Strictly speaking, the physical transport of datagrams, be they TCP/IP or whatever protocol being used, was not within the domain of the work carried out by the author at D2D. Nonetheless, it was important to determine what kind of network transport system was in place and whether or not it might support an Intranet and the consequent increases in bandwidth (traffic volume) demands [48]. This was an important consideration since Intranets tend to be full of images and other similar large files which swell the size of traffic on the local network.
The following information was distilled from a meeting on Friday 18th August with Fran McHugh, the IT manager at D2D and the person ultimately in charge of the local network. As well as the interview, a guided tour of the local network was given and Fran's knowledge and experience in dealing with LAN / WAN's was invaluable to hear.

It became apparent that although the current network was capable of sustaining Intranet traffic, that a considerable re-vamp of the network was in the process of being implemented, this would considerably enhance the capabilities of the local D2D infrastructure. The following paragraphs highlight the differences between the current and future networks.
The engine of the D2D network is the router, this is a 3-COM Netbuilder 8 port device running proprietary software. The router is the intelligent part of the system determining the movement of datagrams, be they TCP/IP, Novell IPX or OSLAN in identity to their destinations.
The router was not a particularly physically large device, much to the surprise of the author ca 40 cm in width and about 15 cm deep. A crucial aspect of a router is not its physical size but whether its software is robust enough to handle the millions of data packets which pass through it every week.
Each of the 8 physical ports at the back of the router support a connection to one of the variety of cables bringing incoming traffic to the router. There are two connections which go out from D2D, these are from port 1 to the parent D2D site at Kidsgrove in Staffordshire whilst port 3 connects with a CISCO router and goes to the outside world via a link through the Salford WAN maintained by CFM. This is an important connection as the Salford WAN is connected to the Internet via a connection to PIPEX, one of the ISP's mentioned in Chapter 1.
Ports 2 and 5 lead out to legacy LAN's which utilise the old ICL OSLAN protocol, OSLAN is non-routable and requires special treatment at the router, OSLAN is basically obsolete.
Port 6 handles Novell IPX traffic whilst ports 7 & 8 are identically split networks built on fibre optic cable, (to which the term FDDI may apparently be applied). These cables run from the router and are boosted via a repeater to a hub after which point, the cable switches to unshielded twisted pair (utp) to the office desktop machines. It should be noted that the although the hub used here serves merely as a junction box and is not intelligent, it is possible to install management software applications on these devices.
The greater part of the office network was run through port 4, after being boosted by a repeater, the network splits off into the section carrying the UNIX mail server and the unused Phase II section. It should be note that the mail server was capable of distinguishing between OSLAN and TCP/IP based mail, thus enabling Internet mail to be used at the Ashton site.
The remaining traffic comes from the first floor office hub, this being the location of the Intranet server. This network uses 10-base-2 thin wire Ethernet, though the connection from the repeater to the offices hub is made using 10-base-5 due to the relatively large distances involved.
It is worth mentioning the different types of transport media briefly here, a great deal of the choice of cabling is influenced by the environment in which the cabling is to be based and the distances which it has to run. Such knowledge comes through intuition and experience and it was interesting to hear Fran's opinions on the matter. The attributes of the different media discussed are summarised in the following table.
Physical transport media and their distance capabilities:
|
Media |
Name |
Max distance over which cable can be used without amplification via booster |
|
Thin wire Ethernet |
10-base-2 |
175M |
|
Thick wire Ethernet |
10-base-5 |
500M |
|
UTP/STP |
10-base-T |
100M |
|
Fibre optic cable |
FDDI |
Large ! |
Needless to say, the cost of the media varies according to its usefulness. Fibre is an excellent media but whilst the cabling itself is not too expensive, the transceivers necessary for the termination of connections are. Shielded twisted pair cable (STP) is necessary when there is a risk of interference of the traffic on the cable and the environment in which the cable is based.
Whilst the network runs quite reliably on a day to day basis, the majority of the desktop PC's are fairly old and are not capable of much more than running older applications such as Office Power.
Within the last couple of years it has become apparent that changes were necessary and so a committee, the Desktop Foundation (DF) was appointed to oversee the overhaul of the current system. After a large study completed in June of this year, a far reaching program of upgrade has been decided upon encompassing overhaul, not only of the desktop PC system as was the initial target, but of the network infrastructure as well.
The DF identified three types of user for the future system:
Browser: For example the shop floor workers in the manufacturing division.
Desktop: Office workers using Microsoft Office, Microsoft Access etc.
Developer: Those people involved in Web development, programming etc.
Running the TCP/IP capable Lotus Notes on an NT operating system, a number of new Pentium 166 powered machines were to be installed, these would be multi-media enabled, being fitted with 32MB of RAM, 1-2 GB hard disks, 12x CD players, sound cards and 15" monitors.
Whilst the current infrastructure has the capacity to handle the greater bandwidth demanded by Intranet traffic, a fibre backboned system as currently deployed at the Kidsgrove site is to be implemented within the next couple of years.
The absolute design of this system has not yet been finalised, it is however possible to say that it will use the FDDI backbone as previously stated and a series of 3-COM Ethernet switches leading off to 'branches' made up of 100 mb Ethernet.
This network will deal solely with the TCP/IP protocol, whilst that might initially seem like a disadvantage, it means that the router will not have to go through the process of evaluating which protocol to use when dealing with data packets, thus saving on processing time and speeding up the efficiency of its operation.
In conclusion therefore, it would appear that the Ashton-Under-Lyne site is to benefit from a large upgrade of its desktop and network capability, this will prove timely indeed in view of the proposed large scale development and use of a company Intranet.
| INTRODUCTION | CHAPTER 1 | CHAPTER 2 |
| CHAPTER 3 | CHAPTER 4 | CHAPTER 5 | CHAPTER 6 |
| 48 | V. McCarthy., Datamation., 1996 (December), 88-91. |