Revision 937644 of "Topic:Computer network/Definition Network" on enwikiversity

#REDIRECT[[Topic:Introduction to Networking]]
=What is a network?=

The concept of a network is pretty simple. A couple computers have some cable strung between them, and send data back and forth using electrical signaling on the cable. More or less the same as telephones do or, in a very rough sense, like two kids speaking into tin cans connected by a string.

But how does the data actually get from computer A to computer B? How does computer A find the physical location of computer B on the network? If they communicate with electrical signaling, so the data is traveling "at the speed of light", why does it take so long to send a big file across the network?

Gaining a scientific understanding of the answers to these questions may take years, but we shall endeavor to crack them open at the surface, and explore the definition of ''network'' in a way that makes as much sense as two kids speaking into tin cans connected by a string.

=Discussion=
Consider the tin can phone analogy. There are three questions about this primitive network that will help us explore the modern equivalent:
#What is it?
#How does it work?
#Why would you use it?
Notice that ''When was it invented?'' and ''Who invented it?'' aren't on the list. Invariably you will encounter such trivia as you study the protocols and technologies involved in networking. While such details may (or may not) help you pass an exam, chances are you'll never need to recall the name of [[w:George Stibitz|George Stibitz]] to troubleshoot a down T1 line; and while your peers may be impressed when you recall the exact date of [http://tools.ietf.org/html/rfc1 RFC 1], the CTO probably isn't - not as much as by your ability to elegantly describe [[w:Address Resolution Protocol|ARP]] to a frustrated junior technician.

There are so many details (and so much corresponding trivia) in the field of Networking that it's easy to miss the forest for the trees. It is likewise often the mistake of even a seasoned engineer to get mired in the details of a problem, and its lengthy analysis. There is an important parallel between the stumped seasoned engineer and the frustrated junior technician: both are defeated by what they don't know, and enlightened by what they do know.

You probably already know how the tin can telephone works. This doesn't require an intimate understanding of the physics of sound or material science. There are only a couple of things that can go wrong with it, and they're easy to fix. While it is by no means as simple or foolproof to implement a large-scale switched or routed network, these three basic tenets of networking remain constant:

==What is it?==
If you want to design and build a network, you need to know what parts to specify and buy, and how to set them up. If you want to troubleshoot the same network, you need to know what parts are involved, and what can go wrong with them.

In the tin can network, there are only two parts: the tin cans (which may or may not be identical), and the string. There are three places where something might go wrong:
#Tin can A.
#The segment of string between the tin cans.
#Tin can B.
''Thought exercise: Should the operators of the tin cans themselves be considered possible points of failure?''

By understanding what components were used to build the network, and what can go wrong with those individual parts, you will have a much better idea for what ''could'' go wrong with those parts, or how they could be misused, and how to prepare for possible failures or even redesign the network in such a way that these shortcomings are mitigated.

==How does it work?==
This is the technical part of Networking where students tend to feel most uncomfortable. Fortunately, a scientific understanding of all a network's parts is rarely required to design or operate one, and many problems can even be solved simply on the basis of common symptoms. Unfortunately, even ''with'' a scientific level of understanding, networks can be complex enough - because there can be hundreds or thousands of parts involved - that difficult problems hinge on an engineer's ability to ''apply'' that knowledge in a practiced, objective way which will produce results to a series of questions, that ultimately produce a clear and reliable solution.

Some technicians are and always will be under the impression that collecting answers to common problems is the measure of your growth as an engineer. This practice of memorization and regurgitation is both dangerous and fruitless. Too often, this is due to the job pressure most technicians face to "fix it yesterday", which promotes scrambling for some documented solution on Google (ironic considering Google's very own Feeling Lucky feature), rather than carefully dissecting the problem and, if necessary, researching the underlying technology so as to better understand how it works.

''Thought exercise: Technical certifications like the CCNA often place a heavy emphasis on the How of some product or technology. This can easily lead to an obsession with details and memorization, but since most tests are administered by software - and in multiple-choice format - it is difficult to design an exam which accurately measures competency. How might the exam and certification process evolve to overcome this problem?''

==Why would you use it?==
Easily overlooked is the Why of the network. What is its purpose? More importantly, what ''business functions'' does it serve? When part (or all!) of the network goes down, how are those business functions affected, whose work will be most hindered, how valuable is that work to the business, and what natural workarounds exist?

Going back to the tin can telephone as an example, the purpose of the network might be described in a business document like so:
#Jake can contact Anne on very short notice, without calling her house.
#Easy to work together on homework, no need to clean the room!
#It's cool.

This document is already fairly valuable just as the start of a disaster recovery plan. Seeing the reasons for the network makes it easy to understand the impact of a minor or major outage, and plan ahead.

''Thought exercise: Is it the job of management or IT to write such a document?''

Larger networks will exist in multiple parts, or ''segments''. The office network might serve the purposes of file server access and network printing. The DSL line connecting the office to the Internet could be considered a separate segment, with separate purposes. There might be a T1 line connecting a branch office to the main office - which could be critical for day to day business, or may be just a convenience.

==Conclusion==
So what is a network? It will take some time to get familiar with the exact What and How of the answer. The Why, however, is easy to express: sharing resources.

One thing above all that should be clear from this discussion is that there is a difference between ''knowing networks'' and ''knowing a network''. Testing and hiring are usually based on the former, but job responsibilities depend mostly on the latter. The What and Why of a network tend to be specific, while the How is ubiquitous. The How and What are technical, the Why is strategic: as a network technician, it is your calling to be both technical and strategic. And most importantly: the Why always [[wikt:preclude|precludes]] the What and How.

Anyone can learn the technical details required to master the How of Networking, given enough time and discipline. It can certainly be discouraging - and not only at the beginning of your career - to be faced with what seems like an overwhelming amount of what you ''don't'' know. But all networks and the technologies that made them possible are ultimately based on simple business and human needs. So long as you don't lose track of ''why'' all of these protocols, parts and practices exist, you'll find that understanding ''how'' they work isn't that much more complicated than two tin cans connected with a length of string.

=Activities=
These exercises are designed to get you to think and do some research. There are ''suggested answers'' included, but these too are designed to encourage reflection or group discussion, not to be the only possible solutions to the problems.

==Buy a network==
You're working as an intern with a small IT consulting business. The lead technician has been giving you easy jobs with detailed instructions to help you learn the ropes, in exchange for a very lean paycheck.

A local record shop is using your business to set up its computers and install a network. Your lead wants to show you how easy it is to buy and build a simple network, and he's sent you an e-mail with these instructions:

 Hey,
 
 These guys don't need anything fancy. The owner is actually pretty hip to HTML and
 wants to run the web server right there in the store, and give the employees
 Internet access right there at the front counter.
 
 They picked up some used computers from a surplus sale at the high school. These
 things don't have network cards. We need three network cards, three pieces of
 cable, and a switch. I already have the Internet router and the cables for that.
 
 I think this would be a neat opportunity for you to see how easy it is to buy the
 parts for a network and build it. Don't bother with the local Circuit City, their
 prices are crazy and they only carry premium parts. You can try eBay or, my
 personal favorite, NewEgg.com.
 
 For the network cards, the spec you're looking for is a VERY BASIC 10/100Mbps
 speed Ethernet NIC (network interface card), not a 1000Mbps speed (gigabit). Any
 brand will work, though I prefer 3Com and Intel when I can find them for $25 or
 less.
 
 For the cable, we need what's called Category 5e (Cat5e), blue or green (depending
 on what's in stock), 20 or 25 feet each. Should be $5 or less each.
 
 For the switch, I'd prefer a NetGear 10/100 (not Gigabit), with 4-5 ports. Again,
 nothing fancy - price should be about $25.
 
 On NewEgg, I know you can find all of these parts in their Networking store, under
 the Wired Networking category I think. Write me back with a shopping list, and
 I'll make the purchase.
 
 Thanks,
 Gary
 Director of Technology
 Io Networking Inc.

===Suggested answer===
This is more an exercise in using an Internet browser and shopper savvy than any technical skill as a network engineer - but that's the point, network engineers make relentless use of the Internet to do research, stay abreast of current events, and even buy equipment. (A bigger company might use [http://www.cdw.com CDW] or have a direct relationship with the vendor.)

Here are three links on NewEgg to a NIC, a cable, and a switch that would fit the bill:
* [http://www.newegg.com/Product/Product.asp?Item=N82E16833122109 NetGear FA311 10/100 NIC]
* [http://www.newegg.com/Product/Product.asp?Item=N82E16812105307 Generic blue 25' Cat5E cable]
* [http://www.newegg.com/Product/Product.asp?Item=N82E16833122029 NetGear FS605 10/100 switch]

''Custodial note: These links are bound to expire. Faculty and reviewers, and even students, are encouraged to keep these links fresh!''

==The speed of light==
While still working for the same IT consulting business, you make a number of visits to the record shop. On one visit, an employee (a self-admitted luddite) asks you the question, "If computers talk to each other using electricity, and electricity travels at the speed of light, why does it take so long for web pages to load on this computer?"

How might you answer the question?

===Suggested answer===
As a practical exercise, there are many considerations to be borne, here. First, consider that in this actual situation (which you must inevitably encounter as a field technician) you may not know the answer right then and there. This is the first part of the problem: how to maintain your professional image (especially with a business customer) without making something up on the spot. There is certainly no perfect answer to this problem, but common sense social ettiquette should serve as your guide. It's okay to say you don't know. It's even better to say that, your own technical curiosity piqued, you can find out, since you'll be back tomorrow anyhow.

Next comes the problem of answering the technical question itself. And in the case of this particular question, there are ''a lot'' of factors involved. Just a few of them are:
*Electricity does not travel at exactly the speed of light ''through copper'' - there is a propagation delay.
*The capacity of the data link is finite - hence, bandwidth.
*Other completely unrelated factors affect your ''perception'' of this process: for instance, a very slow CPU (or even a fast CPU when significant overtasked) may delay the final ''rendering'' of the web page and its graphics, even if the data for the page is received very fast. (Furthermore, the CPU is often used by the NIC to process and receive data packets, hence a burdened CPU contributes to latency.)
*In the case of the Internet, there are many intermediate systems between source and destination. Any and all of these are subject to and will compound all of the factors above.

Finally is the problem of condensing this massive amount of scientific reasoning into a concise explanation that any luddite should be able to follow. And to come full circle, courtesy matters: assume you are speaking to someone intelligent, even if not necessarily a peer in your own field.

To wit, here's an example:
 Well, there's a lot that goes on between the time your computer asks for the web
 page and gets it. Some of it is happening amazingly fast, but other things slow
 down the whole process. It's actually a lot like taking the freeway across the
 country. In some places there are lots of lanes and no traffic, so you can make
 pretty good time. In other places... not so good.
 
 As some technical trivia, electricity only travels at about 65% to 70% the speed
 of light through copper cabling - you need the vacuum of space itself to go at
 full speed.
 
 But the real killer, I think, is these old computers you guys got from that
 surplus sale. They're so slow it takes them a while just to draw a web page.

And here are a few Wikipedia links with technical information you might use to research this topic for yourself:
*[[w:Velocity_of_propagation|Velocity of Propagation]]
*[[w:Latency_(engineering)|Latency]] (also [[w:Lag|lag]])
*[[w:Comparison_of_latency_and_bandwidth|Comparison of latency and bandwidth]]

==The network is down!==
On a particularly busy day, you're called upon to assist a one-man shop with a small emergency. The details you've been given from your lead are:
*He has a very basic network in his own home office: two computers, DSL, and a new printer with its own network card so the computers can use it over the network.
*Everything was working fine last night. Then there was a power outage. Now the computers can't print to the printer, though the printer is still completely operational on its own.
*The customer needs to print checks today using QuickBooks on the main computer, as soon as absolutely possible.

You are tasked with visiting the customer's home office and providing a solution. Although this is only a hypothetical example and you can't troubleshoot the problem interactively, what would you try first?

===Suggested answer===
You may have recognized that the first two activities were examples of What and How. This is the Why exercise. ''Why does the customer need the printer?'' To print checks. And for that, it doesn't need to be networked - the best thing to try first is to skip lengthy troubleshooting and simply connect the printer to the computer directly (via USB), then get the checks printed. The network connection can be diagnosed afterward, or at a later date.

Anything else that focuses on the solution rather than the problem is also a good answer. Can he copy the files to a floppy or CD and print them from somewhere else? Does he have an old spare printer lying around? Could you borrow a printer from your office and use it at the customer location? And so on. The point is that, to the customer, all that matters is getting the checks printed. Not analyzing exactly what went wrong or getting everything working perfectly.

=Resources=
As this introduction highlights, it isn't crucial to dive right into technical research right away. Better to gain an appreciation for the art first, and then the science. [http://en.wikipedia.org Wikipedia] is an excellent resource for the latter. Here's a list of other resources focusing more on introductory and high-level explanations of networking and networks:

==Websites==
[http://compnetworking.about.com/od/basicnetworkingconcepts/ Basic Networking Concepts on About.com]

==Books==
[http://www.amazon.com/How-Networks-Work-7th/dp/0789732327 How Networks Work]

''Custodial note: Please help expand this list.''

[[Category:Networking]]

[[Category:Networking Intro]]