To send a letter to someone, you need to know their address. The same is true on the Internet. To you the address may be frodo@aol.com, but in reality this is translated by the Internet to a type of numeric value. If you wish go to a URL such as http://www.cnn.com, this is translated to a type of numeric value for the actual address, then the Internet sends you the page at this address. More on this in a minute.

Protocols

If you are in a meeting, you have certain rules of order that are used so that everyone isn't talking at once. Most of the time we use the rules defined in Roberts Rules of Order - although you may have never read that book. You do know, however, that if you wish to speak you raise your hand and avoid speaking until the moderator recognizes you.

It is the same with an Ethernet network and with the Internet. The Internet transmits data in something called packets, each 1500 bytes. Each packet has some overhead information about the address to which it is to go, where it fits relative to other packets in your total transmission, and some error-checking information. This system was birthed during the cold war with concerns about a city being destroyed. The packets could be routed through the network in multiple ways and then reassembled at the destination. If one path was down, a packet would automatically be sent another way. This means the packets of a particular message you get (email, web page, or file) could arrive at your system using multiple routes. It is then reassembled just before it reaches you. The system, or protocol, for this transmission is defined as the Internet Protocol, or IP. The Internet addressing scheme is defined within that protocol.

There is a second protocol that is closely related to the Internet Protocol, and this one is called the Terminal Control Protocol, or TCP. It makes sure packets get where they are going and are reassembled in the right order. The TCP and IP protocols are so closely related that they are often referred to as the TCP/IP protocol. For most people doing networking and broadband Internet, this is the only networking protocol you need to install on your computer. Many manufacturers ship their computers with other protocols installed as well (such as NETBUI); but this is done to minimized technical support calls. Turning the other protocols off speeds up the networking process, relieving the computer of having to convert the data for each protocol.

The IP Address

Now let's look at the IP addressing. Our web site (www.netadventures.biz) has an IP of:

208.56.131.130

You can type this(http://208.56.131.130) into the browser instead of the name and you will still see our home page come up. This IP numeric address has two parts. One part defines the network on the Internet, the second defines a computer (hosting system) on that network. As to the question of which part of this number goes to which part - we will answer that in a little bit. Stay with me.

Notice that the address is a 32-bit number, broken into four eight-bit values. Each of the eight-bit values is called an octet, and can have a value from 0-255. Each of the four octets is separated by periods. The first part consists of 1-3 octets and defines a network on the Interent. The rest of the IP defines a computer host on that network.

A given IP (which represents a host on a network) belongs to one of three classes. The value of the first octet determines the class.

There are a few more classes; but this will do for the moment. For a Class A IP, the network IP is determined by the first octet only. For a Class B, the first two octets define the network. For a Class C, the first three octets determine the network. From this you can see that my business IP I just referenced is on a Class C network and that network'ss IP is 208.56.131.0. The network IP is always the octets for the network with the value of 0 for any host system.

The Internet is a network of networks. With millions of messages composed of multiple packets, you can easily see there would be a lot of collisions on the Interent if there wasn't some way to break things down into subnetworks. This is done using a subnet mask.

The subnet mask has two components. The first component defines which part of the IP is Part 1 of the IP and defines the network. The second part, Part 2, defines the host system on that Internet network. A subnet mask of 255.255.255.0, for example, defines a Class C network. A subnet mask of 255.0.0.0 defines a Class A network. At the moment I'm logged into Comcast and the actual IP is 24.20.76.253. The subnet mask is 255.255.240.0. You could think of this as a binary mask that blocks off from the actual IP the part that defines the network. A value of 255 in a mask means all bits of that octet are set to 1 and that entire portion of the IP is masked off. The rest of the IP defines a host system on that network. You are defined on the network by two values: Your IP and your subnet mask.

Network IP Rules:

• All hosts on the same physical network must have the same network number
• All host IDs on the same nutwork must be unique.
• You cannot assign a host ID with all the host bits set to 1 (octet 255 for a Class C). This value is reserved.
• You cannot assign a host ID with all the host bits set to 0 (octet 0 for a Class C). This value is reserved for the network number

Note: You must always use the subnet mask. It works with the IP to identify the network. Neither the IP or subnet mask can stand alone. Routers use it to determine the network to which the packet should be forwarded.

How the Router Works

Now look at a typical wired network for a small office or home. In fact, this is my own with one exception - the router/DHCP/switcher in my network is a wireless gateway, or access point.

The router separates the Internet network from my local network. The DHCP in the router product acts as a small server system. When a computer in my local network needs an address from the Internet, it asks my DHCP for an address and the DHCP assigns a local IP. The NAT in the router then requests and obtains an IP from the Internet DHCP. The Internet sees a completely different address for this system than any address you see here. The router then converts between that address and any local computer address.

Notice that all local addresses begin with 192.168.1. This is the "Part 1" of any local address and defines the network. Any local network using Linksys equipment will have this same "Part 1". It is never seen by the larger Internet network, as the router converts everything. When an Internet network is defined by the first three octet groups it is called a Class C network. A Class C network can have as many as 2,097,152 possible addresses, but supports only 254 possible hosts (the other two addresses are reserved). A Class A network, in comparison, is defined by the single first octet. There can be only 126 Class A networks; but each can support over 12 million hosts. A class B network would be defined by the first two octets.

The router product also blocks any attempt from outside the local network to access the local computers using their 192.168.1.XXX address. The router knows this is a local number only, and protects the network from any outside intrusion using this number. In other words, you have a firewall here as well.

Here are the local addresses permitted for each class. None of these ranges can be seen by the larger Internet.

If you are already using a network, you can find out the IP addresses used by the network by switching your computer to the command mode and entering the following to the DOS prompt:

ipconfig -all

This will give you a lot of information on your network connection. You will see the IP address of your computer, the IP address of the DHCP, and a physical address of your computer. Every addressable component in your network has a physical address. This is a hexadecimal number that is permanently stored in the client adapter card, router, or switcher. It will also be on a label attached to the bottom of the unit. This physical address is unique. No other device on any other Internet network has this physical address.

Another thing you can do is ping a network computer. From the command prompt on my main computer I can do:

ping 192.168.1.103

This will generate a response if the "103" computer is online.

With your own local network with its DHCP host, you can hand out your own IPs on your own local network. You can add IPs for webcams, printers, game consoles, and more.

When you bring your computer up on the Internet, your local network router requests an IP from the Internet's host DHCP server. This gives you a lease on an IP for your local network. At the same time, the Internet server returns to you a subnet mask. The subnet mask defines what part of the IP address that is assigned to you is the network address in the IP and which part of the IP address is that of the host system on that network.

Finding Your Internet IP and Subnet

If you are NOT using a router, this is simple. Go to the command mode on your computer and enter IPCONFIG -all. The screen will display the IP, subnet, DNS server, physical address, and more. The DNS server is the system on the Internet that converts a domain name, such as netadventures.biz, to its IP address.

If you are using a router, this is more complicated as the router shields your system from the actual Internet addressing. Try the above on your system and you will get the IP and subnet assigned by the DHCP in the router to your computer. To find the actual address you are using on the Internet, you will need to access the configuration screen of the router (or wireless access point if you are using that - which as a router in it.) For a Linksys wireless access point, this is http://192.168.1.1/ on your browser. For a Linksys WAP on the configuration screen you would then select the Status option. You will then see the IP, subnet, DNS, physical address and more you are using with the Internet at that time.

One quick troubleshooting trick when your system locks up on the Internet is to close the browser or email program, shut down your computer, modem, and router (any order). After 30 seconds cycle up your system in this order:

1. Bring up the modem. Wait for its' diagnostics to complete by watching the lights on the modem.
2. Bring up the router. Wait for the router diagnostics to complete. The router DHCP requests an IP and then receives this and is ready.
3. Boot up your computer. It will boot and then request an IP from the DHCP in the router. This IP will start with the 192 octet. You computer is now online
4. Repeat the last step with any other computers on the network.

Systems Without DHCP

This addressing scheme works great for networks that have Internet connections and a router. Suppose, for example, you have a local network but no router. In specific, what happens if the network has no DHCP to assign IP numbers to the various network components? For simple computer networks, this isn't a problem. Switchers and hubs still work using the physical addresses on the client adapters (NICs). It is not unusual, however, for a system to be smarter than this and need IPs on the components for the system to work - with no router DHCP to assign these.

Microsoft has already solved this by putting a technology in most versions of Windows that automatically assigns an IP to the computer when there is no DHCP to do the job. The computer gives itself an IP. This is present in Windows 2000, Windows XP, and Windows 2003 Server.

The technology is called APIPA, or Automatic Private IP Addressing. If a computer needs an address and can't find a DHCP server on the network to assign it, the APIPA in Windows assigns one from the block 169.254.0.1 - 169.254.255.254. This range is reserved for Microsoft and does not exist outside of a local network. A Class B subnet mask of 255.255.0.0 is assigned. The APIPA monitors the local network to protect against duplicate addressing. Moreover, the APIPA checks the local network every five minutes just in case a DHCP is temporarily down. If a DHCP is found, the APIPA drops out and the DHCP resumes the IP assignments. You can tell if you network is using the APIPA by entering at a computer's command mode:

ipconfig -all

If the IP starts with 169, you are on APIPA. If it starts with 192, you are using a DHCP.

An example of a network with no DHCP is a wireless ad-hoc network. For example, a a group of people in a conference room with a wireless network with no router or DHCP. The adapters are configured "ad-hoc", which means they have no hub and only communicate with each other.

Addresses and Ports

A single host may be serving many types of software: web pages, email, and file transfers. Each of these has its own host software on the host for serving that application. The servers determine the type of application to use for a request by the port number assigned to the address in the TCP protocol. Web servers normally use port 80, for example. Normally you don't need to be concerned about these ports. The software handles this automatically.

Another type of port is also important. On the back of a switcher or hub, you will see connections for multiple computers. Each of these is also called a port. There is no relationship between this hardware port and the TCP port (such as 80) mentioned in the last paragraph.

Static and Dynamic Addresses

Due to the large Internet growth, it became apparent that the Internet would soon run out of IP addresses for everyone. Fortunately, everyone is not on the Internet at the same time. Your host is assigned a block of IP numbers. When you log into the Internet, your host assigns you a number from their block. This is a dynamic number; that is, the next time you log in it you will probably get a different one. To prevent IP addresses in the pool from accidentally being lost due to glitches, the host will periodically terminate your IP. If you are connected when this happens, the host will simply assign you a new IP and your connection is never broken.

If you have a Web site, it normally has a fixed or static IP. The 208.56.131.130 IP is a static, or fixed, IP of my web site. All of the web sites we host have static, for fixed, IPs. Use caution with other hosting systems. Some use virtual IP addresses for your web site, sharing IPs between multiple web sites. This is a dangerous practice as if another site on the host sharing your IP starts spamming and gets black listed, your own site is black listed as well. For hosting, you always want a static IP.

Media Access Control (MAC) Addressing

There is another address used to identify each component in your network called the MAC, or media control address. You will find this MAC number on any adapter card, switcher, router, Wireless Access Point, or modem. In short, any addressable component of any network. This is a twelve-digit number in which each digit is hexadecimal; that is, the digit can be any value form 0 to F. I have a Dell Ethernet card made by 3Com that has the MAC address 0050045B3CB1. The first half of this address represents the manufacturer, and the second half is a unit from that manufacturer. The numbers are never duplicated. This system permits a given manufacturer to make 16 million devices with unique numbers.

The MAC address is often called the physical address of your system. It is actually burned into the device. You can override it with software; but this isn't advised. Doing so can cause problems on the network if you duplicate an existing MAC.

Each time you log onto the Internet, your IP may have changed but your physical address doesn't unless you change your hardware. This physical address is broadcast into the Internet. You host can track that. If someone manages to come into your wireless network at 2 a.m. and do a little spamming off your IP; their MAC address gets sent into the host for tracking where the spam came from.

Let us help you with your network!

8/19/2004