Atm Cell Frame Format

Published Date: 02 Nov 2017

Abstract

ATM stands for Asynchronous Transfer Mode and is a kind of asynchronous connection oriented packet switching technique by ITU-T for transmitting data in the form of fixed sized packets known as Cell. This standard allows transmission of multiple types of data including video, voice and images.

This paper presents information about ATM with regard to its layered model architecture, cell format and cell transmission from source to destination.

Introduction

Packet switching technique has provided better channel utilization as compared to circuit switching. The two popular standards X.25 and Frame Relay have not only allowed large number of users but also better transmission rates. But as newer and faster technologies are emerging in terms of network devices and transmission mediums, both of them are proved to be underachiever. Today we need such transmission technology which not only provide fast data transfer but also versatile in carrying all types of data. And ATM can be the answer for this.

ATM is similar to X.25 and Frame Relay in many contexts like, it divides the data into packets, but there are usually of the same size and are commonly identified as cell. Besides this, ATM also establishes a logical connection between sender and receiver. Like Frame Relay, it provides end to end error correction and mainly concerns with the lowest two layers that is Physical and Data Link Layer of OSI model.

ATM was introduced in the decade of 1980s and is still used in many WAN technologies. It was designed in order to gain good point of both telephone network and Internet.

Architecture

The diagram below maps the architecture of ATM model to OSI model [1]. This three dimensional model not only show the layers but also the planes namely, User plane, Control plane and Management plane. The physical layer as usual is concerned with the transmission of the data and encoding schemes. Physical layer can be able to transmit data between 25 Mbps and 622 Mbps; however, variant in the transmission rate can be achieved.

CT842707.jpgThe functions of each plane can easily be identified from their names, such as user plane handles the job for data transmission, control plane look after the connection issues and management plane concerns with the coordination of different layers. The main job is done at the data link layer, which in ATM architecture is split into two layers namely ATM layer and ATM adaption layer.

ATM adaptation layer can be further divided into two sub layers, known as CS (Convergence Sub layer) and SAR (Segmentation and Reassembly Sub layer). The main function of ATM adaption layer is to map higher layer data into ATM cells while that of ATM layer are transmission control, switching, congestion control, Cell header processing, etc.

ATM Cell Frame Format

In ATM, as mentioned earlier, the information is transmitted in the form of fixed size packets known as Cells. The length of each cell is of 53 bytes, wherein 5 bytes are for header and remaining 48 bytes are reserved for payload. The main concept behind fixed sized cell is to make ATM switch design simpler and high speed data transmission can be achieved.

The diagram below is the cell format of ATM for User-network interface (a) and Network-network interface (b)

http://ars.els-cdn.com/content/image/1-s2.0-S0140366498001558-gr1.gif

GFC (Generic Flow Control): It is a 4 bit field which is used for local flow control between the network access point, e.g. switch and one or more end stations. This field is only a part of header user-network interface and not of network-network interface.

VPI (Virtual Path Identifier): In user-network interface this field is of 8 bits whereas in network-network this field can be extend up to 12 bits as GFC is absent in this interface. This field is used for internal network routing purpose.

VCI (Virtual Channel Identifier): This 16 bit field is used for end to end routing.

PT (Payload Type): It is a three bit field contains the information regarding the payload field. Different combinations of 0s and 1s in this field suggest different types. For example, 0 indicates user information if it is at first position, second bit suggest congestion, etc.

CLP (Cell Loss Priority): This one bit field is used by source for marking priority which can guide network in case of congestion. 0 indicates high priority and cell should not be discarded unless no alternative is present

HEC (Header Error Control): This eight bit field act as header check sum for error correction.

Working

An ATM connection starts with sending a message to set up a connection and once the connection path is established, all cell will follow the same path. But unlike circuit switching it will be a virtual circuit instead of physical circuit. In ATM connection line is formed by virtual channels and virtual paths which are dealt by VCI and VPI fields respectively in the header of ATM cell.

http://i.technet.microsoft.com/dynimg/IC196038.gif

The figure [3] above shows the sequence in which the two end nodes connect and pass information through ATM network. Here the two nodes in two different ATM network are trying to transmit cells. ATM network adapter and ATM switch will get connected through user-network interface whereas two ATM switches through network-network interface.

Firstly, the virtual channel is established by sending message to the node at another end through series of ATM adapters and ATM switches. Once the end node accepts the message for virtual channel actual transmission begins. Thereafter, all the cells are transmitter through the same virtual channel. This whole process sequence is shown in the diagram above from number 1 through 4.

The information over the physical medium will travel through virtual channels. As more than one virtual channel shares same physical line, they are grouped together as virtual paths. Virtual circuit is used in connecting user-network devices and network-network devices. These channels are mainly responsible for node to node transmission; cells travelling on one virtual channel may travel through another virtual channel after switching. Due to this the values of VCI and VPI change after passing through network devices. In short, the physical path between two nodes in an ATM network is divided into logical path, identified as virtual channel and virtual path. Also, the cells arrive at each node in a sequence.

The figure below shows the diagrammatic representation of VC and VP [2].

https://encrypted-tbn2.gstatic.com/images?q=tbn:ANd9GcRDGq3qBOTpTANxNer12W7jRqDhWoVbX8zez67WoAPXu5VT8MfrQg

Advantages

ATM is independent of any transmission medium.

It can transmit different types of data simultaneously, as it utilizes total bandwidth available in the channel.

Fixed sized cells decrease the overheads in terms of encoding and decoding and also increase in transmission rate.

Disadvantages

ATM NIC and Switches were comparatively costly.

Majority of the network software were designed keeping IP as center point, so it was so difficult to move to ATM network.

Abstract

Fiber Channel is one of the most convenient physical layer standards which provide very high data transmission speed up to 1 Gbps and are able to deal with high quality storage.

Introduction

Since the boundaries of computers and telecommunications have become pale, more and more applications containing large amount of multimedia data including graphics, videos, sound were developed. Speed, reliability and quality have become paramount in the transmission world. Due to such demands a high speed data communication channel is required and it can be fulfilled by fiber channel technology.

Fiber channel not only provide high speed data transmission but is also flexible with large range of network and end devices and also supports almost all protocols.

Features

Fiber channel supports full duplex transmission with two fibers over a single line.

With fiber channels, devices can be placed relatively apart up to 10 kms.

Though it uses small connectors, they provide very high channel utilization.

Support every level of networking devices and flexible with the changes in network architecture.

Architecture

The diagram below shows the architecture of fiber channel network [2]. The fiber network is consists of end systems and switching elements. These switching elements are commonly identified as fabric. All these nodes and fabrics are usually established point to point connection between them.

http://www.cisco.com/en/US/prod/collateral/ps6418/ps6423/ps6429/images/0900aecd805cd8f6_null_null_null_02_07_07-1.jpg

Nodes use N_ports for interconnections while F_ports when connecting with fabrics. As it supports full duplex mode, two way data transmission is possible and all the frames between N_ports are routed through fabric.

Fiber channel works on circuit and packet switching and can easily accommodate new transmission medium and new fabric connections.

Protocols of Fiber Channel

Fiber channel supports five levels of protocol standards which can easily be mapped into four lower layers of OSI as shown in the diagram below [5].

http://techpubs.sgi.com/library/dynaweb_docs/0620/SGI_EndUser/books/FC_XIO_PCI_BD_OG/sgi_html/figures/fcbd2-4.gif

FC-0: This standard describes about the physical transmission media. It supports various transmission cables for different purposes. Typically, for long distance optical fiber is used whereas for short distances coaxial cable (high speed) and twisted pair ( low speed).

FC-1: This standard is concerned with signaling and encoding schemes.

FC-2: This protocol is related with framing and topologies, such as type of topology, format of frames and also flow and error control mechanisms.

FC-3: It talks about the common services such as multicasting.

FC-4: This standard specifies how different kinds of network protocols and channels can be mapped with the standards of fiber channel.

Topologies in Fiber Channel

Arbitrated Loop Topology

This topology supports maximum of 127 devices in a network. We can add or remove and device up to this number, but whenever any device is introduced or removed from the network this topology undergoes LIP (Loop Initialization Process). This process can interrupt any I/O operations taking place.

This figure shows two host systems and two storage units that are connected to each other in a ring configuration.

The figure above represents arbitrated loop topology [4], wherein devices numbered 1 are host systems and numbered 2 are storage units.

Point-to-Point Topology

In this topology ports are connected directly as shown in the figure below [4].

Switched Fabric TopologyThis figure shows a host system that is directly connected to only one storage unit.

This figure shows two host systems and two storage units that are connected to each other using three switches.

The above picture represents switch fabric topology [4]. This topology can be seen as hybrid topology as it supports both arbitrated and point to point topologies as well. It can be extended over long distance connecting different kinds of networking devices including routers.