General Networking Concepts:-Section 2

  OSI Reference Model

 

Network Reference Models

A computer network connects two or more devices together to share information and services. Multiple networks connected together form an internetwork.

Internetworking present challenges - interoperating between products from different manufacturers requires consistent standards. Network reference models were developed to address these challenges. A network reference model serves as a blueprint, detailing how communication between network devices should occur.

The two most recognized network reference models are:

·        The Open Systems Interconnection (OSI) model

·        The Department of Defense (DoD) model

Without the framework that network models provide, all network hardware and software would have been proprietary. Organizations would have been locked into a single vendor’s equipment, and global networks like the Internet would have been impractical, if not impossible.

Network models are organized into layers, with each layer representing a specific networking function. These functions are controlled by protocols, which are rules that govern end-to-end communication between devices.

Protocols on one layer will interact with protocols on the layer above and below it, forming a protocol suite or stack. The TCP/IP suite is the most prevalent protocol suite, and is the foundation of the Internet.

A network model is not a physical entity – there is no OSI device. Manufacturers do not always strictly adhere to a reference model’s blueprint, and thus not every protocol fits perfectly within a single layer. Some protocols can function across multiple layers.

OSI Reference Model

The Open Systems Interconnection (OSI) model was developed by the International Organization for Standardization (ISO), and formalized in 1984. It provided the first framework governing how information should be sent across a network.

The OSI model consists of seven layers, each corresponding to a specific network function:

1. Application 2. Presentation 3. Session 4. Transport 5. Network 6. Data-link 7. Physical

Note that the bottom layer is Layer 1. Various mnemonics make it easier to remember the order of the OSI model’s layers:

Application All Away Presentation People Pizza Session Seem Sausage Transport To Throw Network Need Not Data-link Data Do Physical Processing Please

ISO further developed an entire protocol suite based on the OSI model; however, the OSI protocol suite was never widely implemented.

The OSI model itself is now somewhat deprecated – modern protocol suites, such as the TCP/IP suite, are difficult to fit cleanly within the OSI model’s seven layers. This is especially true of the upper three layers.

The bottom (or lower) four layers are more clearly defined, and terminology from those layers is still prevalently used. Many protocols and devices are described by which lower layer they operate at.

OSI Model - The Upper Layers

The top three layers of the OSI model are often referred to as the upper layers:

·        Layer-7 - Application layer

·        Layer-6 - Presentation layer

·        Layer-5 - Session layer

Protocols that operate at these layers manage application-level functions, and are generally implemented in software.

The function of the upper layers of the OSI model can be difficult to visualize. Upper layer protocols do not always fit perfectly within a layer, and often function across multiple layers.

OSI Model - The Application Layer

The Application layer (Layer-7) provides the interface between the user application and the network. A web browser and an email client are examples of user applications.

The user application itself does not reside at the Application layer - the protocol does. The user interacts with the application, which in turn interacts with the application protocol.

Examples of Application layer protocols include:

·        FTP, via an FTP client

·        HTTP, via a web browser

·        POP3 and SMTP, via an email client

·        Telnet

The Application layer provides a variety of functions:

·        Identifies communication partners

·        Determines resource availability

·        Synchronizes communication

The Application layer interacts with the Presentation layer below it. As it is the top-most layer, it does not interact with any layers above it.

OSI Model - The Presentation Layer

The Presentation layer (Layer-6) controls the formatting and syntax of user data for the application layer. This ensures that data from the sending application can be understood by the receiving application.

Standards have been developed for the formatting of data types, such as text, images, audio, and video. Examples of Presentation layer formats include:

·         Text - RTF, ASCII, EBCDIC

·         Images - GIF, JPG, TIF

·         Audio - MIDI, MP3, WAV

·         Movies - MPEG, AVI, MOV

If two devices do not support the same format or syntax, the Presentation layer can provide conversion or translation services to facilitate communication.

Additionally, the Presentation layer can perform encryption and compression of data, as required. However, these functions can also be performed at lower layers as well. For example, the Network layer can perform encryption, using IPSec.

OSI Model - The Session Layer

The Session layer (Layer-5) is responsible for establishing, maintaining, and ultimately terminating sessions between devices. If a session is broken, this layer can attempt to recover the session.

Sessions communication falls under one of three categories:

·        Full-Duplex – simultaneous two-way communication

·        Half-Duplex – two-way communication, but not simultaneous

·        Simplex – one-way communication

Many modern protocol suites, such as TCP/IP, do not implement Session layer protocols. Connection management is often controlled by lower layers, such as the Transport layer.

The lack of true Session layer protocols can present challenges for high- availability and failover. Reliance on lower-layer protocols for session management offers less flexibility than a strict adherence to the OSI model.

OSI Model - The Lower Layers

The bottom four layers of the OSI model are often referred to as the lower layers:

·        Layer-4 – Transport layer

·        Layer-3 – Network layer

·        Layer-2 – Data-Link layer

·        Layer-1 – Physical layer

Protocols that operate at these layers control the end-to-end transport of data between devices, and are implemented in both software and hardware.

OSI Model - The Transport Layer

The Transport layer (Layer-4) does not actually send data, despite its name. Instead, this layer is responsible for the reliable transfer of data, by ensuring that data arrives at its destination error-free and in order.

Transport layer communication falls under two categories:

·        Connection-oriented – requires that a connection with specific agreed-upon parameters be established before data is sent.

·        Connectionless – requires no connection before data is sent.

Connection-oriented protocols provide several important services:

·        Segmentation and sequencing – data is segmented into smaller pieces for transport. Each segment is assigned a sequence number, so that the receiving device can reassemble the data on arrival.

·        Connection establishment – connections are established, maintained, and ultimately terminated between devices.

·        Acknowledgments – receipt of data is confirmed through the use of acknowledgments. Otherwise, data is retransmitted, guaranteeing delivery.

·        Flow control (or windowing) – data transfer rate is negotiated to prevent congestion.

The TCP/IP protocol suite incorporates two Transport layer protocols:

·        Transmission Control Protocol (TCP) – connection-oriented

·        User Datagram Protocol (UDP) - connectionless

OSI Model - The Network Layer

The Network layer (Layer-3) controls internetwork communication, and has two key responsibilities:

·        Logical addressing – provides a unique address that identifies both the host, and the network that host exists on.

·        Routing – determines the best path to a particular destination network, and then routes data accordingly.

Two of the most common Network layer protocols are:

·        Internet Protocol (IP)

·        Novell’s Internetwork Packet Exchange (IPX).

IPX is almost entirely deprecated. IP version 4 (IPv4) and IP version 6 (IPv6) are covered in nauseating detail in other guides.

OSI Model - The Data-Link Layer

While the Network layer is concerned with transporting data between networks, the Data-Link layer (Layer-2) is responsible for transporting data within a network.

The Data-Link layer consists of two sublayers:

·        Logical Link Control (LLC) sublayer

·        Media Access Control (MAC) sublayer

The LLC sublayer serves as the intermediary between the physical link and all higher layer protocols. It ensures that protocols like IP can function regardless of what type of physical technology is being used.

Additionally, the LLC sublayer can perform flow-control and error- checking, though such functions are often provided by Transport layer protocols, such as TCP.

The MAC sublayer controls access to the physical medium, serving as mediator if multiple devices are competing for the same physical link. Data- link layer technologies have various methods of accomplishing this - Ethernet uses Carrier Sense Multiple Access with Collision Detection (CSMA/CD), and Token Ring utilizes a token.

The Data-link layer packages the higher-layer data into frames, so that the data can be put onto the physical wire. This packaging process is referred to as framing or encapsulation.

The encapsulation type will vary depending on the underlying technology. Common Data-link layer technologies include following:

·        Ethernet – the most common LAN data-link technology

·        Token Ring – almost entirely deprecated

·        FDDI (Fiber Distributed Data Interface)

·        802.11 Wireless

·        Frame-Relay

·        ATM (Asynchronous Transfer Mode)

The data-link frame contains the source and destination hardware (or physical) address. Hardware addresses uniquely identify a host within a network, and are often hard coded onto physical network interfaces.

However, hardware addresses contain no mechanism for differentiating one

network from another, and can only identify a host within a network. The most common hardware address is the Ethernet MAC address.

OSI Model - The Physical Layer

The Physical layer (Layer-1) controls the signaling and transferring of raw bits onto the physical medium. The Physical layer is closely related to the Data-link layer, as many technologies (such as Ethernet) contain both data- link and physical functions.

The Physical layer provides specifications for a variety of hardware:

·        Cabling

·        Connectors and transceivers

·        Network interface cards (NICs)

·        Wireless radios

·        Hubs

Encapsulation and Layered Communication

As data is passed from the user application down the virtual layers of the OSI model, each layer adds a header (and sometimes a trailer) containing protocol information specific to that layer. These headers are called Protocol Data Units (PDUs), and the process of adding these headers is called encapsulation. Note that in the TCP/IP protocol suite only the lower layers perform encapsulation, generally.

For example, a Transport layer protocol such as TCP will add a header containing flow control, port numbers, and sequencing. The Network layer header contains logical addressing information, and the Data-link header contains physical addressing and other hardware specific information.

The PDU of each layer is identified with a different term:

Layer	PDU Name
Application	-
Presentation	-
Session	-
Transport	Segments
Network	Packets
Data-Link	Frames
Physical	Bits

Each layer communicates with the corresponding layer on the receiving device. For example, on the sending device, source and destination hardware addressing is placed in a Data-link header. On the receiving device, that Data-link header is processed and stripped away (decapsulated) before being sent up to the Network and other upper layers.

Network devices are commonly identified by the OSI layer they operate at; or, more specifically, what header or PDU the device processes.

For example, switches are generally identified as Layer-2 devices, as switches process information stored in the Data-Link header of a frame, such as Ethernet MAC addresses. Similarly, routers are identified as Layer- 3 devices, as routers process logical addressing information in the Network header of a packet, such as IP addresses.

OSI Reference Model Example

A web browser serves as a good practical illustration of the OSI model and the TCP/IP protocol suite:

·        The web browser serves as the user interface for accessing a website. The browser itself does not function at the Application layer. Instead, the web browser invokes the Hyper Text Transfer Protocol (HTTP) to interface with the remote web server, which is why http:// precedes every web address.

·        The Internet can provide data in a wide variety of formats, a function of the Presentation layer. Common formats on the Internet include HTML, XML, PHP, GIF, and JPEG. Any encryption or compression mechanisms used on a website are also considered a Presentation layer function.

·        The Session layer is responsible for establishing, maintaining, and terminating the session between devices, and determining whether the communication is half-duplex or full-duplex. However, the TCP/IP stack generally does not include session-layer protocols, and is reliant on lower-layer protocols to perform these functions.

·        HTTP utilizes the TCP Transport layer protocol to ensure the reliable delivery of data. TCP establishes and maintains a connection from the client to the web server, and packages the higher-layer data into segments. A sequence number is assigned to each segment so that data can be reassembled upon arrival.

·        The best path to route the data between the client and the web server is determined by IP, a Network layer protocol. IP is also responsible for the assigned logical addresses on the client and server, and for encapsulating segments into packets.

·        Data cannot be sent directly to a logical address. As packets travel from network to network, IP addresses are translated to hardware addresses, which are a function of the Data-Link layer. The packets are encapsulated into frames to be placed onto the physical medium.

·        The data is finally transferred onto the network medium at the Physical layer, in the form of raw bits. Signaling and encoding mechanisms are defined at this layer, as is the hardware that forms the physical connection between the client and the web server.

IP and the DoD Model

The Internet Protocol (IP) was originally developed by the Department of Defense (DoD), and was a cornerstone for a group of protocols that became known as the TCP/IP protocol suite. 

OSI   Model                 DoD Model

Application Presentation Session Transport Network Data-link Physical

Application Host-to-Host Internet Network Access

   

The DOD developed their own networking model, which became known as the DoD or TCP/IP Model. It consists of four layers:

                      

         

The consolidated DoD model is generally regarded as more practical than the OSI model. Upper layer protocols often provide services that span the top three layers. A converged Data-link and Physical layer is also sensible, as many technologies provide specifications for both layers, such as Ethernet.

The following chart illustrates where common protocols fit into the DoD model:

Layer                           Example Protocols

Application	FTP, HTTP, SMTP
Host-to-Host	TCP, UDP
Internet	IP
Network Access	Ethernet

Despite the practicality of the DoD model, the OSI model is still the basis for most network terminology.

 

General Networking Concepts:-Section 1

                                                

Introduction to Networks

What is a Network?

A network is simply defined as something that connects things together for a specific purpose. The term network is used in a variety of contexts, including telephone, television, computer, or even people networks.

A computer network connects two or more devices together to share a nearly limitless range of information and services, including:

·        Documents

·        Email and messaging

·        Websites

·        Databases

·        Music

·        Printers and faxes

·        Telephony and videoconferencing

Protocols are rules that govern how devices communicate and share information across a network. Examples of protocols include:

·        IP – Internet Protocol

·        HTTP  - Hyper Text Transfer Protocol

·        SMTP – Simple Mail Transfer Protocol

Multiple protocols often work together to facilitate end-to-end network communication, forming protocol suites or stacks. Protocols are covered in great detail in other guides.

Network reference models were developed to allow products from different manufacturers to interoperate on a network. A network reference model serves as a blueprint, detailing standards for how protocol communication should occur.

The Open Systems Interconnect (OSI) and Department of Defense (DoD) models are the most widely recognized reference models. Both are covered in great detail in another guide.

Basic Network Types

Network types are often defined by function or size. The two most common categories of networks are:

·        LANs (Local Area Networks)

·        WANs (Wide Area Networks)

A LAN is generally a high-speed network that covers a small geographic area, usually contained within a single building or campus. A LAN is usually under the administrative control of a single organization. Ethernet is the most common LAN technology.

A WAN can be defined one of two ways. The book definition of a WAN is a network that spans large geographical locations, usually to connect multiple LANs. This is a general definition, and not always accurate.

A more practical definition of a WAN is a network that traverses a public or commercial carrier, using one of several WAN technologies. A WAN is often under the administrative control of several organizations (or providers), and does not necessarily need to span large geographical distances.

A MAN (Metropolitan Area Network) is another category of network, though the term is not prevalently used. A MAN is defined as a network that connects LAN’s across a city-wide geographic area.

An internetwork is a general term describing multiple networks connected together. The Internet is the largest and most well-known internetwork.

Some networks are categorized by their function, as opposed to their size. A SAN (Storage Area Network) provides systems with high-speed, lossless access to high-capacity storage devices.

A VPN (Virtual Private Network) allows for information to be securely sent across a public or unsecure network, such as the Internet. Common uses of a VPN are to connect branch offices or remote users to a main office.

Network Architectures

A host refers to any device that is connected to a network. A host can also be defined as any device assigned a network address.

A host can serve one or more functions:

·        A host can request data, often referred to as a client.

·        A host can provide data, often referred to as a server.

·        A host can both request and provide data, often referred to as a peer.

Because of these varying functions, multiple network architectures have been developed, including:

·        Peer-to-Peer

·        Client/Server

·        Mainframe/Terminal

In a basic peer-to-peer architecture, all hosts on the network can both request and provide data and services. For example, two Windows XP workstations configured to share files would be considered a peer-to-peer network.

Peer-to-peer networks are very simple to configure, yet this architecture presents several challenges. Data is difficult to manage and back-up, as it is spread across multiple devices. Security is equally problematic, as user accounts and permissions much be configured individually on each host.

In a client/server architecture, hosts are assigned specific roles. Clients request data and services stored on servers. An example of a client/server network would be Windows XP workstations accessing files off of a Windows 2003 server.

There are several advantages to the client/server architecture. Data and services are now centrally located on one or more servers, consolidating the management and security of that data. As a result, client/server networks can scale far larger than peer-to-peer networks.

One key disadvantage of the client/server architecture is that the server can present a single point of failure. This can be mitigated by adding redundancy at the server layer.

Network Architectures (continued)

In a mainframe/terminal architecture, a single device (the mainframe) stores all data and services for the network. This provides the same advantages as a client/server architecture – centralized management and security of data.

Additionally, the mainframe performs all processing functions for the dumb terminals that connect to the mainframe. The dumb terminals perform no processing whatsoever, but serve only as input and output devices into the mainframe.

In simpler terms, the mainframe handles all thinking for the dumb terminals. A dumb terminal typically consists of only a keyboard/mouse, a display, and an interface card into the network.

The traditional mainframe architecture is less prevalent now than in the early history of networking. However, the similar thin-client architecture has gained rapid popularity. A thin-client can be implemented as either a hardware device, or software running on top of another operating system (such as Windows or Linux).

Like dumb terminals, thin-clients require a centralized system to perform all (or most) processing functions. User sessions are spawned and managed completely within the server system.

Hardware thin-clients are generally inexpensive, with a small footprint and low power consumption. For environments with a large number of client devices, the thin-client architecture provides high scalability, with a lower total cost of ownership.

The two most common thin-client protocols are:

·        RDP (Remote Desktop Protocol) – developed by Microsoft

·        ICA (Independent Computer Architecture) – developed by Citrix