Networking can be described as the sharing of information between two or more users or devices by some method understandable to all parties involved. looking back into history we can trace the development of modern networking all the way back into the early 1800s. At this time various electric telegraph systems were developed which led to the general acceptance of the Morse system as standard and enabling widespread transferral of data. Around the 1870s harmonic telegraphy was invemted resulting in the first telephone systems. In the early 1900s the first radio signals were sent across the Atlantic Oceans and undersea telegraph cables were laid connecting the continents.
The years around the Second World War led to many advances in technology. The invention of the transistor paved the way for the digital revolution that started in the sixties. IBM introduced a ‘compatible time sharing system’ in the early part of this decade which enabled separate terminals in different offices to access the same hardware. The concept of remote client access to a host computer or server was now realised and the concept of a local area network born.
The first wide area network, created in the mid-sixties, connected computers via a low speed telephone line and exposed the limitations of the public switched telephone network to system developers.
In this decade packet switching was developed and a packet switched network, ARPANET proposed. ARPANET was a network designed to interconnect military installations. 1969 saw the first RFC, RFC001, describe a computer 'handshake'. The first ARPANET network was implemented in 1970 at UCLA after a protocol was designed to allow computers to send and receive messages and data. This protocol came to be known as an interface message processor. The seventies saw rapid development of network technologies. The early part of the decade brought the development of electronic mail messages, gateway technology, Ethernet, host-to-host protocols and TCP/IP. TCP/IP enabled separate networks to communicate. In the early eighties TCP/IP was integrated into Unix and ARPANET, the SMTP protocol and DNS established. DNS enabled the establishment of a true Internet and later years saw the development of the OSI Protocol standards.
As can be seen this development path led to the interconnection or networking of computers over a wide area as well as to the development of 'networkable' devices, such as network printers. These devices could be connected to a network for the benefit of all users having access to the device.
Unfortunately industry was developing interconnection methods completely incompatible with the TCP/IP protocol suite.
Industrial connectivity
As the Internet, TCP/IP Networks and e-mail is the norm to current PC users, the electronic and electrical industry developed their own standards for connecting devices. Communications between electronic devices usually takes place using one of the following methods, RS232 or differential.
RS232 is an electronic interface developed in the early sixties and has remained in wide use throughout the electronic and electrical industries. Here, two discrete voltage levels relative to a common ground represent binary levels. Because RS232 is based on a potential difference, the distance signals that can be transmitted is limited. Modern computers also tend to simplify the interface by accepting 0 V as logic zero and 5 V as logic one. This low voltage level further limits the range of a RS232 interface.
There are also various handshaking lines that may be used to control devices using this interface. There are two types of RS232 devices, DTE (data terminal equipment) and DCE (data communications equipment). RS232 defines a that a common ground should exist between a DTE and DCE device. Many RS232 interfacing problems occur when a potential difference exists between these devices.
For many years the industry standard for connection to and configuration of devices, was asynchronous serial communication. A basic explanation of this concept would be that when data is transmitted at irregular intervals, it is known as asynchronous transmission. As an example, a keyboard does not send characters at regular intervals, but the bits within each character must be sent at regularly timed intervals.
The asynchronous transmission line has two electrical states, on and off, which is represented by 1 and 0. A start bit indicates the start of a character whilst a stop bit marks the end. When the receiving device detects a start bit, it counts off the regularly timed bits that form the character and returns to its passive state once the stop bit is detected. The presence of these start and stop bits allows the time intervals between characters to be irregular, or asynchronous. On the other hand, synchronous communication relies on a timing scheme coordinated between two devices to separate groups of data and then to transmit them in blocks known as frames. Synchronisation between the devices is checked periodically.
Asynchronous serial devices are of the most prolific of all computer peripherals currently in use by the information technology community. Many examples of such devices confront us daily: magnetic card, smartcard and bar code readers, automatic teller machines (ATMs), point of sale (POS) and electronic funds transfer (EFT) terminals, industrial controllers, serial printers and legacy devices such as asynchronous terminals. Some serial devices require a dedicated PC to not only present the management and configuration information, but also to process such information before the user views it.
As technology progressed, local area (LAN) and wide area (WAN) networks expanded, becoming the accepted norm for transferring data between devices and peripherals. Of these LAN technologies, Ethernet is certainly the most popular. Unfortunately the packet-based transfer protocols, as used by Ethernet and Token Ring technologies, excluded vast numbers of existing asynchronous serial devices from direct network connectivity. The world is rapidly moving towards total connectivity based on LAN/WAN systems. Connecting asynchronous serial devices to Ethernet LANs currently require the use of costly and inefficient methods. Typically, serial devices have to be front-ended by a PC whose sole purpose is the conversion between serial and Ethernet communications. It was to address this problem that the EtherPAD was developed. In short, it provides an asynchronous serial socket for a serial device on an Ethernet LAN.
Benefits of networking serial devices include:
* Remote device access and management: the fact that serial devices can now be accessed via an IP address enables remote management using tools such as SNMP. SNMP is a defined network management protocol providing continuous management information derived from continuous device monitoring. There are many third party vendors, such as HP, that have developed application software packages for SNMP network management. The Internet also allows a user to roam the world at will, still having access to a serial device connected to a network.
* Flexible device management: most modern buildings are wired for easy network access from virtually any location in that building. Serial devices can thus be conveniently placed anywhere on a network and accessed from any point on a LAN/WAN. The physical length of the cable connection between the serial device and its management console is thus not constrained by the limits inherent in the RS232 interface.
* Lower costs: device management can be automated and centralised with resulting savings in time and human resources. Hardware and maintenance costs are also reduced in that less equipment is required to monitor individual serial devices.
Basic principles of serial connectivity using EtherPAD technology
At its most basic level, an EtherPAD is a protocol enhanced I/O controller. The serial port collects asynchronous data input for the building of packets of information. These packets are then encapsulated in Ethernet frames and transmitted on the connected Ethernet LAN. In the opposite direction, the contents of Ethernet frames intended for the serial port are unpacked, or disassembled, for output to the serial port. EtherPAD thus functions as a packet, or datagram, assembler/dissembler (PAD) between an asynchronous serial device and an Ethernet LAN. The correct protocol choice for the format of the packets contained within the Ethernet frames was important. A proprietary protocol could have been created, but, for practical reasons, a subset of the TCP/IP set of protocols was selected for implementation. The benefits of using TCP/IP as standard are numerous and include:
* A ubiquitous standard: all major operating systems support TCP/IP. Even DOS has third party TCP/IP support. TCP/IP is the internetworking standard used by the Internet. This means that datagrams containing data collected from or intended for the serial port can be routed to any interconnected network even beyond the connected Ethernet LAN. Effectively the decision to adopt the TCP/IP stack as transmission protocol allows a PC-based application to control and monitor a serial device located almost anywhere in the world. It is a well-established standard that has showed exponential growth during the last decade.
* TCP/IP is a multiple transmission media protocol which can deliver datagrams to hosts on other LAN technologies such as ArcNet and Token Ring.
* The TCP/IP stack includes protocols such as BootP, which may be used by a device to learn its IP address and about its environment. TFTP may be used by a device to obtain configuration data.
* TCP/IP presents no problem co-existing on the same LAN with other protocols such as IPX, NetBEUI and AppleTalk.
EtherPAD, in addition to serving as a means to connecting a serial device to an Ethernet LAN, assigns an Internet, or IP, address to a connected serial device, thereby making it accessible via TCP/IP networks such as the Internet.
Two main methods of application
Firstly, EtherPAD may be used to allow a PC-based application access to over a thousand serial devices connected to a single Ethernet LAN. This may distribute the serial devices over an area equal to that covered. The maximum throughput of such a system will be a function of the throughputs offered by the Ethernet card, the TCP/IP stack on the PC, the hard disk and the CPU. This makes EtherPAD good for applications involving many serial devices, such as access control systems, networks of asynchronous terminals, etc which are connected to a central controller or server.
EtherPAD may be used to geographically separate a serial device and a controlling application thus effectively extending the serial connection beyond its normal distance limitations. The device therefore allows control and/or monitoring of a large number of serial devices located over an area as small as that covered by a single Ethernet LAN or as large as that covered by a WAN or the Internet.
A variation on this method uses virtual port redirector software. Here, a PC running redirector software creates a 'virtual' serial port and redirects any data to – and from the application to the EtherPAD's IP address. The PC COM port can thus be extended across a LAN or WAN to control a remote serial device. A typical application for this software would be a dated application unable to access TCP/IP or Winsock. The life of any application can thus be extended far beyond the limitations of serial ports and DOS. Costs can be reduced as applications do not have to be rewritten and recompiled.
© Technews Publishing (Pty) Ltd | All Rights Reserved
printer friendly version