22 Electronic Mail, Znternet and Electronic Message Services The ability to connect two computers together and, with relative ease, to send information from one to the other, is bringing a revolution in the way in which business and life as a whole is conducted. Today it is possible to run your bank account from home, book your holiday, send electronic messages to your work colleagues or friends, and look up to see what is on at the theatre in London or New York City. Alternatively companies may make their orders to and pay their bills from their suppliers by computer program and ‘electronic data interchange’. A number of technologies have enabled this revolution: videotext, electronic mail, electronic data interchange (EDI) and Internet. We review the telecommunication aspects of these technologies in this chapter. 22.1 VIDEOTEXT Videotext was the first type of device specially designed to allow telephone network customers to use a cheap device to access information from a large public database. The original technology and standards, including special terminals and modem techniques allowing asymmetric transmission (a high bitrate channel for information download to the customer with a low speed control channel for his ordering of different pages of information) have now been overtaken by modern personal computer based technology, but the appearance of the worldwide Internet has stimulated recent rapid growth for the long established videotext service providers (British Telecom’s Prestel, Deutsche Telekom’s Bildschirmtext, BtX or Datex-J, and France Telecom’s Minitel). The idea of videotext is that, using a low cost terminal in the form of a small tele- vision, a customer can make a phonecall to a public central database, where he could access all sorts of pages of information which he could then have displayed on his screen. Thus, for example, he might access tomorrow’s weather forecast, current flight arrival information, information about financial markets, about holiday offers, about dating services or, via an extended connection to the railway company or his bank, might even order a train ticket or pay his bills. 399 Networks and Telecommunications: Design and Operation, Second Edition. Martin P. Clark Copyright © 1991, 1997 John Wiley & Sons Ltd ISBNs: 0-471-97346-7 (Hardback); 0-470-84158-3 (Electronic) 400 ELECTRONIC MAIL, INTERNET AND ELECTRONIC MESSAGE SERVICES Videotext in its original form in the United Kingdom (as Prestel) was quickly accepted by the travel industry, which used it widely as a means for holiday companies to advise travel agents of available itineraries and to book tickets on behalf of clients. It came at a time of revolutionary computerization of the airline industry in particular, when simultaneously the airlines were cooperating to establish and operate common and inter-linked computer ticket and booking systems. In France, videotext became the standard method for telephone directory enquiries. The Minitel terminals were given away free by France Telecom. They justified doing this initially by the cost savings brought about by fewer telephone enquiries to human operators, but nowadays they also make money from the telephone calls which cus- tomers are making to the other information services which have subsequently sprung up. Some services, including the pink pages of dating services, are world renowned. In Germany, the Deutsche Telekom’s (formerly Deutsche Bundespost’s) T-Online service (the new name for the modern version of Bildschirmtext, Btx or Datex-J) had been a commercial non-runner for years. Suddenly, however, during 1995 demand for connections started to increase rapidly. The reason was the cheap availability of access to the Internet using standard personal computers. It is the Internet which has become the world’s electronic noticeboard. We discuss it later in this chapter. 22.2 ELECTRONIC MAIL (E-MAIL) Electronic mail (e-mail) is a reliable means of message communication between human users equipped with computer terminals or personal computers. Large tracts of text, and diagrams too, can be quickly delivered across great geographical distances, and if necessary printed to a high quality paper format using local computer printing resources. To send an e-mail, a user simply calls up the electronic mail software on his terminal. This prompts him for the name of the user that he wishes to send his message to, and for the names of any individuals to whom the message is to be copied. Having filled in this information, it prompts him to type the main text of the message. If he wishes, the sender can add further electronic attachments (say a document previously created using his word processor, his spreadsheet program or presentation software). Once the e-mail is ready for sending, the sender simply clicks his mouse on the ‘send’ button. At this point, he may be asked to set a priority rating for the message, to decide whether confirmation of receipt is required, and (optionally) to set an exact time and date for delivery (if not to be immediate; this feature can be used as a reminder function, for example by sending time-delayed messages to himself). The delivery into the electronic mailboxes of the intended recipients is almost immediate. Should any of the recipients be at their computer terminals at the time of the message receipt in their mailbox, then they may be advised of its arrival (say by a beeping noise and a short message or icon displayed at the bottom of their computer screen). They have the option to read the message immediately or later, depending on how they regard its priority. Those recipients who are not concurrently using their terminals will be advised of the message next time they log on. As messages are read, confirmation is returned to the sender (if required). Each recipient has the choice to reply to the message, forward it, file it, print it out, amend it or delete it. ELECTRONIC MAIL (E-MAIL) 401 Electronic mail systems are usually constructed in a clientlserver configuration. The client is the human user and his personal computer, workstation or other computer terminal. The client prepares and reads electronic messages. The client software prepares the e-mail for submission to the server or post ojice. The post ojice has a distribution function for electronic mail similar to the function of the postal service for letter mail (Figure 22.1). The client software for electronic mail usually runs on a personal computer, and normally has software interfaces to other PC software such as word processing software, spreadsheet software and presentation software, so enabling easy attachment of such software to e-mail messages. Thus, for example the Microsoft Office package of software includes Microsoft Word, Excel, Powerpoint and Microsoft Mail software. Together these allow a personal computer user to prepare various types of different text, spreadsheet and presentation documents and submit them to an electronic mail post office as e-mail messages. The post office function which forms the core of electronic mail usually resides on a server, often nowadays connected to the individual PCs by means of a LAN (local area network), though other types of X.25 packet network or dial-up connection are also possible. The post office is a combination of software and hardware, capable of storing e-mail messages, sorting them and transmitting them. The post office function works both on a store-and-forward and store-and-retrieve fashion. At the sending end, the user submits his message to the post office (post office A of Figure 22.1). This post office temporarily stores the entire message while it examines the address of the intended destination and establishes a communication means to the post office (post office B) serving the destination user. The communication means may take any number of different forms (X.25, telephone connection, internet connection etc.). Once the lower layer connection is established between origin and destination post office, or between origin post office and intermediate (transit) post office, the higher layer protocol (i.e. OS1 layer 7 protocol for electronic mail (e.g. ITU-T X.400, SMTP or equivalent)) takes over, to ensure the appropriate relay of the entire message to the next post office. In each post office along the way, the message is stored and forwarded in this way. 402 ELECTRONIC MAIL, INTERNET AND ELECTRONIC MESSAGE SERVICES Finally, at the destination post office, the message is stored in the electronic mailbox of the destination user until that user retrieves it. This is done when he logs into his own electronic mailbox account to read or send messages. ITU-T, in its X.400-series of recommendations, specifies a message handling system (MHS) for use as an OS1 layer 7 protocol in support of electronic mail. This is described in detail in Chapter 23. MHS defines P1 and P3-interfaces for relay and submission/ delivery respectively. In reality, however, the P1-interface has proved to be the more important of these two, since this is the interface which allows electronic mail servers (post offices) to inter-communicate with one another. Thus the electronic mail systems of different companies, and maybe supplied by different software manufacturers are able to transfer e-mails between themselves. This is the X.400-interface talked of by electronic mail software suppliers. The P3-interface is only needed when a remote electronic mail user (client) wishes to submit or receive electronic mail messages directly from a post office operated by a public X.400-based electronic mail service provider. As, however, most electronic mail systems and post offices are based on corporately operated sytems, it is usual for client and post o@ce (server) software to be provided by the same supplier (e.g. Microsoft mail or Ccmail). In this case, it is not necessary for the software manufacturer to use the P3-interface, and instead a proprietary interface is employed. An alternative to the use of the X.400 interfaces for inter-post-office interfaces is the use of the Internet. The Internet interface looks likely to become the predominant interface. E- mail based on the Internet is already firmly established in North America, and is fast supplanting X.400 where this already exists (mainly in Europe). In addition to the X.400 and Internet-based electronic mail systems, a number of other proprietary systems have emerged. One of the first was introduced by Compu- serve, one of the world market leaders in on-line services. Compuserve created the capability for companies and private individuals, from their personal computers, to access database information and send messages between one another. Its success has been largely based on being the first company to establish a worldwide user group for electronic communication. Currently it appears to be re-positioning to encompass Internet technology. Microsoft, on the other hand, has chosen to build a Microsoft mail capability into its Windows95 operating system software for personal computers. This client capability can be used either in conjunction with a corporate post ofice, or alternatively, may use the Microsoft Network (MSN), a public post o@ce network operated worldwide by Microsoft Corporation. The initiative lays down a tough challenge for other X.400- and Internet-based electronic mail service providers. 22.3 ADDRESSING SCHEMES FOR ELECTRONIC MAIL Three main types of addressing scheme are used nowadays for electronic mail. These are e proprietary addressing schemes e Internet-based electronic mail addresses e X.500 addresses (for the X.400 message handling system) THE ADVANTAGES AND DISADVANTAGES OF E-MAIL 403 Proprietary addressing schemes are normally employed within a corporation’s electronic mail environment. The form of these addresses is often determined by the corporation itself, within the bounds set by the software product on which the system is based. Internet-based electronic mail addresses (correctly SMTP addresses) usually take a numerical form, punctuated with dashes (-), dots (.) and @ signs (e.g. 12345.67890 @compuserve.com). More commonly nowadays, an alias address appears on business- men’s visiting cards, e.g. Martin.Clark@serviceprovider.corporation. This form allows for easy recognition and memory of the address. X.500 addresses are less common, but form the basis of the ITU and IS0 directory service and are used in X.400 electronic mail networks. We describe the X.500 address- ing scheme in Chapter 28. 22.4 THE ADVANTAGES AND DISADVANTAGES OF E-MAIL E-mail messages are delivered quickly. Broadcasting of messages is also quickly and easily achieved. Editing of text and returning or forwarding the amended version can be achieved with minimal re-typing. Messages can be filed and quickly retrieved later. Messages can be posted for exactly timed delivery, and can be prioritized according to the urgency with which they need to be dealt with. Furthermore, users are able to check their electronic mailboxes even when they are away from their normal offices, either by using somebody else’s terminal or perhaps by dialling in to the post office using a portable laptop computer from a hotel room. The Internet, in particular, has done a lot to release the full power of electronic mail, by enabling a large worldwide community of computer users to inter-communicate elec- tronically, thus destroying the previous communication boundaries between companies and overcoming all the barriers of geography and time zones. Companies who have successfully introduced electronic mail have observed a beneficial change in the whole culture of how they do business. Questions and responses have been much quicker and more direct, messages have been much shorter and less formal; they have been typed by the managers rather than by their secretaries. Workgroups composed of members in widespread locations have evolved, and it is possible to draw together new teams for previously impossible tasks. 22.5 EDI: CORPORATE COMMUNICATION WITH CUSTOMERS AND SUPPLIERS VIA E-MAIL Where electronic mail is used as the basis of formal communications between companies for orders, payments and confirmations, it is termed instead electronic data interchange (EDZ). EDI between companies has become an important way of doing business. Many manufacturers and retailers demand that their suppliers accept orders for goods electronically. This gives the scope for more frequent ordering (in the case of supermarkets, for example, daily based on yesterday’s sales) and so maintain their shelf stocks (and value of stocks) at a minimum level. This is just-in-time (JZT) provision. 404 ELECTRONIC MAIL, INTERNET AND ELECTRONIC MESSAGE SERVICES Not only the buyer benefits, but also the supplier can ensure that products on the shelves with his name on are fresher and closer to their best. Even more important for the supplier, EDI allows him to bill for the goods more quickly. The most important of standards defining the format and content of messages for elec- tronic data interchange (EDI) are those published by the United Nations as EDIFACT (electronic data interchange for administration, commerce and transport), but there also are a number of regional, national and industry-specific standards (e.g. Odette used by the automobile industry). EDI products are available in ready-packaged software solutions for integration into a company’s existing computer and electronic mail set-up. Specialized service providers have also appeared in several countries concentrating on the communication needs of companies within a specific industry segment. In this way, a community of interest may be established for communication between customers and suppliers (for example, within the retailing industry or the car manufacturing industry). Setting up a network and software for EDI is actually quite straightforward. What is much harder, and what the company therefore needs to be prepared for, is the adjust- ment of its culture and practices to an entirely new way of business, one requiring much quicker reaction to customer demands. 22.6 INTERNET The Internet emerged from a United States government and military initiative to enable the interconnection of different, mainly UNIX-based computer systems for intercom- munication. As UNIX was proclaimed to be the first ‘portable’ operating system for computers, enabling software developed on a particular manufacturer’s computer hardware to be easily ported (i.e. transferred for operation) to another manufacturer’s hardware, it was natural also to develop means for easy transfer of data between systems. This led to TCPjlP (transport control protocollinternet protocol). TCP/IP was quickly adopted by the academic community in the United States, and soon afterwards by academics worldwide, because it allowed for rapid sharing of scientific information and the electronic mail communication necessary for its rapid discussion and analysis, both within and between university campuses. A worldwide community of inter-linked computers rapidly emerged. Finally, as businesses recognized the potential of the Internet as a large interconnected group of electronic mail users and noticeboard readers, they began to exploit it for communication and marketing purposes. Though the Internet does not have the rigid structure, network management and security controls of other public telecommunication network services, it has a very persuasive appeal; there are already plenty of people to communicate with. This has driven the explosive growth in numbers of registered Internet users and pages of information. In its original form, the Internet and the internet protocol (IP) provided a means for interconnecting computer servers (typically UNIX computers) together. The internet addressing scheme allowed individual workstations, personal computers or software applications running on either the server or on any of the workstations to direct and send information to other applications on other servers or distant LANs. Being a unique TCP/IP PROTOCOL STACK 405 address, the Internet address allowed an end user to be identified, no matter how many transit servers, routers or networks would have to be traversed along the way (Figure 22.2). The single network address of the destination port in the destination network does not suffice, because the intermediate networks are unable to recognize this address. A suite of new protocols arrived with Internet. Among others, these included SMTP (simple mail transport protocol), TFTP (trivial $le transfer protocol) and FTP ($le transfer protocol), which together form the basis of the Internet electronic mail service. As the popularity of Internet has grown, so have the number of servers and routers making up the network. New Internet service providers have provided for dial-up access to the servers from private individuals using their PCs at home, and new information providers have provided more Internet pages of information. The Netscape browser, a computer software allowing users to ‘surf the Internet and World Wide Web (WWW), by seeking information from any of the connected servers by means of a menu-driven screen software which drives a hypertext search and browse capability. However, the reason for the rapid growth of the Internet also provides a major challenge for the next stage of its development. The factor enabling rapid growth was the possibility to add further servers and routers to the network at almost any point in the network without considering a master plan. The number of Connected devices could therefore rapidly increase, as the routers (Chapter 19) are always able to find the desired destination somehow. The problem is the lack of control (and even lack of knowledge) of the path taken. The traffic flo& in the network as a whole are therefore largely unmanaged and unmanageable. The only solution to slow response or congestion can be to add more capacity. Whether this capacity is added at the most appropriate point is a matter of chance. Messages within the Internet may go undelivered without trace for any number of reasons, and there is no record of whether electronic copies of information have been retained by any of the intermediate parties. Much attention is currently being focussed by the computer industry on solving the security and network management difficulties of the Internet, to stimulate a further surge in demand. 22.7 TCP/IP PROTOCOL STACK Figure 22.3 illustrates the various protocols going to make up the TCPIIP stack. At the heart are the Internet protocol (IP), which is equivalent to an OS1 layer 3 network 406 ELECTRONIC MAIL, INTERNET AND ELECTRONIC MESSAGE SERVICES OS1 layer 7 (Network (Remote (Trivial file (Simple (Simple (File windows NFS RPC TFTP SNMP SMTP FTP TELNET x- 6 transfer mail UDP TCP 4 ent protocol 5 file server) procedure transfer network 3 ARP ICMP RARP Gateway protocols IP BGP EGP 2 SNAP SLIP (PPP) protocol) call) protocol) managem transfer protocol) LLC (e.g. Ethernet LAN) Serial line Frame Relay etc. 1 physical network I ARP = BGP = EGP = ICMP = IP = LLC = PPP = RARP = SLIP = TCP = UDP = Address resolution protocol Border gateway protocol Edge gateway protocol Internet control message protocol lnternet protocol Logical link control (for LANs) Point-to-point protocol Reverse address resolution protocol Serial line internet protocol Transmission control protocol User datagram protocol Figure 22.3 TCP/IP protocol stack and associated applications protocol, and the transport control protocol (TCP) an approximate equivalent of OS1 layer 4. Also very important for some common multivendor computer-networked applications is the UDP (user datagram protocol). The Internet protocol is a network layer-like protocol, typically running across an ethernet LAN or between LANs via a router network (say comprising inter-router con- nections running on frame relay). The benefit, however, of IP over OS1 layer 3 protocols is that Internet addresses are widely used in, and are unique across, all computer networks worldwide. This gives the potential for a software application or a user connected anywhere in the world to the Internet to access any other computer or software application. The transport control protocol (TCP) is usually used in conjunction with IP and ICMP (internet control message protocol) to guarantee reliable transmission. On an end-to-end basis, TCP ensures correct sequencing of arriving frames of data, and requests retransmissions when necessary. ICMP is an addition to the basic IP allowing network problems to be reported back to a communicating device. Thus ICMP is able to report the inability to deliver a message and the cause, or the need for fragmentation. An alternative to TCP is UDP (user datagram protocol), a simpler protocol which does not perform retransmissions. Instead this job is left to the application if necessary. The address resolution protocol (ARP) is used in association with the IP to translate IP addresses into physical hardware addresses. Thus, for example, ARP is capable of determining the appropriate ethernet LAN address (LLC). This address corresponds to COMMON APPLICATIONS USING TCP/IP 407 a given IP address. RARP performs the reverse function. When, for example, a diskless computer workstation boots, it obtains its hardware address from the network interface card to which it is attached. It does not, however, know its IP address, this is resolved by RARP. The gateway protocols (BGP, border gateway protocol and EGP, exterior gateway protocol) are used to connect together different sub-networks of the Internet. They provide for inter-network routing information and communication exchange. SNAP (sub network access protocol) is an extended version of the LAN logical link control (Chapter 19) which enables the internet protocol (IP) to be carried over LANs. SLIP (serial line internet protocol) or its successor, PPP (point-to-point protocol) are equivalent to OS1 layer 2 protocols. They enable transport of IP packets across a simple serial communications line (such as a telephone connection or leaseline). The protocol does not include an address field. The main function is simply to ensure the delineation of packets. Other well known routing protocols of the TCP/IP suite (e.g. RIP, OSPF) are discussed in Chapter 28. 22.8 COMMON APPLICATIONS USING TCP/IP A number of the most common applications used in association with TCP/IP are also shown in Figure 22.3. These are X-windows TELNET FTP (file transfer protocol) SMTP (simple mail transfer protocol) SNMP (simple network management protocol) TFTP (trivial file transfer protocol) RPC (remote procedure call) NFS (network file server) We discuss each of these briefly in turn. X- windows This is a windows-based computer operating system intended to allow computer workstation users anywhere within a network of computers supplied by multiple vendors to access software applications running on remote hardware platforms. TELNET This is a simple TCP/IP-related application that enables a remote log-on from one com- puter to an application running on another; if you like, a simple version of X-windows. Thus a user of of personal computer using Windows95 software has the option to 408 ELECTRONIC MAIL, INTERNET AND ELECTRONIC MESSAGE SERVICES commence a TELNET log-on, enabling him to conduct a session on a remote UNIX computer server. It is the simplest form of TCP/IP connection. As such, it is worth mentioning at this point the widely used PING (Packet Internet Groper) technique for confirming the availability of a IP connection. A PING is an IP-message that contains the address of its source and the destination. On receipt by the destination terminal a confirmatory PING is returned. The procedure confirms not only that the network connection is available but also that the destination device is ‘alive’. FTP (file transfer protocol) This is a simple but efficient protocol to transfer data files from one computer to another. Although nowadays it is also possible to include data files (e.g. Microsoft word data or Microsoft Powerpoint presentation) as an attachment to an e-mail message, this can be a very slow and cumbersome means of transferring large files. Drawing files from a remote computer can be effected much more quickly using a remote log-on and then the FTP. SMTP (simple mail transfer protocol) SMTP has become the basis of the now world-renowned Internet electronic mail system (with the addresses that run president@whitehouse.gov). It provides for transfer of mail messages between mail servers running on the Internet. SNMP (simple network management protocol) SNMP has become the de facto standard method of transporting network management information around computer and router networks. We discuss it in a little more depth in Chapter 27. TFTP (trivialJile transfer protocol) This is intended to be a simple protocol, and it is particularly well suited for the downloading of software to and initiation of a remote device. RPC (remote procedure call) This provides for a software routine or other application to be called and executed on a remote computer or server. NFS (networkJile server) This application makes files on remote servers and other computers appear to reside on the local user’s workstation. It enables him to read and process the data as if it were residing in his own machine. 22.9 THE INTERNET PROTOCOL (IP) Figure 22.4 illustrates the fields comprised in an IP frame (or datagram). The version indicates the format of the IP header, specifically which version of the protocol is in use. (For example, the 1993 version was version 4). The internet header length (ZHL)