TCP/IP First-Step [Electronic resources] نسخه متنی

اینجــــا یک کتابخانه دیجیتالی است

با بیش از 100000 منبع الکترونیکی رایگان به زبان فارسی ، عربی و انگلیسی

TCP/IP First-Step [Electronic resources] - نسخه متنی

Mark A. Sportack

| نمايش فراداده ، افزودن یک نقد و بررسی
افزودن به کتابخانه شخصی
ارسال به دوستان
جستجو در متن کتاب
بیشتر
تنظیمات قلم

فونت

اندازه قلم

+ - پیش فرض

حالت نمایش

روز نیمروز شب
جستجو در لغت نامه
بیشتر
لیست موضوعات
توضیحات
افزودن یادداشت جدید






The Big Picture


TCP/IP, as a suite of protocols, is divided into several large components, each of which focuses on a different set of related functions. Remember the example from Chapter 3, where you grouped the related car-driving tasks into layers for a reference model? That exercise was a nice way of providing a framework for understanding how TCP/IP is constructed.

Just like driving a car, those sets of related functions can be broken down into individual tasks. Preceding sections use the word tools, but the words tools and tasks are functionally equivalent. In TCP/IP, those tools or tasks are known as protocols. A protocol is a small piece of software designed to perform one small, specific task.

As you might have guessed, those original small utilities created by the Internet's first group of engineers became the protocols that now comprise TCP/IP! Of course, a lot of other protocols representing new features and functions have been added over time. The result is a comprehensive and complicated suite of communications rules and tools.

Dissecting TCP/IP


A good way to understand how something works is to take it apart. OK, so that's my personal rationalization for all of those things I've disassembled in my life and reassembledwith varying degrees of success. Without getting into too much detail, the electric toy trains proved unsalvageable, but I hear the GTO did eventually run again! But, I learned a LOT by taking them apart, so dissecting things is good way to learn…if you are prepared to lose the patient during the process.

Unlike cars or toy trains, you can't physically disassemble software. However, you can logically dissect TCP/IP. The TCP/IP protocol stack includes four functional layers. These four layers follow, from the highest level of functionality to the lowest:

Process/application
Provides tools for users

Host-to-host
Provides tools for users' applications

Internet
Provides tools for the network

Network access
Is the network itself and not a native part of TCP/IP


Figure 4-9 contains the TCP/IP reference model, including how these four layers stack up.

Figure 4-9. TCP/IP Reference Model

If you look closely at the layers in Figure 4-9, you see that the functional layers referred to in this chapter correlate nicely to the actual layers of the TCP/IP reference model. The one noticeable difference is that this chapter has mentioned three layers of related functions; the TCP/IP reference model contains four! Don't worryI can explain that discrepancy! You see, even though the TCP/IP reference model identifies all the functional layers required to support open communications through a network, TCP/IP itself doesn't directly provide tools for all those layers. The layers of tools that TCP/IP does provide are the following:

Process/application layer
The process/application layer is equivalent to tools for users. When you stop and think about it, that's exactly what an application is.

Host-to-host layer
The host-to-host layer is where tools for applications reside. This layer contains all the mechanisms applications need to talk to each other on different computers.

Internet layer
The internet (lowercase i intentional) layer is equivalent to tools for the network.


A simple process of elimination reveals that network access is the difference between the TCP/IP reference model and the functional layers that were introduced to you. The reason for this disparity is simple: Even though TCP/IP's reference model includes a layer for physical network access, it is not a native component of the TCP/IP protocol suite! That protocol suite assumes you already have a physical network.

Now that you've seen the big pictureTCP/IP's architecture and its major componentstake a slightly closer look at each of its four functional layers. As you walk through these layers, you probably notice the strong similarities between these layers and the groups of functions explained earlier in this chapter. That was not an accident!

Process/Application Layer


The process/application layer contains just what you might expectprocesses and applications! As mentioned earlier, TCP/IP contains some basic applications. These applications live in this functional layer. However, as the name process/application layer indicates, some processes live there, too! The difference between a process and an application lies in their functionality. An application lets you do something using a native component of TCP/IP. A process allows you to use other applications that aren't native components of TCP/IP and have those applications communicate using TCP/IP.

Some utilities examples in this layer include those mentioned earlier in this chaptersimple, but useful tools that let you send and receive files, check on the reachability of a remote computer, log on to a remote computer, and check the communications path through a network.

Later, you are introduced to these tools and shown how to use them. For now, continue looking at the architecture and how the groups of functions are related.

Host-to-Host Layer


The IP host-to-host layer is where Transmission Control Protocol (TCP) and User Datagram Protocol (UDP) live. Why do you need two different host-to-host layer protocol groups? The answer lies hidden because applications' needs can vary widely. So widely, in fact, that the two main approaches to conveying data through a network are almost exact opposites.Chapter 2, explains, applications tend to be one of two types: those needing reliable data delivery ("good") and those needing timely ("fast") data delivery. You can't have reliable and timely, so pick the one that's best for your application's needs. Applications that require reliable delivery for their data must use TCP. Applications that require fast or timely delivery use UDP.

Internet Layer


Continuing with the dissection of TCP/IP's layers, the internet layer is next. This layer consists of all the protocols and procedures necessary to allow data communications between computers (regardless of how near or far apart those two machines might actually be)!

Classifying this group of tools for the network stops short of actually including the network. The network is yet another set of completely separate functions that lies outside TCP/IP boundaries. This distinction is subtle but critical: TCP/IP is not a network. It is a communications protocol. As such, it absolutely needs a network before it is useful. Because it doesn't include a physical network, it is imperative that TCP/IP spell out clearly all the rules for communicating across a network. All networks that use TCP/IP for communications must follow the same set of rules. The Internet Protocol (IP) spells out these rules.

The heart and soul of IP is its address system. The IP addressing system enables each machine in the world to have a unique address. More importantly, it enables networks to find different paths between source and destination machines, compare them mathematically, and pick the best path. That science is known as routing. Entire books are dedicated to routing; one paragraph of one chapter in a TCP/IP book cannot do that topic justice. Suffice it to say that IP offers invaluable capabilities that transcend the limited host-based interactions described here.

Network Access Layer


Different types of networks exist, including local-area networks (LANs) and wide-area networks (WANs). These networks are self contained in that they hold all the logic and tools needed to operate. They stop there! None contain the facilities for working with an application to get data to or from the network. That's why you need TCP/IP.

The network access layer is included in the TCP/IP reference model because it contains some critical functions. This layer describes the physical and logical mechanisms you need to actually send data across a LAN or a WAN. Just because TCP/IP needs this set of functions, however, doesn't mean it has to re-create them.

Many standards and technologies provide this capability. TCP/IP merely has to acknowledge they exist and develop a way to work with them.

This layer winds up being a placeholder as far as TCP/IP is concerned. TCP/IP contains no tools or mechanisms that correlate to the network access layer. However, TCP/IP does expect that some other technology (such as an Ethernet LAN) will provide these functions.


/ 133