infrastructure, maintaining privacy through the use of tunneling protocol and security procedures. For example, a company could contract with an ISP to set up a VPN to use the Internet to connect two geographically separated sites, rather than set up a dedicated WAN or use a leased line.

W

WDM: wavelength-division multiplexing. A mechanism to allow multiple signals to be encoded into multiple wavelengths.

WFQ: weighted fair queuing. Per-flow packet scheduling in network elements that automatically categorizes traffic flows into high and low priority, based on volume of packets seen by a router or switch. Low-bandwidth traffic has effective priority over high-bandwidth traffic, and high-bandwidth traffic shares the transmission medium proportionally according to assigned weights. Like class-based queuing (CBQ), WFQ is designed to prevent any one traffic type from entirely eclipsing another. By default, WFQ favors lower-volume traffic flows over higher-volume ones (for example, a routine e-mail over a large FTP download).

WLAN: wireless local-area network. WLANs use radio frequency to transmit and receive data over the air.

WRED: weighted random early detection. A congestion-avoidance and QoS mechanism for TCP/IP-based networks.

X

XML: eXtensible Markup Language. A standard maintained by the World Wide Web Consortium (W3C). It defines a syntax that enables you to create markup languages to specify information structures. Information structures define the type of information (such as subscriber name or address), not how the information looks (bold, italic, and so on).

X terminal: Terminal that enables a user to simultaneously access several different applications and resources in a multivendor environment through implementation of X Windows.

Reference Used in Glossary

Cisco Systems, Internetworking Terms and Acronyms. http://www.cisco.com/univercd/cc/td/doc/cisintwk/ita/index.htm