TechTips; Networking your facility
It’s nearly impossible to build a facility without networked computers – but networking doesn’t take genius.
Nearly every facility built today includes computers. The more computers you have, the more likely it is that your facility will employ a network, possibly more than one. On one level, networks allow resources to be shared and can even simplify facility wiring. On another level, the complication has just begun.
Network fundamentals A variety of topologies and cabling/connectors exist for computer networks, from rings wired with twinax to fiber optic connections. All are designed with the same basic goal: to move digital data from point to point. For the most part, that digital data includes source and destination information, and can navigate from one to the other without intervention. This is different from a traditional routing switcher that requires a crosspoint selection. Within a computer network that “crosspoint selection” occurs automatically.
Today, the most common network topology is a star arrangement that requires a single cable connection from each device to a central hub. More than one type of network can be wired this way, but the most common is Ethernet. Three common flavors of Ethernet exist: 10BaseT, 100BaseT (or Fast Ethernet), and Gigabit Ethernet. Of these, 100BaseT and Gigabit Ethernet are suited for use in video facilities. 10BaseT can be used for control signals and within small business or home environments, but the bandwidth is realistically limited to 1Mb/s. Most businesses of any size use 100BaseT connections between the desktop systems and hubs. Connections between hubs are typically done using a fiber backbone, possibly using Gigabit Ethernet.
Several different network operating systems and protocols exist, but because of the Internet most systems use TCP/IP (transmission control protocol/Internet protocol). TCP/IP makes it easy to integrate Internet communications into your network. This can be done through a high-speed Internet connection and a firewall or firewall software. For most large businesses, Internet connectivity through the network is almost uni-versal. Unfortunately, smaller business don’t always have the know-how or the budget to provide full-time Internet connections to their employees’ desktops.
Network assembly Building a network requires at least two network devices. Some 1394 (FireWire) networks commonly use only two devices, but most networks use more. In these networks, each device is given a unique address and connected to a hub. Low-end hubs do not require an address, although some high-end hubs may. The speed of the connection is determined by the network interface card (NIC) and the hub. Low-priced hubs and NICs are available for 10BaseT and 100BaseT. Most 100BaseT systems are capable of switching from 10 to 100 as needed. Gigabit Ethernet equipment is still relatively new and can be quite pricey, especially in a large installation. Ironically, those large installations are typically the ones that need the bandwidth that Gigabit Ethernet affords.
The current addressing scheme for networks using TCP/IP is IPv4, which uses 32-bit addresses. However, the next-generation addressing scheme, IPv6, uses 128-bit addresses. IPv4 addresses are written in a dotted quad format as four numbers (0-255) separated by periods (dots). An example address is 22.214.171.124. Originally, addresses were denoted by class. Class A addresses started with 0 to 127. There were only 128 Class A networks, but each had the capability of more than 16 million addresses. Class B addresses began with 128 to 191 and could handle more than 65,000 devices. There were 2 million Class C networks, but each offered only 254 addresses. As networking and the Internet grew, it became apparent that a better scheme was needed, as many addresses were not being utilized even though there was (and is) a shortage of IP addresses.
Classless addressing offered a solution. The combination of an IP address with a network mask (a second 32-bit number) allows network addressing to be managed more effectively. An example of a Class C network is 192.168.234.0. This network would use a network mask of 255.255.255.0, or 24 ones followed by eight zeros. Within a network, the netmask simplifies routing, as everything “masked” by a “1” can be ignored. This can also be written as 192.168.234.0/24. Class B addresses can be written as x.x.x.x/16, and Class A addresses as x.x.x.x/8.
Because each device on a TCP/IP network must have a unique address, several addresses have been set aside for private networks. These addresses can be used inside a firewall and will not conflict with any addresses on the Internet. These addresses are (Class A) 10.0.0.0 through 10.255.255.255; (Class B) 172.16.0.0 through 126.96.36.199; and (Class C) 192.168.0.0 to 192.168.255.255. Additionally, the address 0.0.0.0 is used to indicate a default route in routing tables, and 127.0.0.0 is used for the local host (loopback).
Knowing all those numbers allows you to set up a network that can be connected to the Internet without causing a problem. Though you could choose any of the addresses set aside for private networks, most facilities will only require an address pool along the lines of a Class C network. Choosing an address of 192.168.0.1 for your server, and 192.168.0.2 and 192.168.0.3 for your network printers would leave 192.168.0.4 to 192.168.0.14 for workstations. The netmask would then be 255.255.255.240 (28 ones followed by four zeros). In this case, 192.168.0.15 would be the highest address allowed, which would then become the broadcast address. Normally, 192.168.0.255 would be the broadcast address on a Class C network, but that assumes a netmask of 255.255.255.0. The broadcast address contains all ones where the netmask has all zeros.
Servers Servers are used as a central resource available to the network clients. Clients include desktop machines, network printers, and other resources. At the simplest level, a Win95/98 machine with a single shared directory could be used as a server. This machine could even be used by someone as a desktop machine, although performance would be reduced. More often a machine or group of machines is dedicated as the server. Custom-built servers typically have redundant systems and extra drive slots for increased storage.
Like their desktop counterparts, servers must have operating systems. Common network operating systems include Windows NT Server, Novell Netware, and Linux/Unix. The server’s task is to manage the network resources as appropriate. This includes ensuring that network logins are authenticated and that access to network resources such as files and printers is restricted as needed. For instance, only some members of accounting should be able to access payroll records. Or you might have a project that only a few members of your team are allowed to work on. In either case, the server is set up to manage these access rights.
Depending on the demand, one machine may provide several functions. These functions might include file server and print server for several printers. If performance is an issue, those functions might be divided across several machines. One machine could serve some database files, while another handles such everyday business activities as accounting and word processing.
In the same manner, each printer could have a dedicated server. In an environment that prints large, graphic-intensive files, servers can pay for themselves quickly. This is because the print job can be spooled off the artist’s machine and onto the server much more quickly than it can be sent to a printer. Once the file has spooled to the server, the artist can again be productive, rather than wait for the job to be sent to the printer. A similar scenario could occur in a facility handling large video files.
DNS and DHCP Two acronyms you are likely to encounter when setting up a network are DNS and DHCP. Neither is required, but each can make your life a little easier. DNS stands for domain name service, while DHCP is dynamic host configuration protocol. Humans have a difficult time remembering a long string of numbers, but we find it quite easy to remember something like www.videosystems.com. DNS provides a mechanism to equate the IP address with the alphanumeric name.
Within most network-ready computers is a file called “Hosts.” This file contains the IP addresses and names of several computers on the local network and possibly on the Internet. For instance, the file might contain the names and addresses of the following: the local machine, the local gateway to the Internet, and one or two DNS servers on the Internet. You can set up a DNS server for your local network, but this can be a large task, depending on your network OS.
Among other things, DHCP is used to assign IP addresses dynamically. This is done so that someone doesn’t have to manually keep track of which address is used for which machine. Although this is quite simple in a small network, it gets harder and harder to manage as the network size increases. DHCP allows a group of addresses to be assigned by the server. As machines are turned on, each is given an address. Later, when the machine is powered off, the address returns to the pool of available addresses.
Today, it’s relatively easy to set up a computer network. Windows software makes it easy and network hardware is inexpensive and easy to find. Configuring a small network is fairly straightforward, even for a novice. Nevertheless, to get high performance and reliability out of any large network, it’s a good idea to leave the job to a professional network engineer.
COPYRIGHT 2000 PRIMEDIA Business Magazines & Media Inc. All rights reserved.
COPYRIGHT 2003 Gale Group