This page lists all of the answers to the focus questions which are included in each unit. Scroll down or click the unit in the menu on the right to see the answers.

UNIT 1 INFORMATION SHARING

1)Distributed Processing - data processing that is done on several computers, rather than done on one central computer. A local area network facilitates processing within a geographically bounded area. Distributed processing may be wide area networking spanning continents.

2)Time-sharing can also take place on a file server when users must queue up to share the processors resources. Usually, though, the server can facilitate the entire request in one time. True time-sharing takes place on a mainframe when a portion of a user's job is completed within one time-window and continues when that users time-window occurs again.

3)Many IS managers see no real savings in a LAN over a mainframe environment, but there is often greater flexibility in a network, where it is usually easier to "swap" components in and out. Technical support is needed in both environments, but a major advantage of a LAN over a mainframe is that often a user does not have to be "on-line" to the network to use their computer, which can function as a stand-alone computer when not networked.

4)One major disadvantage is loss of one central point of control. Users with their own modems can dial in and out and access other systems, often downloading a virus which can then spread to the network when they log back on. The initial cost of converting from a LAN to a mainframe may be very expensive, and as users install their own versions of software, it may be a burden to the IS department to support.

5)A data communications line must be available to transmit data from one node to another. Usually the telephone system is used, although a private line may be provided by other carriers.

Return to Unit 1 Focus Questions

UNIT 2 FILE SERVERS AND WORKSTATIONS

1) Peer-to-peer networks connect workstations with no central control point. Client-server networks have a file server that directs network traffic and usually is the main "post office" site for e-mail(until the network is large enough to warrant a separate mail server). The file server is also the central repository for shared data and applications.

2)If only e-mail is needed I would recommend a peer-to-peer network because of its ease of implementation and maintenance, as well as relatively low cost.

3) The NIC(Network Interface Card) is the piece of hardware that contains the node address, and the electronics to make the PC behave as a workstation. It works with the client software, the NOS, to provide connectivity to the network. It is the NIC that intercepts the messages, sends them to the user software, and transmits them out of the workstation to the network.

4) A diskless workstation might be the workstation of choice for this type of data. The user could still work on the data, but could not save it to a floppy diskette where it could be removed from the premises, and where it could not be saved to a hard-drive for later copying. The only place the data could be saved would be on the network hard drives via the file server. In addition, I would recommend a firewall be installed between the database server and the rest of the network.

5) A good file server would have to have a LOT of RAM to accommodate all the requests from clients. A fast processor, certainly nothing less than a 200Mhz Pentium, and a fast bus would be priority items. In addition, fast disk drives, equipped with RAID technology and a good backup system, plus a UPS would be essential.

6) If there is a lot of FAX traffic, a FAX server would offload the CPU on the file server. Often an enterprise needs paper copies of signatures or specific forms that might not be available on-line, or when e-mail or voice mail is not feasible.

7) I would estimate that a minimum of 200 Mhz today would be ideal. INTERNET traffic is intense today, and thus generates a lot of message traffic on the server. Coupled with more and more GUI(Graphical User Interface)-based software, requiring more memory and data transmission time, the CPU has to be very, very fast.

8) Yes, the user can store private software if a hard-drive is installed on the PC/Workstation. The user might have a favorite package that the enterprise does not support.

9) A network provides better support, as the enterprise can control which software packages and versions are available, thus restricting a proliferation of software and training. The network supports e-mail and data transmission, facilitating communication.

10) There would be 10 nodes, as the workstation, printers and fax server each count as one node if each is individually connected to the network. If the printers and the fax server can only be accessed through a workstation, then there would only be 6 nodes, one for each workstation. The users are not nodes, only the hardware.

Return to Unit 2 Focus Questions

UNIT 3 CABLING

1) Coaxial cable is the choice if the distance to be covered exceeds the maximum distance of twisted-pair, or if the topology is such that there are devices that could generate interference.

2) Monomode has a wide bandwidth, with a small core that is difficult to handle. It also uses a laser, so it costs more. Multimode has a smaller bandwidth, however it is easier to work with. Graded index multimode is very expensive, but has very high bandwidth and can be extended the furthest.

3) Often the signal is obstructed, and it is confined to one room, as the signal does not penetrate walls. However it is great for small areas where one does not want connectivity with another network.

4) I would recommend coaxial cable as it is strong, can travel the distance, and is good for future implementation. Certainly twisted-pair could be used, but coax would be my choice.

5) I would use coaxial cable of fiber-optics, as twisted pair would be prone to pick up interference from the machinery.

6) Twisted-pair is more flexible and may be easier to "pull" through tight places.

Return to Unit 3 Focus Questions

UNIT 4 HUBS AND ARCHITECTURE

1) A hub is a device that connects many nodes to a central point, which is then connected to the file-server. A hub may be passive and serve merely as a connector, much the way an extension cord connects many appliances to the electrical outlet in your house. An intelligent hub can send information to a software package that the LAN administrator uses to monitor the network.

2) A bus topology is usually likened to a data highway, as the data flows freely along it, and all nodes "hear" the broadcast, but only the destination node acts on it.

3) If I were the LAN administrator, I might first send messages to another workstation to determine if the network is "sane". If it is, I would then disconnect the workstation and reconnect it to another port in the MAU. If a message is then delivered, it appears that the original port is bad.

4) Ethernet would be a good choice, supported with repeaters, and possibly routers and bridges. Token-ring would also work if it was a large enough network with enough small-message traffic to warrant the cost.

5) Token-ring is best for lots of small message traffic, as it is deterministic.

6) Ethernet would be excellent for this application, as there would be little contention.

7) MAU is the IBM term for a concentrator on a Token-Ring network. Essentially, they mean the same thing.

Return to Unit 4 Focus Questions

UNIT 5 MODEMS, PRINTERS, UPS, BACKUP, RAID

1) The good news would probably be that my VISA account would never have to be paid, as they would lose all the account information.

2) Citicorp would probably go down, down, down.

3) I would store them in a salt mine in Utah.

4) If my print job wasn't there, I would probably print the material again. If I were the LAN administrator, I would train the users to monitor the print queues so they could see for themselves that the printers are simply very busy, and they wouldn't reprint the job.

5) I'd want the UPS to stay up long enough to send me a message that a problem exists, and that the system should be "gracefully" powered down. How long that should be depends on how many devices and how large a staff is available at the "worst case" time. If the system is not gracefully powered down, a lot of damage could be done to equipment and data.

6) Depending on the amount of traffic, each workstation with its own modem guarantees the user access to a larger network, but a modem pool might be cheaper in the long run, and would provide more security as it can be disabled when "prime time" is over.

7) A LAN with 20 users is not very big, and RAID is fairly expensive. Unless it was highly sensitive defense department type of data, it is probably not feasible to use RAID. A LAN with 200 users might be justifiable. If the data was highly volatile, such as brokerage information, super-secret design data, or security data, it might be worth the cost.

8) College of DuPage would probably be backed up early Sunday evening, when the labs are closed and most users are not on the system. Citicorp in New York probably does an incremental backup, with a piece of the data being completely backed up each day, but with daily backups done all the time, if RAID is not used, or some other redundant backup system in place. The same would apply to the Pentagon.

9) A dedicated line would be used by a high-volume transaction-based system, perhaps a credit card company, a major credit bureau, anyone doing constant EDI. Yes, it is costly, but the need might justify that cost.

Return to Unit 5 Focus Questions

UNIT 6 CONNECTIVITY AND MAINFRAME-MICRO

1) A repeater would be used when the segment length has been exhausted.

2) Only four repeaters may be used in each segment.

3) In reference to routing messages via bridges, not routers, source-routing has the information needed to route the message at the source, transparent routing means the exact route is not known at the source, but is established as it travels the network.

4) SNA is the means by which data is transmitted from a microprocessor to a mainframe.

5) I would expect to use it on a network connecting to an IBM AS400.

6) The message would go through a bridge or router, and finally on to the "cloud", i.e., the X.25 protocol, which would probably be controlled by a common carrier, such as AT&T.

7) A static bridge has routing tables that must be created and maintained by the LAN administrator, while a dynamic bridge builds its own tables by polling the devices on its segment of the network to store the addresses.

8) The database should contain the users id, types of peripherals, a control number of some sort, serial numbers, model numbers, date acquired, date installed, any information about upgrades, software installed.

Return to Unit 6 Focus Questions

UNIT 7 PROTOCOLS AND STANDARDS

1) The network designer would have to design within the parameters of the standard. At the same time, she would not have to worry about compatibility, so would be able to spend time being creative.

2) Standards guarantee compatibility, so one device could be swapped out and replaced by another. Generic NICs, for example, would be guaranteed to work as well as a brand name.

3) Oh, I would bet the ranch on the standards. Why take a chance? I know products will work together if the standards have been adhered to, I don't know about those products in plain, brown-paper wrapped boxes.

4) Never.

Return to Unit 7 Focus Questions

UNIT 8 OPERATING SYSTEMS/NETWORK OPERATING SYSTEMS

1) An error log should contain the date, time, device name, id, and possibly the user id, if known. It should also contain any and all diagnostic or informational messages that would aid the administrator in preventing future recurrences.

2) A standard script would enable the administrator to set up one script for everyone in a department to share.

3) A group of users can be given the same set of rights, so the work would have to be done only once. Software can also have rights attached to it, so all can share or be locked out.

4) I think standards make life easier, because of all the connectivity attached to it. Open architecture is great, as features can be added-on at will.

5) I would put security on the files to keep non-need-to-know out of the secure data, then give permission to the right users to use/read/update it.

6) Peer-to-peer is an excellent choice for sharing e-mail and documents, and for an administrator not having to have a great deal of technical knowledge. Client-server is considerably more complex and requires more technical skill from the administrator.

7) The network would have to come "down", the hard-drive on the server would have to be reformatted to handle the directory structure that the new NOS uses. The LAN administrator would have to be educated on the new system, some files would have to be converted, drivers would have to be installed, login scripts would have to be written. It would be a LOT of work. In addition to installing the new NOS, great care would have to be taken to give the users transparency, as they should not have to change their commands. This is not a real-world scenario, but pie-in-the-sky, as the user would have to change some habits. Educating the user would have to be carefully planned.

Return to Unit 8 Focus Questions

UNIT 9 NETWORK APPLICATIONS

1) E-mail is saved in the file-server or in a mail-server. When the user logs in and accesses the e-mail feature, that software detects unread messages.

2) E-mail is encrypted and should not be able to be read by anyone but the addressee. The law is hazy at best. E-mail is addressed to the user, but often the facilities and equipment are owned by the enterprise, so it is a vague area.

3) E-mail messages are routed through the mail-server, where they are logged to files for the individual users, who may request to read, save, or delete their mail. The e-mail server software checks the passwords when the user logs in, validates that that user has permission to the e-mail, and logs the date, time, and status of the messages and receipts.

4) A boiler plate could be created and defined solely by an individual responsible for a company function, or it could have to be reviewed or inspected by a team assembled for that function, to determine the correctness of the template.

5) Yes, it is possible that there is a system limit on the number of messages and/or disk-space for e-mail messages.

6) The LAN administrator should be proactive in monitoring the size of the e-mail files, and the dates last accessed, and should periodically "garbage collect." One way to do this is to request that the user purge his/her files every quarter or every six-months to a specified limit of saved messages. If not, then the files can be archived, that is, removed from the immediately available disk and saved for later retrieval.

7) Some NOS software comes with e-mail. Often a third party e-mail package is added to the system.

Return to Unit 9 Focus Questions

UNIT 10 LAN ADMINISTRATION

1) The administrator should have a checklist for review weekly, monthly, quarterly, every six-months, and once a year tasks. In this manner the administrator can be proactive and often prevent problems before they become serious.

2) A firewall can prevent unwanted access, a call-back modem, passwords are all ways to prevent unwanted access.

3) A computer center is a cost center and often an enterprise wants to "charge-back" various departments for their usage. The accounting department might use 40% of the capacity, marketing 40% and various smaller departments the remaining 20%.
Management can make more informed decisions about purchases, leases, upgrades, etc.

4) The administrator can use surveys to discover the satisfaction level of the customer. She can also make random phone calls to see if problems were properly resolved.

5) The LAN administrator should first gather all the available data on the time, date.

6) The time between backups should only be the length of the window of time as long as you are willing to lose data. If you are willing to lose the last 24 hours of data, then back up only every 24 hours.

Return to Unit 10 Focus Questions

UNIT 11 LAN SELECTION & FUTURE TRENDS

1) It depends on the size of the LAN and the complexity of the installation process. If there were no original problems and the cabling was straight forward, the second installation might be the same. Another consideration is if any of the existing components are to be reused. Do they have to be reconfigured? Can you reuse cable? The answer is subjective.

2) The best way is to contact other administrators whom you have met at classes, seminars, product shows, conferences, etc. Have the RFP carefully reviewed if you are the one who wrote it, or be sure at least one other pair of eyes reviews it before it is submitted. Be objective, and get the bid and any promises in writing.

3) Install a mail server. The cost and trouble are well worth it.

4)Yes, they actually are. The NOS encrypts messages before they are sent, they are decrypted when they are received, and because transmission distance is limited, and the walls of the room obstruct signals from leaving, it is usually a secure medium.

5) By adhering to standards, and buying from a reputable dealer, connectivity usually guaranteed.

6) She can subscribe to the "trade rags", the network magazines, many of which are free, or she can read them on-line. She can attend trade shows, workshops, seminars, create her own "network" of professionals.

Return to Unit 11 Focus Questions

C.O.D. Online