Microsoft Encyclopedia of Networking
ISBN: 0735613788
EAN: 2147483647
EAN: 2147483647
Year: 2000
Pages: 37
Pages: 37
Authors: Mitch Tulloch, Ingrid Tulloch
See
A Microsoft Windows 2000 service that can be used to centrally designate when, how, and by whom shared network segment resources will be used.
How It Works
QoS Admission Control Service (QoS ACS) is based on the Subnet Bandwidth Management (SBM) specification defined by the Internet Engineering Task Force (IETF). QoS ACS operates at the network layer and can service all transport protocols in the TCP/IP protocol suite, including Remote Display Protocol (RDP), User Datagram Protocol (UDP), and Transmission Control Protocol (TCP). A QoS ACS host (a server running Windows 2000 with the QoS ACS service installed and configured) uses the Resource Reservation Protocol (RSVP) as a message service for sending and receiving priority bandwidth requests.
A QoS ACS host controls the bandwidth for the subnet to which it is connected. The QoS ACS host uses multicasting to send out messages called beacons to inform clients on the subnet that it is ready to receive bandwidth allocation requests. Clients on the subnet that desire access to shared network resources (such as multimedia servers) first submit their bandwidth request to the QoS ACS server so that it can determine whether sufficient bandwidth is available to allocate to the clients. Bandwidth is then allocated based on the current state of resource and bandwidth availability on the subnet and the QoS ACS policy rights of the requesting user. These policy rights are defined in Active Directory.
A client’s request for bandwidth will be rejected if the QoS ACS host determines that the user does not have the right to reserve bandwidth on the subnetwork or the subnetwork does not have sufficient resources to support the request at that time. If the client’s request is rejected, the client must decide whether to try accessing the resource using a best-effort service level or wait until later when priority bandwidth becomes available and can be allocated to the client. If the request is approved, the QoS ACS host logically allocates the requested bandwidth and forwards the client’s resource request to the appropriate server on the network.
No configuration is required for Windows 2000 or Windows 98 clients, and non-Windows clients can request bandwidth, provided they are running suitable SBM client software.
Also called Series Q protocols, a set of protocols developed by the International Telecommunication Union (ITU) that includes the following:
Q.920/921: Specifies the User-Network Interface (UNI) data-link layer for Integrated Services Digital Network (ISDN), including the ISDN D channel’s Link Access Procedure D (LAP-D) data-link protocol.
Q.922A: Specifies the ITU encapsulation method for frame relay networks.
Q.931: Specifies mechanisms for establishing, maintaining, and tearing down ISDN connections. This includes working with connections in the LAP-D data-link protocol, which runs on the ISDN D channel.
Q.93B: The Asynchronous Transfer Mode (ATM) signaling protocol, which specifies mechanisms for establishing, maintaining, and tearing down broadband ISDN (B-ISDN) connections.
Generally, any networking technology that has predictable latency and data loss. More specifically, any mechanism that allows either absolute or relative performance requirements to be defined for different traffic streams carried over a network. In other words, a quality of service (QoS) network can guarantee a certain level of throughput for a specific path, connection, or type of traffic. This makes it possible to ensure that critical network applications receive priority handling.
How It Works
Networks that support QoS mechanisms can generally control both the quality of a network transmission and the availability of bandwidth to ensure this quality. These are different from ordinary networks, which guarantee only a best-effort delivery—traffic flow cannot be controlled and bandwidth cannot be reserved. If traffic congestion occurs during a period of intense network communication, QoS features can kick in to ensure that certain data streams receive preference for users and applications that need consistent data flow. For example, networks carrying real-time audio or video require a high level of QoS to ensure that reception is smooth and free of errors. (Latency in delivery of packets for a real-time multimedia stream can produce pauses and dropouts that are highly undesirable from the user’s point of view.)
You can control the following network properties in a network that supports QoS functions:
Throughput (total bandwidth used)
Latency (traffic delay)
Priority (among types of traffic)
Peak traffic, burstiness, and jitter (to smooth traffic flow)
Packet or cell loss and retransmission
Asynchronous Transfer Mode (ATM) is one networking technology that can deliver data at specific levels of QoS. A look at the QoS features in ATM will give you an idea of the scope of issues that QoS can address. The QoS mechanisms of ATM function in the following areas:
Traffic contracts: Specify QoS parameters for a data stream, such as peak and average bandwidth and maximum burst size. Each ATM end node in an ATM network negotiates a traffic contract with that network. QoS parameters can be defined for the end node to control the type and amount of traffic that the host can send over an ATM circuit. If an ATM switch in the circuit cannot meet these requirements, it sends a message to the transmitting host indicating that the connection is refused. Each virtual channel or virtual path that is established by an ATM end node in an ATM network has its own traffic contract.
Traffic shaping: Involves ATM devices using data queues to smooth out traffic bursts and limit the peak data rate so that the conditions of the negotiated traffic contracts are met.
Traffic policing: ATM switches enforce traffic contracts so that the agreed-upon QoS conditions of each data stream are met. If these conditions are violated, ATM switches set the cell-loss priority (CLP) bit of each offending cell, which indicates that the offending cells might be dropped when congestion occurs at the switch.
The specific QoS parameters that can be negotiated in an ATM network are called traffic parameters. They include the following:
Cell Delay Variation (CDV): The tolerance per second for cells that should not exceed the peak cell rate. CDV is specified at the network, not at the end station. Cells within the CDV are accepted.
Maximum Burst Size (MBS): The maximum number of cells per burst.
Peak Cell Rate (PCR): The maximum number of cells that can be transmitted per second based on the specified peak bandwidth per second.
Sustainable Cell Rate (SCR): The maximum number of cells that can be transmitted per second based on the specified sustainable bandwidth per second.
The implementation of QoS in ATM allows ATM networks to support four classes of QoS, each with scalable levels:
Constant Bit Rate (CBR): The traffic requires a guaranteed rate of transport and does not tolerate cell loss. The ATM end station informs the ATM network of the required QoS parameters at call setup. The network then performs admission control by reserving the necessary bandwidth or refusing the connection. The end station is responsible for complying with the agreed-upon peak data rate—if it exceeds this rate, the network drops the offending cells.
Variable Bit Rate (VBR): Similar to CBR except that it negotiates a maximum burst size and maximum sustainable rate in addition to a peak data rate.
Unspecified Bit Rate (UBR): Does not involve reserving bandwidth or establishing cell-loss ratios. UBR is commonly used in LAN emulation (IP-over-ATM networks that are typically used in campus backbones). If one cell is dropped because of congestion, the entire Internet Protocol (IP) packet to which the cell belongs must be retransmitted.
Available Bit Rate (ABR): Uses periodic polling of the ATM network to adjust the data transmission rate. ABR is also used in LAN emulation (LANE) implementations.
NOTE
The QoS features of ATM give it an advantage over competing technologies such as Fast Ethernet and frame relay, but ATM is more difficult to implement because of its more complex architecture. However, you can use ATM backbone technology to backfill QoS features into connected Fast Ethernet and frame relay network structures, albeit awkwardly. Gigabit Ethernet also includes QoS features.
In best-effort TCP/IP internetworks such as the Internet, QoS in IP data streams is difficult to achieve. The Internet Engineering Task Force (IETF) has proposed two new mechanisms for providing QoS for the Internet and for TCP/IP internetworks: integrated services and differentiated services.
Integrated services focuses on making it reasonable to use TCP/IP internetworks for a variety of services, such as audio, video, real-time data, and classical data traffic. It regulates the flow of data based on computations made by the router. It also makes use of the Resource Reservation Protocol (RSVP), which can be used to reserve circuits to maintain an even flow of data. Controlled-load service, another aspect of integrated services, ensures that priority data keeps moving even when the router reaches its capacity. Because integrated services requires new routers, it would require considerable rebuilding of the core technology of the Internet.
Differentiated services aggregates IP traffic into three classes of priority serviced by routers. This would enable ISPs to offer varying levels of priority to their customers. Differentiated services uses the DS field in the IPv4 packet header to define how the packet will be forwarded (the “per-hop behaviors” of IP traffic). This solution would mean minor changes to the Internet’s backbone routers and would leave the complexity of implementing QoS features to the edges of the network, especially the signaling methods for establishing QoS-enabled links between nodes on the network.
A signaling method used in 100VG-AnyLan (100BaseVG) networks. Signals are transmitted over all four pairs of wire in voice-grade unshielded twisted-pair (UTP) cabling. Quartet signaling makes possible the transmission of data at a speed of 100 Mbps while using the same transmission frequencies that are used on standard 10BaseT networks. Quartet signaling also enables 100VG-AnyLan to leverage existing installations of category 3, 4, and 5 UTP cabling for 100-Mbps transmission.
How It Works
10BaseT Ethernet networks use only two pairs of wires in a four-pair UTP cable—one pair for transmitting data and the other pair for receiving data and for detecting collisions on the network. 100VG-AnyLan uses a demand priority method for controlling access to the media, which prevents collisions from occurring. As a result, 100VG-AnyLan can use all four pairs of wires for data transmission—hence the term “quartet signaling.” In addition, quartet signaling uses a different line coding technique than the traditional Manchester coding method used in Ethernet networks. Quartet signaling uses the 5B/6B NRZ line coding method, while Manchester coding uses a 1B/2B scheme whereby 1 bit of data is encoded using two binary symbols. The 1B/2B algorithm is reliable and simple to implement but inefficient. The 5B/6B method encodes 5 bits using six binary symbols, which allows two and a half times as much information to be transmitted per wire compared to 10BaseT Ethernet (as shown in the following table).
Calculations for Line Coding Data Rates
| Manchester Coding | Quartet Signaling | |
| Line coding | 1B/2B | 5B/6B |
| Line frequency | 20 MHz | 30 MHz |
| Data rate per pair | (1/2) x 20 = 10 Mbps | (5/6) x 30 = 25 Mbps |
| Number of pairs used | 1 | 4 |
| Total data rate | 1 x 10 = 10 Mbps | 4 x 25 = 100 Mbps |
NOTE
The 100BaseT4 form of Fast Ethernet also uses all four pairs of wire in twisted-pair cabling.
See also 100VG-AnyLan, demand priority
A collection of items waiting to be processed in a specific order. Examples of queues in computer and networking technology are numerous and include the following:
A print queue, which consists of print jobs waiting to be sent to a print device
A messaging queue (on a mail server such as Microsoft Exchange Server), which consists of messages waiting to be sent
A backlog of packets waiting to be forwarded over a specific interface by a router
Information, function calls, or transactions sent by one application and forwarded to another by Microsoft Message Queue (MSMQ) Server in Microsoft Windows NT or Message Queuing in Windows 2000
A collection of fax messages waiting to be processed and sent by a fax server
A series of system messages, such as key presses and mouse clicks, sent by applications to an operating system for processing
A resource used in Microsoft clustering services—Microsoft Cluster Server (MSCS) for Microsoft Windows NT Server 4.0, Enterprise Edition, and Windows Clustering for Windows 2000—that guarantees that data necessary for recovery is maintained consistently among all members of a cluster. The quorum resource must be a physical disk and can be owned by only one node in a cluster. The quorum resource stores the cluster log, which is maintained by the clustering software. If a node fails and then recovers, the quorum resource is responsible for updating that node’s configuration information so that it matches all other nodes. If the nodes in a cluster cannot communicate with each other, the quorum resource determines which node can continue operating. If the node owning the quorum resource fails, another node takes ownership of the resource. Each cluster can have only one quorum resource.