Error Control

Error control
is responsible for checking each transmission and verifying that all of the PDUs
are contiguous and not erroneous. If there are missing or damaged PDUs, the destination will
not send an ACK packet for the previous transmission. (Refer to Figure 36.6.)
Once all of the data is transferred without errors, the originating system sends a termination
request, which tells the destination system that no more data needs to be transmitted. The destination
system then responds with a termination acknowledgment.
As you can see, both systems do a lot of communicating, aside from the exchange of data.
From the connection request to the termination acknowledgment, every exchange is accompanied
with control information that keeps the data transfer reliable and error free. Table 36.1
gives examples of several connection-oriented protocols. 1102

Flow Control

Although flow control was briefly described earlier, this section contains more detail.
Flow control
is responsible for ensuring that the transmitting station does not send data faster than the
receiving station can process it. This is done by establishing a window size for the transmission.
Look at Figure 36.6 to see how windowing works. Notice that the originating system sends
out a specified number of PDUs. Once that number is reached, the originating system waits for
a response from the destination system. After the response is received, the system continues to
transmit data.

Sequenced Data Transfer

Systems send protocol data units (PDUs) to one another, and each level of the OSI model has
its own type of PDU. Figure 36.4 shows the PDU names for all seven OSI layers. For example,
the Application layer’s PDU name is layer 7 PDU. Although this convention can be used for all
layers, some layers use other names as well. For instance, a layer 3 PDU is called a
packet
and
a layer 2 PDU is called a
frame
. When a system sends data to another system, the data has to
be fragmented so that it fits the MTU (maximum transmission unit). Therefore, several frames
may be needed to transfer the original data. Connection-oriented protocols assign a sequence
number to each outgoing and incoming PDU. This is
sequenced data transfer
.
Figure 36.5 shows you how sequencing works. There is a possibility that the destination system
will receive the PDUs out of order. If this happens, the protocol on the destination system
uses the sequence numbers to put the PDUs back into the correct order so that the original data
is obtained. 1101

Connection-Oriented Protocols

Connection-oriented protocols
contain inherent functions that control the connection as well as
data transfer. These functions are very detailed in the procedures that are followed to enable
reliable and error-free data transfer. When a source open system needs to transfer data to a destination
open system, the connection-oriented protocols actually establish a communication
pipe. The
pipe,
as it is called here, is nothing more than a logical connection between two open
systems. A great deal of information is used to establish this communication pipe, however.
In order to establish a connection, the two open systems must share certain information that
allows them to negotiate terms and finally establish a link. The information includes the common
protocol that will be used, required resources, and available resources. Look at Figure 36.3. This
figure shows the steps taken as communication is established between two open systems when
using TCP, a connection-oriented protocol.

The originating system first sends a connection request to the destination system. This
request contains information that the two systems need to agree upon before the connection can
be established. Some of the information includes the common protocol, protocol parameters,
and required resources.
Protocol parameters
are the window sizes and other possible parameters.
The
window size
is the amount of data that a station can transmit before needing an
acknowledgment from the destination system that all the data was received without error, or
that errors existed and part of the data will need to be retransmitted.
Required resources
can
include necessary bandwidth, specific port numbers, and other network resources.
The destination system receives this connection request; if it can accommodate the common
protocol, protocol attributes, and required resources, it replies with a connection accept. If, for
some reason, the destination system cannot accommodate any of the requirements sent by the
originating system, the destination system responds with a connection deny. A denied connection
can result from a blocked port on the destination system, insufficient bandwidth between
the systems, or other unavailable requested resources.
Assuming that a connection is established between the two systems, data and control information
is exchanged during the life of the connection. This data exchange can be considered a
dialog
. First, the originating system sends data until the window size is reached. That system
then waits for a response from the destination system. The destination system sends control
information that informs the originating system what needs to happen next. The transmission
can be an acknowledgment that all data in the transmission was received without error and that
the originating system can send the next batch of data. In addition, the destination system can
send a message informing the originating system that some of the data was missing, corrupted,
or had other errors that require the data to be retransmitted.
The foregoing procedure can be summarized with the description of three processes. You will
learn more about each of these processes in the following sections:
Sequenced data transfer
Each packet of a session is assigned a sequence number.
Flow control
Acknowledgments are required after a specified amount of data has been sent.
Error control
Verification of contiguous and nonerroneous packets. 1099

Global Protocol Classifications

As mentioned, each layer of the OSI model utilizes specific protocols that enable the layer to perform
the necessary functions and communicate with adjacent layers. Each protocol has specific
properties based on the functions that it needs to accomplish. Throughout all seven layers, there
are two major protocol classifications: connection-oriented and connectionless.

The OSI Reference Model

This section is a review of the OSI model, which was originally discussed in
CCNA: Cisco Certified
Network Associate Study Guide, 4th ed
., by Todd Lammle (Sybex, 2004). The
OSI model
(the Open Systems Interconnection reference model) is the template used to design applications
or protocols that allow nonhomogenous computers or networks to communicate with one
another. The ISO (International Organization for Standardization) developed the OSI model.
The OSI model consists of seven layers. Each layer communicates directly with its adjacent layers,
as well as with the corresponding layer of the destination system (depicted in Figure 36.1).
Communication between layers facilitates the transfer of data up and down the OSI model. Communication
between the corresponding layers of the source system and the destination system
enables two heterogeneous networks or computers to understand each other.
The OSI template defines the services and roles that each layer is to provide. Because each
layer provides different services and functions, the layers need to communicate so that the data
can be transmitted up and down the seven layers and onto the destination system. The following
list summarizes the responsibility of each of the seven layers, starting from the Physical layer
and working up to the Application layer:
Physical
This layer sends and receives bits with values of 1s and 0s. The Physical layer is
in charge of determining how it sends these values. If the physical connection between two
machines is fiber-optic, then the Physical layer has to use light to transmit the 1s and 0s. If
the connection is electrical, then electrical signals are sent to represent the 1s and 0s.

Data Link
This layer takes all the data that is accumulated as packets are handed from one layer
to the next and then packages it into frames. The Data Link layer equates the Network layer address
(IP address) to a data link address, or MAC address, of the next hop. Once the physical address is
known, the frame is sent to that address. The receiving interface uses the Data Link layer to extract
the packet from the frame, discards the frame, and then sends the packet up to the Network layer.
Network
This layer defines the topology of the network through the use of logical addressing.
Routing protocols use this information to route packets.
Transport
This layer takes care of end-to-end communications. It is responsible for the
connection to the destination system, as well as for packet segmentation and assembly. The
Transport layer includes both connection-oriented and connectionless protocols (for example,
TCP and UDP).
Session
This layer is responsible for coordinating communication among applications, which
it does through dialog-control methods.
Presentation
This layer negotiates syntax, so it is responsible for the proper method of presenting
the data to the Application layer. Some of the Presentation layer functions are compression/
decompression and encryption/decryption of data.

Application
This is the user and application interface. The Application layer is responsible for
data exchange and job management. It also handles file, print, message, database, and application
services.
You saw how the logical data flow of the OSI model works, but look at Figure 36.2, in
which you can see the actual data flow. This figure depicts data that is handed from the Application
layer all the way down to the Physical layer. At that point, the data is transmitted
across any variety of physical media to the next hop, or destination system. Once the 1s and
0s arrive at the Physical layer of the destination system, the information is sent to layer 2 (the
Data Link layer). This layer discards the frame, and then the extracted packet is handed up
to the Network layer. The network packet header is stripped off, and the resulting packet is
handed up to the Transport layer. This process is repeated for each layer until it arrives at the
Application layer.
Now that each layer of the OSI reference model has been explained briefly, you need to
focus on the functions of each layer in detail. This detail provides the necessary background
and information to effectively troubleshoot network problems that occur within specific layers
of the OSI model.

Protocol Attributes

THE CCNP EXAM TOPICS COVERED IN THIS
CHAPTER INCLUDE THE FOLLOWING:

Verify network connectivity.

Use the optimal troubleshooting approach in resolving
network problems.

Minimize downtime during troubleshooting.

Use Cisco IOS commands to identify problems.

Determine the layer or layers on which a problem is occurring.

As you know, to successfully troubleshoot network problems, it is
important to have a good understanding of how network components,
including PCs and servers, communicate with each other.
Without this basic knowledge, troubleshooting a network problem is like trying to read a book in
a foreign language. The information is there, but it just isn’t comprehensible. Although the troubleshooting
model discussed in Chapter 33, “Troubleshooting Methodology,” provides the method of
retrieving all the necessary information, the data is useless without an understanding of the information
presented.
This chapter is a review of the protocols used by layers 2, 3, and 4 of the OSI model. We briefly
review the seven layers of the OSI model, and then discuss how they communicate with one another.
We then discuss layer 2 and layer 3 protocols. More specific information on some of the material
covered here can be found in later chapters and is cross-referenced here where appropriate.
1095