The ATM Cell Header

The ATM cells can be found in one of two formats, depending on whether the endpoints are a UNI or an NNI connection. The two differ in one way: The NNI header does not contain a Generic Flow Control (GFC) field. The NNI header has a Virtual Path Identifier (VPI) that occupies the entire first 12 bits. A cell header for a UNI cell is shown in Figure 8.3.

Figure 8.3: An ATM UNI cell header.

Let’s take a look at the fields in a cell header:

∙Generic Flow Control (GFC)—An 8−bit field that is used to provide information to identify multiple stations that share a single ATM interface. The GFC is typically not used.

∙Virtual Path Identifier (VPI)—An 8−bit field used with the VCI to identify the next destination of a cell as it passes through a series of ATM switches on its way to its destination.

∙Virtual Channel Identifier (VCI)—An 8−bit field used in conjunction with the VPI to identify a cell’s next destination.

∙Payload Type (PT)—A 3−bit field that indicates whether the cell contains user data or control data. If the cell contains user data, the second bit in the user data indicates congestion, and the third bit indicates whether the cell is the last in a series of cells that represent a single AAL5 frame.

∙Congestion Loss Priority (CLP)—A 1−bit field that indicates whether the cell should be discarded if it encounters extreme congestion as it moves through the network.

∙Header Error Check (HEC)—An 8−bit field that indicates a checksum calculated only on the header itself.

The ATM Switch and ATM Endpoints

ATM networks use one of two types of devices for each end of the network: ATM switches and ATM endpoints. An ATM endpoint is a device that has an ATM network interface adapter, such as a workstation, router, Data Service Unit (DSU), or LAN switch. These devices in turn transmit data to an ATM switch, which is responsible for receiving this data, updating the header information, and then sending the data out the proper interface port to its intended destination.

As mentioned earlier in the chapter, in a UNI, the ATM interface connects an endpoint to a switch. In an NNI, the interface connects two ATM switches together. The UNI and NNI connections can be used to further divide the network into private and public networks. As the name implies, a private network connects the ATM endpoint to a private network, whereas a public network connects an ATM endpoint to a public switch (possibly owned by a telephone company or other WAN service provider).

The ATM Reference Model

The ATM standard uses a reference model to describe the functions of the protocol. The ATM Reference Model has three layers (which roughly correspond to those in the OSI model) and three management planes. The ATM Physical layer is similar to the Physical layer of the OSI model. The Physical layer controls the transmission and receipt of bits on the physical medium.

The ATM layer and the ATM Adaptation layer (AAL) are similar to the Data Link layer of the OSI chart. The

164

ATM layer is responsible for establishing connections and passing cells through the ATM network. The ATM Adaptation layer translates the different types of network traffic. Four AALs are defined, but only three are actively in use:

∙AAL1—Used to transport timing−dependent traffic such as voice

∙AAL3/4—Used by network service providers in Switched Multimegabit Data Service (SMDS) networks

∙AA5—The primary AAL used for non−SMDS traffic that doesn’t require the pacing AAL1 would provide

Figure 8.4 shows the mapping of the ATM Reference Model compared to the OSI Reference Model.

Figure 8.4: The ATM Reference Model layers compared to OSI Reference Model layers.

The Physical Layer

The ATM Physical layer controls transmission and receipt of bits on the physical media. This layer also tracks the ATM cell boundaries and packages cells into the appropriate frame type. This layer is divided into two sublayers: the physical medium dependent (PMD) sublayer and the transmission convergence (TC) sublayer.

The PMD sublayer is responsible for sending and receiving a continuous flow of bits with the timing information to synchronize the transmission and reception of data. ATM does not care about the physical media being used, and all widely used physical topologies are capable of supporting ATM cells. Existing high−speed topologies capable of supporting ATM cells include Synchronous Optical Network (SONET), DS3/E3, Fiber Distributed Data Interface (FDDI), and unshielded twisted pair (UTP).

The TC sublayer is used to maintain the ATM cell boundaries, verify the validity of data, maintain data synchronization, create and check header error control to ensure valid data, and put the cells into a format that the physical media can use. This sublayer also extracts and inserts ATM cells within either a Plesiochronous Digital Hierarchy (PDH) or a Synchronous Digital Hierarchy (SDH) Time Division Multiplexed (TDM) frame and passes this frame to and from the ATM layer.

The ATM Layer

The ATM layer is responsible for establishing connections, passing cells to and from the AAL, inserting the ATM header, and extracting the ATM header. This layer is also responsible for multiplexing and demultiplexing data through the ATM network. To do this, ATM uses information contained in the header of each ATM cell. ATM switches use a VPI and a VCI field inside the ATM cell header to identify the next network segment a cell needs to transit on its way to its final destination.

A VCI is also known as a virtual channel. It is an identifier for the physical connection between the two ends that form a logical connection. A VPI is the identifier for a group of VCIs that allows an ATM switch to perform operations on a group of virtual connections (VCs).

165

The ATM Adaptation Layer

The ATM Adaptation Layer (AAL) provides the translation between the larger service data units of the upper layers of the OSI Reference Model and ATM cells. It works by receiving packets from the upper−level protocols and breaking them into 48−byte segments to be dumped into the payload of an ATM cell. The AAL has two sublayers: segmentation and reassembly (SAR) and the convergence sublayer (CS). The CS has further sublayers: the common part (CP) and the service specific (SS). Like protocols specified in the OSI Reference Model, Protocol Data Units (PDUs) are used to pass information between these layers.

The AAL translates between the different types of network traffic—such as video streams, data packets, and voice packets—of upper−layer processes and ATM cells. In other words, the AAL receives packets from upper−level protocols and breaks them into the 48−byte segments that form the payload field of an ATM cell. Several types of AAL standards are defined for this layer. Which AAL you use will largely depend on the type of traffic and what you are trying to do with the traffic. The characteristics of each AAL are as follows:

∙AAL1 (Class A)—This layer is a connection−oriented service that provides end−to−end timing provisions. It maintains a constant data transfer rate, which is used for transporting telephone traffic and uncompressed video traffic. This is known as a constant bit rate (CBR) service. It is appropriate to use AAL1 to transport voice and video traffic or another type of timing−sensitive data.

∙AAL2 (Class B)—This layer is reserved for data traffic that requires variable bit rates (VBR) and timing sensitivity, such as multimedia. It multiplexes short packets from multiple sources into a single cell with end−to−end timing and connection orientation.

∙AAL3/4 (Class C)—This layer was designed for network service providers; it closely aligns with SMDS. This layer uses no VBR and has no timing requirements. It supports both connection−oriented and connectionless data for WAN links using Frame Relay or X.25. This layer is perfectly suited for use in environments that need to send or receive large files. AAL3 is identical to AAL4, with the exception that the AAL3 layer is connection−oriented only, whereas AAL4 is both connection−oriented and connectionless.

∙AAL5 (Class D)—This layer is the primary AAL used to transfer non−SMDS data. It supports both connection−oriented and connectionless data. This layer is used for such applications as classical IP (CLIP) over ATM and LANE. Catalyst switches use this layer to provide LANE services for ATM.

ATM networks provide the transport method for several different independent emulated LANs. When a device is attached to one of these emulated LANs, its physical location no longer matters to the administrator or implementation. This process allows you to connect several LANs in different locations with switches to create one large emulated LAN. This arrangement can make a big difference, because attached devices can now be moved easily between emulated LANs. Thus, an engineering group can belong to one ELAN and a design group can belong to another ELAN, without the groups ever residing in the same location.

LANE also provides translation between multiple−media environments, allowing data sharing. Thus, Token Ring or FDDI networks can share data with Ethernet networks as if they were part of the same network.

Specifying ATM Connections

ATM networks manage traffic by establishing and configuring each connection. When establishing the connection, the connection type and the resources required to support the connection are specified as a class of service. This class is used to provide users with a guaranteed QoS.

The classes of service for QoS are defined by the ATM Forum and are as follows:

∙Available bit rate (ABR)—Supports variable−rate data transmissions without preserving any timing relationships between the source and destination nodes. This connection type provides for a best effort service above a specified minimum cell rate (MCR).

∙Constant bit rate (CBR)—Typically used to carry constant rate traffic and represented by fixed

timing. CBR is typically used for circuit emulation, uncompressed voice, and multimedia data traffic.

166