Randy Plessor

An Amateur’s Crash Course On Setting Up a Home Network

Since Microsoft’s release of Windows for Workgroups (3.11), networking has become very simple and commonplace for even home computer users. Everything one needs to setup a network at home is built right into the operating system that came installed in the machine. It just has to be turned on and setup. With Win95/98 and its ‘plug & play’, setting up a network is very simple.

A simple ‘local area network’ or LAN consists of two computers each containing a network adapter board, a network capable operating system like Win95/98 or NT and a special ‘cross-over’ cable to connect them together. With the addition of a third machine, another component called a hub must be added. A hub is a box with holes called ‘RJ-45’ jacks. They look similar to the jacks on a telephone (and wall plate where the pone line connects) except they are a little bit wider and have eight conductors in them instead of four. The wires look just like the wires you buy at the phone accessory rack in places that sell phones. Telephone cables, however, are not usable for networking for a number of reasons. First, they generally are not a high enough grade to handle data especially at 10 and 100 megabits per second (Mbps.). They are also wired differently in the RJ-45 plugs at each end.

Hubs and network adapter boards can be purchased at your local retail computer store and are quite inexpensive. They can even be found now at computer shows where you can expect to pay even less. Boards run all over the lot in price. I would suggest to the beginner to look for cards with the Win95/98 ‘Plug & Play’ logo on the box. This will greatly simplify the configuration process in a machine running Win 95/98. If you use Windows NT 4.0, although it is not a true ‘Plug & Play’ (PnP) operating system, it does have a program called NTDETECT that will identify ‘PnP’ network cards.

Installation of the card will be the most involved part of the job. If you can use a screwdriver, you can install a network interface card of "NIC’. Turn the machine off. Unplugging the machine is up to you; buy leaving it plugged in does supply a great ‘earth ground’ through the lime cord to protect the sensitive components from static charge. Take the covers off of the machine. Place a hand on the chassis to ground yourself. Check to see what type of slots your system board has available and purchase a NIC that matches that slot type. Remove the screw holding the slot’s metal plate on the back of the chassis (or break it off and scrounge around for a screw that fits if you have a ‘cheap chassis clone’). Simply insert the card into a matching slot. I suggest not the one at the extreme right or bottom (looking from the back of the machine). This will make it easier to remove the network cable from the card jack when you want to (the ‘clip’ like on a phone cable can be difficult to get to when the card is in that slot… you’ll have difficulty every time you want to disconnect it). That’s it… provided you are using a ‘Plug & Play’ card. Put the covers back on.

Now the board has to be recognized by the system. With ‘Plug & Play’, the additional hardware will be detected automatically and should configure itself as far as the ‘interrupt’ and addresses are concerned. If you do have a problem, you will need to make changes to the ‘interrupt’ setting because it may be conflicting with another device already in the machine. Doing this will require a bit more knowledge about the workings of a PC. When it is recognized, software will be loaded automatically if it can be found on your system. If the files cannot be found, you will have to tell Windows where they are or put he Win 95/98 CD into the CD-ROM drive. If you have a NIC that is newer than the operating system you are using, the specific software the NIC uses may not be available. You can use the ‘Have Disk’ option if the NIC came with a diskette supplying the necessary software.

The following pertains to Windows 95/98 machines. If you are using NT, it will be a little different, but basically the same things must be done. When the machine is finished booting up, click the START button. Click SETTINGS, CONTROL PANEL and then click the NETWORK applet icon. It will open to the ‘configuration’ panel. You should find several lines in the window. Windows 95/98 installed these when it identified the ‘new hardware’ during boot-up. You may also find listed here an item called ‘Dial-Up Adapter’ if you have Internet access through an ISP or Dial-Up service. If these lines are there and Windows re-booted by itself after it found the new hardware, you are close to complete. If nothing is listed in the window, or just a line for ‘Dial-Up Adapter’, you will have to load the software manually. This will be the case with a non-PnP NIC too. That’s OK. We’ll check to see what these things are all about anyway.

Click the ADD button on the NETWRK applet screen. Another window will open and will list ‘Client’, ‘Adapter’, ‘Protocol;’ and ‘Services’. Each will be done one at a time. First, the Adapter; double click the adapter line and after a program runs to load the NIC database, a window will open. The panel on the left lists NIC manufacturers in alphabetical order. The panel on the right will list the models of NIC’s made by the manufacturer highlighted in the left panel. Scroll down to the manufacturer of your NIC and click the name. Choose the exact description of your card. This is critical because the wrong ‘device driver’ will not turn your board on at all or it will not allow it to do all of its capabilities. Again, if it is necessary, you may need to use the ‘Have Dick’ option, which will add the addition drivers to the list for future use.

Click back to the ‘configuration’ window and click ADD again. Now double click ‘Protocol’. A protocol is like a language that ALL of the cards on the network will use to communicate. Like people, the cards can communicate only if they use the same language or ‘protocol’. For a small home network to start with, choose ‘Microsoft’ then ‘NetBEUI’. Later, you may want to choose ‘Microsoft’ TCP/IP also from this window. If there was a "Dial-Up Adapter’ line in the configuration windows chances are TCP/IP was there already.

Go back to the ‘Configuration’ window again and this time double click ‘Client’. This determines what login screen will come up when you sign on to your network. The job of the client is to determine if a command that you give (or one that is given by software) is to be executed locally or if it is for a machine on the network, and handles the re-direction. Sometimes the client is referred to as ‘the shell’, the ‘re-director’ or the ‘requester’ of the networking software. Click ‘Microsoft’ and then ‘For Microsoft Networks’. The ‘For NetWare Networks’ would be the choice if you were connecting to a NetWare server.

One more time, return to the configuration page. The last choice is ‘Services’. There are two things to set here if necessary. They are optional. In Win 95 they are separate and the choice to share files and share your printer, if you have one attached to your machine, can be set. IN Windows 98 they are combined in the first choice. Simply put, the settings here determine if your machine will also act as a ‘server’ (be able to share its resources) besides its normal role of ‘client’ (able to access the resources of other machines). That does it for the configuration part. The rest is easier.

Assigning Machine and Workgroup Names

Go back to the screen with configuration and click the center tab "Identification’. In order for a machine to be identified in a network, it has two identifying entities. The first is the ‘machine name’. Type in a name for your machine in the machine name field that others on the network will know your machine by. The second is the workgroup or domain name (with an NT server) that the machine is in. For a simple home network using Win 95/98, you would want to name a workgroup. You can leave the default ‘workgroup’ there if you wish or change it to anything you want. All of the machines should have the same workgroup name.

Finally, click the ‘Access Control’ tab. Here you will determine the type of security, real simple for Windows 95/98. You simply click the top choice, ‘Share Level’ because these operating systems do not really have any security. The type of network being setup here is what is known as a ‘peer to peer network’. This means that every machine can be used to view and access what other machines on the network ‘share’, or access what is called a ‘resource’. A machine can also share its own resources and particular arrangement has drawbacks. It’s fine for right now. That is explained further below in ‘Different Types of Login’. Close out the NETWORK applet and the machine will update the registry and will need to be re-booted (the machine should do so by itself).

The rest of the installation depends on where and how far apart the machines are physically located. Connect one end of a cable into the ‘RJ-45’ jack on the back of each machine. Connect the other end into an available jack on the hub. Make sure that you do not use a hub jack that is marked differently than the others. This is a ‘cross-over’ jack. Basically, this is the same thing as the ‘cross-over’ cable or a ‘null modem’ cable. When only networking two machines without a hub, a ‘cross-over’ cable forces the ‘transmit’ conductor on one machine to match up with the ‘receive’ conductor on the other and vice versa. This crossover is not needed when using a hub except for when you are connecting two hubs together to expand your network. In this case, one, not both hubs, has a cable plug snapped into the specially marked ‘RJ-45’ jack to provide the forced connection described above machines plugged into the other hub. Remember: only use the crossover jack on one hub. Some hubs may have a switch for this purpose. Again, only switch one. Otherwise, you will crossover twice which would be like not having crossed over at all.

If all has been done correctly to this point, when you reboot the system, it will pause with the Microsoft Networking Login screen. For a workgroup, this is primarily a local login for the purpose of identifying one of perhaps several users of that machine who have their own profiles. In the case of a domain, it is truly a security check where the user’s login ID and password are ‘authenticated’ by either a Novell NetWare server or a Windows NT server.

The access type set in the Access tab in the NETWORK applet determines how one accesses the various resources on the network. In a workgroup where the ‘Share Level’ security model type login is used, each individual resource has a password. If the person whose machine is ‘sharing’ a resource has given you the password to that resource, you can access it. The user of that machine administers each machine’s resources that are shared. When you attempt to access a resource, a login dialog box will appear at that time. If you know the password, you can gain access to the resource by simply typing it in. The password dialog box will not come up again during that session. However, you will be presented with a password dialog box for each resource when you first attempt to access it.

For a domain, which is referred to as a ‘Client/Server’ type network, a single login dialog box is presented when one first logs on to the network. This would occur if the ‘User Level’ access type were selected in the NETWORK applet ‘Access’ tab. At this time, the Primary Domain Controller or PDC will check the ID and password combination and verify that the combination is legitimate. If so, the user is ‘authenticated’. The PDC, an NT server, (Novell servers can also ‘authenticate’ users) contains a database of all of the users and what shares they can access and to what extent they can access them. All of the shares for the entire network are administered here as opposed to being administered on each machine. Obviously, on a larger network, this would be both the more secure and also the more easily managed arrangement.

In order for users on the network to be able to access information or a printer on your machine, you must create a shared resource. Again, this being a home network, we are talking about ‘peer to peer’ with each machine managing its won networked resources. To create a share, you simply open EXPLORER, scroll down to the directory you wish to share, and right click on the highlighted directory. A menu window will open. Choose ‘Sharing’ and the share setup windows will appear. By clicking the ‘Shared’ button, the share name field will light up and will contain the name of the directory as it appears in EXPLORER. You can leave it this way or perhaps rename it. It can be described in the next field. Finally, at the bottom you have a place to set the permissions for who can access the shared resource. Your share will appear under your machine’s name in NETWORK NEIGHBORHOOD on the desktop. The NETWORK NEIGHBORHOOD icon was placed on the desktop when the operating system added networking to your machine earlier in the installation.

When you create a share in EXPLORER, you are in effect, creating a temporary root directory for a drive that will appear as a mapping on other machines. That user will then have access to everything under that ‘root directory’. If you select the entire C: partition for example, the share will be the entire partition. The share name, in this case, comes up C$. The $ will cause the name of the share to be hidden on other machines so delete it when naming your share. This way, the share name will be listed in everyone’s NETWORK NEIGHBORHOOD. Check here to see if what you think should be showing as available shares seem to be missing.

In order for you to access a share on another machine, you must do what is referred to as ‘map-ping’. This is done by clicking the NETWORK NEIGHBORHOOD icon on the desktop and seeing the other machines (and perhaps other workgroups and domains as well). Click on a listed machine and the resources available on that machine appear. If you right click on one of the available folders, a menu windows appears. Choose ‘Map a Drive’ and the first available drive letter will appear in a field with the name of the share. More specifically, the Universal Naming Convention or UNC name will appear. The UNC starts with two backslashes (\\), the name of the machine that the resource is physically located on, and another single backslash followed by the name of the shared resource. You can click the ‘Reconnect on Login’ check box to make it a ‘permanent mapping’ and click OK. That’s all there is to it. That drive letter will now appear in MY COMPUTER and also in the directory tree in EXPLORER. When you click it, you will pull up a window with the directory and all of the files and sub-directories that the resource consists of. You can do with those files whatever the person who created the share has set your permissions for.

Welcome to networking. In no time you will become very comfortable in the Microsoft networking environment, finding another level of organization available to you. Multiple people are able to access and share information from one common source, or from each other. If you have an older machine that has been shutdown because wasn’t fast enough, the addition of a NIC and a little time spent in the operating system can bring that machine back to a useful level again when attached to your network.

For a small home network, even fileservers can be on slow machines. Again, speed isn’t the issue, but hard disk space is. There’s probably good space on the drive in that old machine. Reformat the drive and BINGO, live backup space. Experiment. Have fun. Make use of older PC’s that you have sitting around collecting dust. I have several 386/33’s on my network busy storing information. One happens to be a NetWare server. I rarely type directly to them (although I can) but rather I talk to them through mappings from my much faster ‘workstation’ Pentium machines. Remember,

if you decide to use NT Server for things like DHCP or creating a gateway to a NetWare server, etc., a 486DX is the minimum. NT prefers lots of RAM, not necessarily a quick processor. But it also LOVES Pentiums.

So far we have only used NetBEUI as a protocol. NetBEUI was a protocol developed by IBM and Microsoft around 1984. It is fast and takes up little memory and is fine for small (less than 10 machines) networks. There are limitations to using NetBEUI on large networks. But modern networking allows for the use of two or more protocols at one time on one network card. If you have a ‘Dial-Up-Adapter’, you are probably using a protocol call TCP/IP or Transmission Control Protocol/Internet Protocol. This is a suite of ‘languages’ that were developed by the military that allows communication between many types of dissimilar machines. It is the protocol used on the Internet for this very reason. If you have Internet access, your machine must have TCP/IP running in it to operate the ‘Dial-Up Networking’ or "DUN". By the way, the ‘DUN’ is a logical device, not a physical card in your machine. It works in conjunction with your modem.

Once you have two machines setup with TCP/IP protocol and their unique IP addresses, click START PROGRAMS and click the MS-DOS icon. In the DOS window, type "ping 100.100.100.xxx" where xxx is the IP address can be for either your machine or any other machine on your LAN that has had TCP/IP installed. If TCP/IP is running correctly on your LAN, you should get ‘reply’ back. If not, the cursor will sit while the address is searched for and will eventually time out if it cannot find the IP address requested or state that it is a bad IP address. When you have TCP/IP running correctly, you could remove the NetBEUI protocol and just use TCP/IP. This is an option, but a LAN will perform better with fewer protocols running. ON a home network thought, there won’t be much difference.

Now that you have ICP/IP on your network, each machine has the ability to gain Internet access. Or, with an NT server on your LAN, you can have both PC’s and MAC’s on your network. Out on the Internet, you can find software companies that offer affordable proxy servers. A proxy server is a software layer loaded into a machine that has either an internal or external modem configured in to. The proxy server allows a serial port with a modem connected to it, to be seen on the network much like parallel ports with printers are. Otherwise, a request for a serial port will only get a reply from serial ports on the local machine. They cannot be offered as shared resources. With a proxy server, you can have multiple connections, on multiple machines, to the Internet through a single modem and phone line connection at the same time. You simply go into the ‘proxy’ setup portion of your browser and enter the IP address you assigned to the machine with the modem.

In Netscape for instance, click EDIT, PREFERENCES, ADVANCED, PROXIES. For a normal modem connect, ‘Direct Connect’ would be set. For access through the network, which is an option now in the initial setup of Netscape (and also IE), you click the second choice. Then type in the IP address for the machine that has the LAN modem and the port address (try 8080) in the first four lines and save it. Now when you click the Netscape icon on your desktop, instead of going out to the local modem, it will go out to the LAN modem by way of the proxy server. You can fiddle with it so that the local modem can be used, or you can allow it to start to dial and then click the ‘Cancel’ button causing the dial-out to stop on the local modem and the LAN modem will dial-out instead. Most proxy server programs are licensed for the number of machines that can be attached at one time.

Some additional number of licenses can be added. And if you are one who runs several sessions

of your browser at one time, you can still do that too. Protect your LAN with a ‘Firewall Commun-ications Server’, described above, can be an old machine but it must be able to run Win 95/98. If

you have an NT Server, many of the proxy server programs will run under RAS (Remote Access Services).

Another use of a ‘junk machine’ is a ‘firewall’ machine. This is used when you either have your own full-time on-line Internet server if you have cable modem access. The purpose of the firewall is to place a controlling device between your network and the data on it, and the Internet. The idea of a firewall is to monitor what is coming in from outside in the way of unwanted intrusion either by human or by code. Depending on the software, the amount you are willing to pay for it, and the time you spend on tuning it, the software basically lets in responses that it determines are valid responses for the requests it watched you send out. To set it up, you put two NIC’s into one machine. One NIC is the network card coming in from the modem that your ISP is leasing to you. The other NIC is like any of the other NIC’s you have setup on your LAN.