UGA Technical Network Liaison Handbook

(last modified 2005-03-04) 
Send comments/suggestions to 
dmm@uga.edu

Table of Contents

0. TNL OVERVIEW
1. BASIC CONCEPTS
1.1 TCP/IP Fundamentals
1.1.1 Client/Server Model
1.1.2 Local Servers
1.1.3 TCP/IP Stack
1.1.4 IP Address
1.1.5 Registered Domain Name
1.1.6 Configuring the TCP/IP Stack
1.1.7 TCP/IP Services and Applications
1.2 Campus Backbone
2. USER EDUCATION
3. NETWORK INFRASTRUCTURE
3.1 Basics
3.1.1 Token Ring
3.1.2 Ethernet
3.1.3 Extended/Combined Topologies
3.2 Infrastructure Documentation
3.2.1 Networking Database
3.2.2 Cable Labeling
3.2.3 Cable Diagrams
3.3 Design Specifications
3.3.1 Structured Wiring
3.3.2 Hub Specifications
4. TROUBLESHOOTING & DISASTER PLANNING
4.1 Network Troubleshooting
4.1.1 Pre-Troubleshooting Checklist
4.1.2 Troubleshooting Fundamentals
4.1.3 Troubleshooting Assistance
4.2 Disaster Recovery
4.2.1 Disaster Recovery Plan
4.2.2 Critical Network Components
5. SYSTEM ADMINISTRATION
6. CLIENTS & SERVERS
6.1 Basics
6.1.1 Communications Software Diagrams
6.1.2 Configuration Documentation
6.2 Network Parameter Management (DHCP)
6.3 Client/Server Networking
6.3.1 Client/Server Diagram
6.3.2 Client/Server Implementation Issues
6.4 Virus Security
7. NETWORK PRINTING

0. TNL OVERVIEW

Technical network liaisons (TNLs) are the staff responsible for managing the networking infrastructure for a campus unit (e.g., a department, center, college or school). The unit's networking infrastructure includes the following components:

Some of the responsibilities of the TNL may include the following:

If you need to contact a TNL in another department, click here to search the list of TNLs at UGA.

1. BASIC CONCEPTS

1.1 TCP/IP Fundamentals

TCP/IP (Transmission Control Protocol/Internet Protocol) is a suite of protocols (or rules) that computers use to transfer data across the Internet. It is not within the scope of this section to discuss how these protocols work. Instead, this section will explain many of standard TCP/IP applications, and what information you will need when configuring them. Although much of this section discusses communications on the Internet, the information can be applied to communicating with other computers on campus, as the campus can be viewed simply as a subset of the Internet as a whole. [NOTE:For a more detailed presentation of TCP/IP communications, you are encouraged to attend the TCP/IP Fundamentals Seminar offered regularly through the Staff Training and Development Center.]

1.1.1 Client/Server Model

Most of the TCP/IP applications that you will use across the Internet operate on a client/server model. In this model, the actual machines and applications that you use to get information (World-Wide-Web browsers, electronic mail programs, News reading programs, IRC chat programs, etc.) are the clients. The machines and programs that provide the information are the servers. It is important to understand that when using the client/server model, a specific machine can be both a client and a server. For example, if you are using your PC to browse the World Wide Web, your PC is a client. When a colleague of yours connects to that same machine to copy a file to his machine, your machine is a server.

1.1.2 Local Servers

 University Computing and Networking Services (EITS) provides several servers on campus that offer many standard Internet services:

1.1.3 TCP/IP Stack

In order for your computer to connect to the Internet, it must be installed with a correctly configured TCP/IP stack. A TCP/IP stack is the software that provides an implementation of the TCP/IP suite of protocols. Conceptually, this software works in "the middle", between the TCP/IP applications running on your computer and your computer's network hardware and associated drivers. Some TCP/IP applications come with their own built-in TCP stack, however most applications expect that you will be running a stand-along stack. The later method is preferred, as you need only configure the stack once, and many applications can use it.

On Macintosh computers, the standard TCP/IP stack is implemented within a multiprotocol communications specificationn called Open Transport. Some older Macintoshes use MacTCP as their TCP/IP stack. There are many stacks available for IBM compatible computers. Most TCP/IP applications that run under Microsoft Windows are written to be Winsock compliant. This means that they should work with any stack that provides standard Winsock compatibility. For Windows 9x, NT or 2000, the Microsoft TCP/IP stack that comes as part of the package is sufficient. Computers using the older versions of Microsoft Windows will need a TCP/IP stack. At the University of Georgia, Peter Tatum's Trumpet Winsock is very popular for Windows 3.x desktops. EITS has purchased a University site-license for this TCP/IP stack. Windows for Workgroup computers need Microsoft's add-on TCP/IP stack which can be found on the University of Georgia's anonymous FTP server ftp.uga.edu.

1.1.4 IP Address

In order to communicate on the Internet, each network adapter card on a computer must be assigned an unique set of numbers. This is a unique four-part, period delimited, address called an Internet Protocol (IP) address or IP number of the form:

www.xxx. yyy. zzz

At the University of Georgia in Athens, the first two parts www.xxx of all IP addresses will be 128.192. In fact, the prefix 128.192 is reserved for University of Georgia use only, so any computer with an IP address starting with 128.192 is a University of Georgia address. The third part yyy of the address identifies what subnet the computer is in. On our campus, this number is determined by what building the computer is in. Although more than one subnet number may be in use within a single building, no single subnet number can be used in one building if it is already in use in another. The final part of the address zzz is assigned to a given machine in the local subnet so that the entire number is unique.

For example, one IP Address on the Athens campus is 128.192. 6. 28. The first and second parts of the address, 128.192, identify the computer as a University of Georgia machine. The third part of the address, 6, identifies the computer as one located in the Computer Services Annex. The final part of the address, 28, is an arbitrary number. By using this system, the number 128.192.6.28 pinpoints a specific machine at the University of Georgia.

The UGA Network Information Center (NIC) is the central authority on campus for assigning IP address ranges to units (usually departments). The NIC will assign to each Domain Network Liaison (DNL) one or more subnet ranges, depending on the size of the network that the DNL supervises. For example, the NIC might assign a DNL the IP range 50-150 within subnet 10. This assignment would give the DNL authority to use use addresses 128.192.10.50, 128.192.10.51, 128.192.10.52, ..., and 128.192.10.150. It is important that each DNL keep track of the IP addresses assigned to machines, making sure that no two machines are ever assigned the same address.

DNLs can contact the NIC through the online MyDNL page https://db.uga.edu/network/dnl/ or by calling the EITS Helpdesk at 2-3106 and asking to speak to the Network Address Space Coordinator.

1.1.5 Registered Domain Name

Even though each machine has an unique IP address, a name is normally assigned to each machine, since remembering names is easier for most people. The name is assigned meaning with the most general part on the right, whereas the IP address is assigned meaning with the most general part of the address on the left. The general form of names for machines at the University of Georgia is as follows:

machine. subdomain. uga.edu

where machine is the specific name for a single machine, subdomain is the name associated with a department or campus unit, uga is the University of Georgia, and edu is an educational institution.

It is also desirable to publish Internet resources via names rather than IP addresses so the actual IP address can be changed without affecting how the user accesses the resource. This idea is particularly important for service names like news.uga.edu. If a user accesses the campus News resource by using the service name news.uga.edu and for some reason the machine that hosts the service goes down, the service can be moved to another machine with another IP address. The service name news.uga.edu can then point to the new IP address and the user will see no change.

Most machine names at the University have 4 parts. Examples of machine names on campus are rigel. econ. uga.edu, pollux. cs. uga.edu, and dogwood. botany. uga.edu. The last three parts of the names reference the machine's subdomain. In the above examples, the subdomains are econ.uga.edu, cs.uga.edu, and botany.uga.edu. The first part of the subdomain usually refers to the department or unit that the machine belongs to. All subdomains at the University of Georgia will end with uga.edu. Although more than one building cannot use the same subnet, more than one building can use the same domain.

The first part of a machine's name, also called the machine's short name, is purely arbitrary. It is often chosen with a bit of whimsy. The short name can only consist of letters, numbers, and the dash character. When a machine's short name is put together with its subdomain name, the long name is said to be the machine's fully qualified domain name (FQDN). The above examples of rigel.econ.uga.edu, pollux.cs.uga.edu, and dogwood.botany.uga.edu are all FQDNs.

In order for the rest of the world (and the campus) to be able to use a machine name to refer to a particular machine, the machine name must first be registered with the Domain Name Service (DNS). The DNLs are given the authority to register the machine names in their area. DNLs can refer to the UGA Domain Network Liaison Handbook for instructions on registering machines in the campus DNS.

It should be noted that machine names are not required to be registered with the DNS. However, many services require that a machine be registered before that machine can make use of their services. For example, a machine that is not registered with the DNS cannot post news to the UGA news server or access the UGA FTP server anonymously.

1.1.6 Configuring the TCP/IP Stack

When configuring a TCP/IP stack, you will be asked to provide several pieces of information that will tell the stack how your machine will operate in its environment. Most of this information will be the same on all machines running at the University of Georgia. Following is an explanation and description of all of the information you should need to properly configure a TCP/IP stack at the University of Georgia.

IP Address

When configuring a TCP/IP stack, you will always be asked to provide the unique IP address for the computer that you are using. This is the address assigned by the domain network liaison for the building in which the machine is located.

Domain Name

When configuring a TCP/IP stack, you will sometimes be asked to provide the domain name for the machine that you are configuring. In most cases, this will not be the fully qualified domain name, instead it would be the subdomain name like econ.uga.edu, cs.uga.edu, or botany.uga.edu.

Network Mask

When configuring a TCP/IP stack you will be asked to specify a network mask (also referred to as the netmask). Simply put, the network mask is used to specify which part of your IP address is the subnet. Your domain network liaison (DNL) can look up the appropriate netmask via the URL https://db.uga.edu/network-bin/dnl/dnlview.cgi (Note: DNLs must supply their UGA MyID credentials to see this information.)

Default Gateway

When configuring TCP/IP you will be asked to specify a default gateway. At the University of Georgia, this information is used to tell your computer where to send packets when they are destined for an address on the other side of an IP router. Your domain network liaison (DNL) can look up the appropriate netmask via the URL https://db.uga.edu/network-bin/dnl/dnlview.cgi (Note: DNLs must supply their UGA MyID credentials to see this information.)

Domain Name Server

When configuring TCP/IP you will be asked to specify one or more addresses for a name server. These are the addresses of the machines on campus providing the mapping between IP addresses and registered machine names. The primary name server address at the University of Georgia is 128.192.1.9. The secondary address is 128.192.1.193. If possible, enter both addresses. If the TCP/IP stack that you are configuring only has room for one address, enter the address of the primary name server.

Time Server

When configuring a TCP/IP stack, you will sometimes be asked to provide the address of one or more time servers. These servers are machines that have been programmed to have very accurate time, and to provide that time to any client requesting it. The time servers provided by EITS for use on campus are time1.uga.edu and time2.uga.edu. If possible, enter both names. If the TCP/IP stack that you are configuring only has room for one name, enter time1.uga.edu. Unfortunately, some stacks will not take a name, and will insist that you enter an IP address. If this is the case, please contact EITS Help Desk helpdesk@uga.edu for assistance.

1.1.7 TCP/IP Services and Applications

Below are some network services that every system administrator is likely to encounter. This encounter may be installing a server or providing the user the ability to access such services.

Electronic Mail

Electronic Mail (e-mail) is a basic service that is used to communicate over the Internet. There are numerous mail programs that the user can utilize to create and read electronic mail messages. Popular ones are Pine and Elm on Unix systems and Pegasus Mail and Eudora Lite for microcomputer desktops. Many TCP/IP applications also include e-mail components. For example, almost all World Wide Web browsers include e-mail components.

For more info on e-mail see the following:

A Beginner's Guide to Effective Email
RFC 821 - Simple Mail Transfer Protocol

Post Office Protocol

The Post Office Protocol (POP) is used to download and read mail on a remote machine. It is typically used by PC and Macintosh users who prefer to read mail in their local environments without actually making a login connection to the environment where they receive mail. Popular POPmail clients for the Microsoft Windows environment are Pegasus and Eudora. A popular client for the Macintosh environment is Eudora.

Internet Mail Access Protocol

The Internet Message Access Protocol (IMAP) is used to manage mail on a remote machine. IMAP differs from POP in that mail is typically stored on a central IMAP server and viewed and managed by the IMAP client program, allowing the mail to be accessed from multiple desktops. It is typically used by PC and Macintosh users who prefer to read mail in their local environments without actually logging in to the mail server as an interactive user. The UGA ARCHES system uses IMAP to deliver electronic mail to graphical desktop clients. EITS supports a number of graphical IMAP clients. For more information about these clients, please see IMAP Client Software Information

FTP

The File Transfer Protocol (FTP) is the standard TCP/IP service for transferring files between computers. When using FTP to connect to another computer, you will typically be asked for a login id and password for the remote computer. However, many computers on the Internet offer what is called anonymous FTP. With anonymous FTP, you can use the word "anonymous" or "ftp" as the login id and your personal e-mail address as the password. After making a successful connection to an anonymous FTP server, you will be able to transfer to your machine only files that have been designated by the owner of that server.

For more information on FTP see:

RFC 765 - File Transfer Protocol
UGA's FTP server - Collection of software for microcomputers, workstations, and other systems. Contains some UGA local programs.

GIL

GALILEO Interconnected Libraries (GIL) is a state-wide library information distributing system (library catalogues, databases, etc.)

For more information on GIL see:

GALILEO Interconnected Libraries

OASIS

The Online Access to Student Information System (OASIS) service provides students a convenient way to register for courses. Students are able to view course availability, register for courses, drop courses, and perform other functions via OASIS.

For more information on OASIS see:

UGA OASIS page

News

News (sometimes referred to as Usenet News) is a software system that distributes articles covering a wide range of topics to many sites around the world. News is similar to electronic mail except that articles that are stored on a "news server" are publicly readable. Articles posted to News are categorized into newsgroups. Each newsgroup is dedicated to the discussion of one particular topic. The topic of the newsgroup may be very general (such as: comp.ai) or very specific (such as: sci.physics.computational.fluid-dynamics). Each newsgroup carries a distribution restriction as well. Most are labeled as "world" so that any articles posted to the group will be distributed to all news server accepting the group you posted to. However, if you want to restrict the audience of your article to just Georgia, you could use the distribution "ga" instead of "world." Every newsgroup belongs to a hierarchy. In the example above, comp.ai belongs to the comp hierarchy (computer related and source code distribution newsgroups). There are many others but the largest are: alt (alternative), comp (mentioned above), rec (recreational activities), sci (scientific and mathematical topics), and soc (social type groups).

There are several local UGA groups available. The most active groups are: 

uga.general            Discussion of general interest items.
uga.tech               Technical discussion (normally computer related)
uga.tech.linux         Technical discussion of the Linux OS
uga.forsale            Items forsale locally
uga.computer-security  Computer security and ethnics discussion
There are a number of other local groups including class groups.

For more information on News see:

USENET NEWS (part of Zen and the Art of the Internet)
Usenet News for Electronic Information Sharing from the Fall 1993 EITS Computer Review

TELNET

TELNET is the standard TCP/IP service for terminal connections on remote computers. With TELNET, you can easily connect to a remote computer and initiate a login session that computer. When using TELNET to connect to another computer, you will typically be asked for a login id and password for the remote computer. Popular TELNET applications are TCP3270 for Microsoft Windows computers, CU-TCP for DOS computers, and NCSA Telnet for Macintosh computers.

TN3270

TN3270 is a special implementation of the TELNET protocol that supports IBM 3270 terminal emulation. IBM 3270 emulation is usually necessary when connecting to IBM mainframe operating systems. At the University, this would include connections with CONTAC, CMS, TSO, MUSIC, and OASIS. TCP3270 (now

WWW

The World Wide Web (WWW) consists of many independent but interlinked servers. The building blocks of the web are hypertext documents. As you browse documents on one server you can be seamlessly shifted to another server. Connections to a WWW server are made with a Web browser (see Netscape and Mosaic) and the locations are provided through a Universal Resource Locator (URL). URLs need not just specify a hypertext document but they can consist of ftp and gopher sites as well. Sample URLs are:

http://www.uga.edu -- Connects to UGA's WWW server
ftp://ftp.uga.edu -- Connects to UGA's FTP server

For more information on WWW see:

W3C - World Wide Web Consortium

Netscape

Netscape is a graphical navigator that allows retrieval and viewing of World Wide Web documents. Netscape is available for most platforms.

For more information on Netscape see:

Welcome to Netscape

MS Internet Explorer

Internet Explorer (IE) is Microsoft's graphical navigator that allows retrieval and viewing of World Wide Web documents. Internet Explorer is available for most platforms.

For more information on IE see:

Internet Explorer Home

Lynx

Lynx is a text mode Web navigator that allows retrieval and viewing of World Wide Web documents and connecting to ftp and gopher servers. The main difference from Netscape and Mosaic is that Lynx allows you to access Web documents from a text only connection. Lynx is available for Unix and VMS platforms. It is the Web access software provided on ARCHES.

1.2 Campus Backbone

Broadband Network Gigabit Network

2. USER EDUCATION
The importance of user education should not be underrated. In the best case, educated users can become support allies in a networked computing environment. This can be achieved when they know enough to keep themselves out of trouble. The following are two key points of information the educated user should have as a resource.

Acceptable Use and Netiquette

The educated user should understand that the net is a place much like other human dwellings -- full of rules and limits which support the community's needs. This is the link to the University of Georgia Policies and Use of Computers document. It is a rather lengthy description of the do's and don'ts vis-a-vis the University's access to the Internet. Here is a summary of this official policy document.

User's Guide and FAQ's

A user's guide can be a key point of reference and communication with the users. It should be developed as a community effort, and as much as possible, the policies and procedures should be widely agreed upon and understood on the front-end of the development process.

Design a User's Guide for future updates

Please note that in designing the guide, or in any documentation design, that these pieces of work require frequent updates to retain their value. It is recommended that once you've determined what information is pertinent to your users that this information be modularized so that it can be referred to in subsequent documents, and so that single updates can be reflected in any related documents.

3. NETWORK INFRASTRUCTURE

3.1 Basics

3.1.1 Token Ring

 IEEE 802.5 standard
4 or 16 megabits/sec
Token Ring Hardware Diagram

3.1.2 Ethernet

 IEEE 802.3 standard
10, 100 or 1000 megabits/sec
Thick Ethernet Hardware Diagram

Thin Ethernet Hardware Diagram

Twisted Pair Ethernet Hardware Diagram

An *excellent* technical resource on Ethernet is Charles Spurgeon's Ethernet Web Page, http://www.ethermanage.com/ethernet/ethernet.html. You are also encouraged to read the UGA Network Design and Implementation Guidelines document for further information regarding the deployment of Ethernet networks.

3.1.3 Extended/Combined Topologies

3.2 Infrastructure Documentation

3.2.1 Networking Database

3.2.2 Cable Labeling

 Suggested cable label: MEDIA_PREFIX - PORT_INDICATOR - LAN_NUMBER (optional)

3.2.3 Cable Diagrams

 Sample LAN Diagram

3.3 Design Specifications

You are encouraged to read the document UGA Network Design and Implementation Guidelines for detailed information regarding the deployment of Ethernet networks.

3.3.1 Structured Wiring (Highly Preferred)

Structured Wiring Diagram

3.3.2 Hub Specifications

 Please read section 3. Ethernet Hub Guidelines in the UGA Network Design and Implementation Guidelines document for a detailed discussion of hub considerations.

4. TROUBLESHOOTING & DISASTER PLANNING

4.1 Network Troubleshooting

4.1.1 Pre-Troubleshooting Checklist

Before problems occur on your network, you need the following on hand:

4.1.2 Troubleshooting Fundamentals

There are six basic steps in troubleshooting a problem. These steps are:

  1. Observe symptoms & write them down in the trouble log
  2. Review network documentation, noting recent changes
  3. Analyze the symptoms (ask questions) to systematically isolate the problem down to the network component level (server, workstation, cabling)
  4. Break down the component into assemblies (cards, disks, software, power supplies, etc.)
  5. Test each assembly to isolate the fault
  6. Take corrective action and note the solution in the trouble log

(Note: Troubleshooting is an art that is developed with experience. The above steps are merely a general guideline to the troubleshooting process.)

4.1.3 Troubleshooting Assistance

4.2 Disaster Recovery

4.2.1 Disaster Recovery Plan

A network disaster is any network outage that costs your users considerable downtime and lack of productivity. The only way to successfully deal with disasters is to develop a disaster recovery plan BEFORE the event occurs. An excellent overview of site security and disaster planning can be found in the Internet Request for Comments (RFC) document 1244. (Select this pointer to download RFC 1244 [253 KB, text].)

A disaster recovery plan should answer the following questions:

4.2.2 Critical Network Components

The critical components in the operation of a network include:

Most of the critical network components can be generally repaired in a short period of time, especially if spare parts are readily available. A file server hard disk crash, however, can take some time to recover from since it requires reinstalling the server operating system and restoring the application and data files.

5. SYSTEM ADMINISTRATION
There are the following general tasks that any system administrator must deal with regardless of the operating system they are supporting:

Account Management

 The administrator has to be concerned with the following items:

Backup Strategy

 Here there are several facets listed below:

Security

Software Installation and Maintenance

Hardware Maintenance

This will normally consist of obtaining some sort of Maintenance support for your hardware to provide repair in case of hardware failure and to provide upgrades of vendor firmware required for later versions of software. The system manager usually will do the research to determine the best plan of maintenance for his circumstances and reccomend it to his management for funding and implementation.

Capacity Planning

The system manager should be keeping records of processor utilization and disk space utilization to be able to project when additional processor capacity or disk space will be needed to meet the needs of the user community. The system accounting system can be used to collect much of the data needed.

System Tuning

This is a catch all for the various tasks needed to keep the system running well and responsively.

References

 Novell References

Windows NT References

Unix References

A general reference for all administrators

6. CLIENTS & SERVERS

6.1 Basics

6.1.1 Communications Software Diagrams

 Windows Software Diagram

Macintosh Software Diagram

6.1.2 Configuration Documentation

6.2 Network Parameter Management (DHCP)

The primary method for managing IP addresses centrally within a *single campus backbone* is DHCP. DHCP clients get their IP information (IP address, netmask, default gateway, nameservers, etc.) from a respective server, keying off the clients' network interface card (nic) addresses.

6.3 Client/Server Networking

6.3.1 Client/Server Diagram

6.3.2 Client/Server Implementation Issues

6.4 Virus Security

7. NETWORK PRINTING