Del Mar College
CIS 306 - Managing NOVELL® Networks
Instructor: Michael P. Harris
4. Controlling Data Transmission
Once you have the hardware we've discussed so far, you can start connecting the various pieces into a network.
But simply connecting hardware doesn't make a computer network. Even though the hardware is capable of generating signals and transmitting them across a medium, it must be told when and how to do this. There must be network communication software to tell the hardware when and how to transmit. The software and hardware on all parts of the network must work together to enable the transmission of data from one networked computer to another. We'll explore various networking software a little later. First, let's look at the communication model that is the basis for controlling data transmission on computer networks.
To guarantee reliable transmission of data, there must be an agreed method that governs how data is sent and received. For example, how does a sending computer indicate which computer it is sending data to? And, if the data will be passed through intervening devices, how are these devices to understand how to handle the data so that it will get to the intended destination? And, what if the sending and receiving computers use different data formats and data exchange conventions—how will data be translated to allow its exchange? These are only a few of the questions that must be answered before data can be reliably transmitted and received across a computer network.
Understanding the Open Systems Interconnection (OSI) model will allow you to understand how data can be transferred between two networked computers, regardless of whether they are on the same network, or are the same type of computer, or use the same data formats and exchange conventions.
ISO and the OSI Model
The OSI model was developed by the International Standards Organization (ISO) as a guideline for developing standards to enable the interconnection of dissimilar computing devices. It is important to understand that the OSI model is not itself a communication standard. In other words, it is not an agreed method that governs how data is sent and received; it is only a guideline for developing such standards.
The OSI Model: What It Is and Why It's Important
It would be difficult to overstate the importance of the OSI model. Virtually all networking vendors and users understand how important it is that network computing products adhere to and fully support the networking standards the model has spawned. The reasons are logical.
First, when a vendor's products adhere to the standards the OSI model has spawned, connecting those products to other vendors' products is relatively simple. Conversely, the further a vendor departs from those standards, the more difficult it becomes to connect that vendor's products to those of other vendors. Second, if a vendor were to depart from the communication standards the model has spawned, software development efforts would be very difficult because the vendor would have to build every part of all necessary software, rather than often being able to build on the existing work of other vendors.
The first two problems give rise to a third significant problem for vendors: A vendor's products become less marketable as they become more difficult to connect with other vendors' products unless the introduction of the vendor's products is well ahead of the introduction of other such products into the general marketplace.
Now, keeping in mind the purpose of the OSI model, let's take a look at its structure.
The Seven Layers of the OSI Model
Because the task of controlling communications across a computer network is too complex to be defined by one standard, the ISO divided the task into seven subtasks. Thus, the OSI model contains seven layers, each named to correspond to one of the seven defined subtasks.
Each layer of the OSI model contains a logically grouped subset of the functions required for controlling network communications. The seven layers of the OSI model and the general purpose of each are shown in Figure 15.
Figure 15: The OSI model
Standards and Protocols
National and international standards organizations have developed standards for each of the seven OSI layers. These standards define methods for controlling the communication functions of one or more layers of the OSI model and, if necessary, for interfacing those functions to the layer above and below.
A standard for any layer of the OSI model specifies the communication services to be provided and a protocol that will be used as a means to provide those services. A protocol is a set of rules network devices must follow (at any OSI layer) to communicate. A protocol consists of the control functions, the control codes, and the procedures necessary for successfully transferring data.
For every layer of the OSI model, there is more than one protocol standard. This is because a number of standards were proposed for each layer and because the various organizations that defined those standards—specifically, the standards committees inside these organizations—decided that more than one of the proposed standards had real merits. Thus, they allowed for the use of different standards to satisfy different networking needs.
Network Communications Through the OSI Model
Using the seven layers of the OSI model, we can explore more fully how data can be transferred between two networked computers. Figure 16 uses the OSI model to illustrate how such communications are accomplished.
Figure 16: Networked computers communicating through the OSI model
Our figure represents two networked computers, each of which is running various pieces of software (most not shown). Running together, the various pieces of software implement the seven OSI layers. These computers are identical: They are running identical software, and they are using identical protocols at all OSI layers. Above the OSI application layer, each computer is running an E-mail program. The E-mail program enables the users of the two computers to exchange messages. Our figure represents the transmission of one brief message from computer A to computer B.
The transmission starts with the user of computer A pressing a key to send a mail message to the user of computer B. The E-mail application is designed to talk to the OSI application layer—it knows the proper protocol for doing so. The E-mail application transfers the message to the OSI application layer. Using the functions built into its protocol, the application layer accepts the message data and adds an application-layer header to it. The application-layer header contains the information necessary for the application layer in computer B to correctly handle the data when computer B receives it.
After adding its header, the application layer in computer A passes the data to the presentation layer below. The presentation layer treats everything received as data, including the application-layer header, and appends its own header (the technical term for this is "encapsulation"). The presentation-layer header contains the information necessary for the presentation layer in computer B to correctly handle the data. After adding its header, the presentation layer transfers the new data unit to the session layer.
This process is repeated through all layers in computer A until a final header is added at the data-link layer. After the data-link-layer header is added, the data unit is known as a "frame." The data, or frame, is passed from the data-link layer to the physical layer and is transmitted across the transmission medium connecting the two computers.
When the signal reaches computer B, layer one in computer B (the physical layer) copies the data. Now the process is reversed. The physical layer in computer B transfers the data to the data-link layer. The data-link layer removes the header information that was attached by the corresponding layer in computer A, acts upon the information the header contains, and transfers the data unit up to the network layer. This process continues, with the headers being stripped off at each layer and the instructions contained therein carried out, until the original data from computer A (the message) is finally passed from the application layer to the E-mail application in computer B. When the E-mail application receives the message, it displays the message on the screen for the user of computer B to read.
Now look at Figure 16 and imagine what would be possible if the software implementing different layers of the OSI model were able to handle not just one communication protocol at any one layer, but almost any communication protocol used at any layer, by any computer—there would be no limits to the interconnection of dissimilar computing devices. This is the kind of power that will be the basis for a global network—the networking of all kinds of business and personal devices into the Information Superhighway. And this is the kind of power built into NetWare® products.
Commonly Used Standards
When you read about NetWare products, you will find references to various standards and communication protocols supported by NetWare networks.
To understand the capabilities of NetWare products, it will help to know the OSI layer at which a particular protocol operates and why the standard is important. As you shall see later, by converting protocols or using multiple protocols at different layers of the OSI model, it is possible to enable different computer systems to share data, even if they use different software applications, operating systems, and data-encoding techniques.
Figure 17 shows some commonly used standards and the OSI layer at which they operate.
Figure 17: Important standards at various OSI layers
Layer-One Standards: Physical
Standards at the physical layer include protocols for transmitting a bitstream over media such as baseband coaxial cable, unshielded twisted-pair wiring, and optical fiber cable. The most commonly used are those specified in the Institute of Electrical and Electronic Engineers (IEEE) 802.3, 802.4, and 802.5 standards and the American National Standards Institute (ANSI) Fiber Distributed Data Interface (FDDI) standard. Figure 17 shows the transmission media included in each of these standards. An emerging standard for this layer is the Synchronous Optical Network (SONET).
Layer-Two Standards: Data Link (Media Access Control and Logical Link Control)
The most commonly used layer-two protocols are those specified in the IEEE's 802.2, 802.3, 802.4, and 802.5 standards and the ANSI FDDI standard (all shown in Figure 17). Many microcomputer networking products use one of these standards (or the virtually identical ISO version).
Important recent technologies at layer two include 100Base-TX (IEEE 802.3u), 100VG-AnyLAN (802.12), and Asynchronous Transfer Mode (ATM). The ATM standard is not yet fully defined. Also, frame relay is an important layer-two WAN technology. These technologies are treated in greater detail in a later section.
Layer-two standards encompass two sublayers: media access control and logical link control.
Media Access Control
The media access control (MAC) protocol specifies how workstations cooperatively share the transmission medium.
The IEEE 802.3 standard specifies a medium-access method known as "carrier sense multiple access with collision detection (CSMA/CD)." This medium-access method is the same as the contention method described in the earlier discussion of topologies, under the heading "Logical Bus."
The IEEE 802.4, 802.5, and FDDI standards all specify some form of token passing as the media access control method. The basics of the token-passing method were also described earlier, also under the heading "Logical Bus."
In general, using a form of token passing for the media access control works best when large numbers of computers frequently send small amounts of data—for example, when a number of workstations continually read and write small records to and from a database. Contention schemes work well when computers send large amounts of data intermittently—for example, during desktop publishing or document imaging.
Logical Link Control
The function of the logical link control layer is to ensure the reliability of the physical connection.
The IEEE 802.2 standard is the most commonly used logical link control standard.
The Point-to-Point Protocol (PPP) is an important standard at this OSI level. PPP is used for communications across point-to-point links such as T1 and T3 lines. It is an important protocol for wide area networking, which will be covered later.
Layer-Three Standards: Network
The function of the network layer is to manage communications, most importantly the routing and relaying of data, between workstations.
One important network-layer standard is the Department of Defense (DOD) Internet Protocol (IP) specification, which is part of the Transmission Control Protocol/Internet Protocol (TCP/IP) standard developed by the DOD. This protocol has become extremely important recently because it is the basis for the Internet and for all intranet technology. Also, the Department of Defense will often not purchase networking products that cannot communicate using this protocol.
Because Novell commands a large share of the networking market, its native Internetwork Packet Exchange™ (IPX) protocol, is also an important network-layer standard. IPX is a connectionless datagram protocol. A connectionless protocol does not need to establish a connection between two networked computers to transfer information between them. Packet acknowledgment, or connection control, is provided by protocols above IPX, such as Novell's Sequenced Packet Exchange™ (SPX). SPX will be explained in more detail in a later section. Because IPX is a datagram protocol, each communication packet is treated as an individual entity. IPX does not have to establish a logical or sequential relation between packets. Thus, IPX is very efficient—it addresses and transfers data with minimum control overhead.
IPX uses other NetWare protocols that work at the network layer to accomplish internetwork routing. These protocols, the Routing Information Protocol (RIP), the Service Advertising Protocol (SAP), and the NetWare Link Services Protocol™ (NLSP), will be explained in more detail in a later section.
The Consultative Committee for International Telegraph and Telephone (CCITT) X.25 standard is another commonly used network-layer standard. It specifies the interface for connecting computers on different networks by means of an intermediate connection made through a packet-switched network (for example, a common carrier network such as CompuServe, Tymnet, or Telnet). The X.25 standard includes the data-link and physical-layer protocols shown below it in Figure 17.
Apple Computer, Inc. has established a set of protocols for its products, referred to collectively as AppleTalk. At the network layer of the OSI model, the Apple protocol is called Datagram Delivery Protocol. Figure 18 shows how the set of AppleTalk protocols fits within the OSI model.
Figure 18: Where AppleTalk protocols fit in the OSI model
Like Novell's native protocols, Apple's standard protocols are important because of Apple's wide acceptance in the microcomputer market.
Layer-Four Standards: Transport
Standards at this OSI layer provide for the reliability of the end-to-end communication link. This layer isolates the upper three layers, which are all concerned with user and application requirements, from knowing the details required to manage the end-to-end connection.
The ISO has issued a transport-layer standard that is simply called the Transport Protocol (TP). Because it is an ISO standard, it is of worldwide importance.
At the transport layer, Novell's native protocol is SPX™. SPX provides guaranteed packet delivery and packet sequencing. Although it is basically a transport-layer protocol, it also includes session-layer functions. The NetWare Core Protocol (NCP) and SAP also provide transport-layer functions. SPX, NCP, and SAP will be treated in more detail in a later section.
The AppleTalk protocol set has a number of protocols that operate at the transport layer, including Routing Table Maintenance Protocol, AppleTalk Echo Protocol, AppleTalk Transaction Protocol, and the Name Binding Protocol.
IBM's NetBIOS protocol (not shown in Figure 17) is also an important protocol at this layer and at the session layer above.
The DOD's Transmission Control Protocol (TCP), which is part of the TCP/IP standard, is important at the transport layer to the same degree (extremely important) and for the same reasons as the IP standard at layer three. This protocol provides all functions required for this layer (transport) and part of the functions for the session layer above.
Layer-Five Standards: Session
The function of the session layer is to establish, manage, and terminate the connections of individual network users.
The ISO session standard, named simply "session," has the same worldwide importance as the ISO transport standard. The DOD's Transmission Control Protocol performs important functions at this layer.
In a NetWare environment, the NetWare Core Protocol™ (NCP) provides most of the necessary session-layer functions. SAP also provides functions at this layer.
Layer-Six and Layer-Seven Standards: Presentation and Application
The presentation layer performs general data transformations useful to a variety of applications, thus providing a useful common interface. Presentation-layer services include data encryption and text compression. The application layer provides user applications with basic (yet complete) services such as file transfer and network management functions.
Two important OSI protocols encompassing both the presentation and application layers are File Transfer, Access, and Management (FTAM) and Virtual Terminal Protocol (VTP). Each of these protocols is exactly what its name implies. FTAM provides user applications with useful file transfer and management functions. VTP supports applications by converting specific terminal characteristics to a general (virtual) terminal model shared by applications.
X.400 is an important CCITT standard that encompasses both the presentation and application layers. X.400 provides message handling and E-mail services. It is an important standard because it is the basis for a number of pervasive E-mail packages as well as for other widely used messaging products.
An important DOD standard at this level is File Transfer Protocol (FTP), which, again, is named for the service it provides.
The NetWare protocols that provide presentation- and application-layer functions are NCP™ and SAP. All NetWare protocols will be treated in more detail in a later section.
Further Perspective: Standards and Open Systems
You probably noticed from looking at Figures 17 and 18 that most accepted standards are not neatly packaged to include all (and only) those services specified for any OSI layer. In fact, most common standards encompass parts of multiple OSI layers. This includes most standards adopted by the various government agencies that develop them.
Product vendors' actual implementation of OSI layers is even less neatly divided. Vendors implement accepted standards, which already include mixed services from multiple layers, in different ways.
So why go to all the trouble to agree on a model and then define standards if you are not going to be exact when fitting the standards to the model or in implementing the standards when building a product?
Actually, standards development and implementation have proceeded more or less as expected. The OSI model was never intended to foster a rigid, unbreakable set of rules. It was expected that in implementing the OSI communication model, networking vendors would be free to use whichever standard for each layer they deemed most appropriate. They would also be free to implement each standard in the manner best suited for the purposes of their products.
As noted earlier, however, it is clearly in a vendor's best interest to manufacture products that conform to the intentions behind the OSI model. To do this, a vendor must provide the services required at each OSI model layer in a manner that will enable its system to be simply and easily connected to the systems of other vendors—in other words, vendors must develop open systems. The consequences of not doing so are severe and unavoidable.
Which leads to the next issue—how do you determine if a system is an open system? You can start by getting answers to simple questions such as: (1) Can you establish communications using virtually any accepted communication standard? and (2) How easily can you do this? For example, can you communicate with other networks that are using the TCP/IP protocol, even if your network uses some other protocol at that layer? If you can communicate, what kind of effort is required? And how reliable are such communications?
As you begin asking questions like these, you will find that Novell has the answers you need. NetWare products support every standard we have presented, as well as virtually every other accepted standard. The more you understand NetWare products, the more you will understand that no system is more open than a NetWare system.
Prior Index Next This page is maintained by: Michael P. Harris
Last Updated: Aug 29
Copyright © 1998