OSI Model
2:45 AM | Author: Siddhartha Ghosh

Application Layer: Developer creates application that can plug into network socket – means they can use the services of networking components defined by the Application layer. Main job defined in Application layer is identifying partners, look for resources & synchronization. For eg. Email client. There are various type of Applications that may require different kind of networking services. For. eg. File & Print, Database, Email, Name resolution, Internet access, Terminal emulation, and Management services (like SNMP)

Presentation Layer: Data Encoding/Decoding & Conversion is it’s primary job. For e.g ASCII to EBCDIC, IBM world to LAN PC vise versa. It’s also responsible for Encryption / Decryption for eg. SSL/TLS (Secure Socket Layer/Transport Layer Services used by web based application for security), Certificates, Compression & Decompression.

Session Layer: It establishes a session between client and server or two hosts, so that they can communicate over the protocol stack. It’s primary job is to Setup Dialog (establish connection), Dialog control (manages connection), Termination of connection.
Note:- Setting up Dialog means identifying a type of connection, like – Full Duplex, Half Duplex, Simplex. A Dialog is written within the application itself.

Transport Layer: It’s primary job is Data transport Reliability; Establishing, Managing & Terminating Virtual Circuit; Fault Detection & Recovery; Flow Control
Now it’s the time to talk about the protocol TCP (Transmission Control Protocol), which has its major impact in the world of networking.

There are two reliable protocols that comes under Transport Layer, they are:-



    TCP (Transmission Control Protocol) – It’s a connection oriented protocol, that means it establishes virtual circuit to make sure data is verified and acknowledged, so that data is guaranteed reliable. As every segment is acknowledged, Overhead is more in TCP. Therefore, to minimize this overhead, TCP has come out with couple Flow Control mechanism. One is Buffering (with Source Quench message) and other is Sliding Window. FTP, HTTP, SMTP, DNS uses the services of TCP.

      Buffering: A buffer is predefined in both source and destination host. Once the buffer is full, it starts discarding packets and send source quench message to source host to slow down. Transmission is stopped until the discarded packets are transmitted again and receiving host is unable to process even the available packets. Therefore, this mechanism doesn’t seems to be efficient and thus we basically use Sliding Window concept.


      Sliding Window: In this type of mechanism, every host has a predefined send & receive Window. It’s usually 8 segments of size but may vary from hosts to hosts, and that’s why 3-Way Handshake is necessary.

    Sliding Window idea is to keep data streaming with less overhead





    UDP (User Datagram Protocol): It is a connection less protocol, that means it doesn’t look out for acknowledgement after sending the segments, thus no Overhead. It follows a best effort delivery for reliable data transfer. DNS and TFTP uses the services of UDP.


Ports are also implemented at Transport Layer to identify application. Every application has a port defined. IANA (Internet Assigned Numbers Authority) has assigned port numbers in three categories.

Public Application (Well Known) – 1 -255
HTTP : 80
FTP : 21
FTP DATA : 20
TELNET : 23
SMTP : 25
DNS (Zone
Transfer/TCP) : 53
DNS (Name
Lookup/UDP) : 53
TFTP : 69
SNMP : 161
SNMP Trap : 162
Private Application - 256-1023
Eg. Microsoft Applications
Dynamic Client (Higher Order) – 1024-65535
For e.g Target port of FTP client back from server


Network Layer: Addressing, Actual connectivity and Path Detection is the main job of Network Layer. It’s the layer where routing starts.

    Routed protocols: (IP, IPX, AppleTalk) – Determines the best path based on routing table. It performs routing function and carries data through network.

    Routing protocols: (RIP, IGRP, EIGRP, OSPF, IS-IS, BGP4) – Determines distance, cost and populates routing table. It does the route selection. These protocols run exclusively on routers.

Determining the Switching method (Circuit {Telephone}/ Message {Store & Forward} / Packet {Frame Relay & T1}) is also one of the tasks of Network Layer.
This Layer has 2 supporting protocols:- ICMP, ARP

    ICMP: The Internet Control Message Protocol is one of the core protocols of the Internet Protocol Suite. It is chiefly used by networked computers' operating systems to send error messages—indicating, for instance, that a requested service is not available or that a host or router could not be reached. For e.g. Ping. It send source quench message for buffering, thus do the flow control.

    ARP: Address Resolution Protocol is the method for finding a host's link layer (hardware) address when only its Internet Layer (IP) or some other Network Layer address is known.


Data link Layer – Its main job is Framing (or Encapsulation) and physical addressing (48 bit MAC address). It’s the layer where the actual encapsulation takes place. It has two sub layers – LLC and MAC. Earlier there was no layers of abstraction between Layer 3 protocols and Layer 2 MAC addresses, because of which we could only use one NIC per machine and could bind only one protocol per NIC. To eliminate this drawback, LLC & MAC sublayers were introduced to provide layers of abstraction between upper layer protocols and underlying interface hardware. This layer also provides access to the topology.





It has support protocol RARP: Reverse Address Resolution Protocol is a Link layer networking protocol used by a host computer to obtain its IPv4 address given only its link-layer address (such as an Ethernet address).

Other standards and protocols of Data-link Layer are as below:

    HDLC: High-Level Data Link Control is a bit oriented synchronous protocol developed by ISO. It provided both connection-less and connection-oriented services. It is used for point to multipoint connection. A WAN technology. A serial interface on every CISCO router.

    SDLC: Synchronous Data Link Control is created by IBM to facilitate interconnectivity between IBM mainframes hosts. It’s exclusively used in SNA networks.

    X.25 / LAPB: It’s a standard for packet switched networks. Good in encapsulation. Provides built-in error correction. Understand circuit and packet switching technology as it is developed by telephone company. It’s a WAN technology and is predecessor to Frame Relay. It defines the DTE (Data Terminal Equipment) /DCE device (Data Communication Equipment). X.25 uses LAPB (Link Access Procedure Balanced) protocol, which provides maintenance component for X.25 like sequencing and retransmit.

    Frame Relay: It’s successor of X.25 and is much faster (upto 45 Mbps). It’s used for high quality digital technology and has acknowledgment function, hence guaranteed reliable.

    SLIP: Serial Line Internet Protocol is an original standard for TCP/IP transmission over low speed serial links upto 64 Kbps. It was used by Unix systems for Shell connectivity over dial-up.

    PPP: Point to Point Protocol is a new standard and is a successor of SLIP. It is used by Dialup as well as T1 connections. It supports both synchronous and asynchronous transmission. It adds multiple protocol support, encryption & error correction.

    ISDN: Integrated Services Digital Network connected telephone systems to transfer digital voice & data over these modified analog lines. It’s defined by 802.9

Physical Layer: It’s responsible for binary transmission, clocking (i.e., setting the rate at which data samples are passed across the wire), and determining physical topology (i.e., identifying how devices connect to the network physically)



There are five topologies:-

    Bus: It’s a LAN topology, also called Ethernet with CSMA/CD. It’s both half and full duplex capable. The popular standards for Ethernet are:-

    Ring: One directional data transmission, hence no collision. It’s once again a LAN topology. It uses UTP, STP or Fiber cables. Examples are:- IBM Token Ring and FDDI.

      IBM Token ring: Available in two speeds, 4Mbps & 16Mbps. Token is passed around a logical ring of network devices. System must acquire token to transmit data. Data travel in single direction, hence no collision takes place. Each host is connected with a relay connector called MAU (Multistation Access Unit), which then knows both it’s upstream and downstream. If a problem appears, the upstream sends a beacon towards opposite direction that determines the failure and that failed host is eliminated from the ring. MAU then updates itself with new upstream and downstream again.


    Source Wikipedia under creative commons license

      FDDI: Fiber Distributed Data Interface uses optical fiber as primary transmission medium. The FDDI LAN can range upto 200Km and offers 100Mbps speed. It contains two token rings, one for possible backup in case the primary ring fails. It also falls under Fast Ethernet category.




    Star: It’s primarily a LAN topology but is also a WAN topology.
    E.g. ARCnet –LAN
    Frame Relay – Logical WAN (Main office & Branch Office)

    Point-to-Point: It’s a WAN topology, used by switched network such as X.25 & Frame Relay or Leases line connection such as T1

    Full mesh / Partial Mesh: It’s a WAN topology used by Packet Switched networks.


Standards of Physical Layer:


    EIA/TIA-232 (= RS-232) – Supports connection speed upto 64 Kbps.
    EIA/TIA-449 (= RS-449) – Supports connection speed upto 2Mbps
    V.24/V.35 – ITU-T (International Telecommunication Unit) standard, very similar to RS-232

All of the above standards define serial communication and are all the standards of WAN topology.
History of Networking
2:24 AM | Author: Siddhartha Ghosh
History of Networking:


Early in 1960’s, there were standalone mainframe computers with unintelligent terminals attached. It was an age of centralized processing, where terminals could only avail the services that its mainframe was providing. Every mainframe vendors used to provide different kind of services and there were no standards. It led to Island of Automation, where one mainframe (of say Vendor 1) couldn’t exchange services with other mainframe (of Vendor 2). Department of Defense being a major customer thought of having a kind of standard and thus commenced ARPA (Advanced Research Project Agency) network called ARPAnet. The first protocol developed in 1969, called NCP (Network Connection Protocol), which was however capable of doing only simple file exchange. Therefore, TCP/IP research began at Stanford in 1969 funded by Department of Defense. It was published in a public domain RFC that led to its popularity. At this point the concept of Internetworking was born.

In 1970, ISO (International Organization of Standardization) started it’s work on OSI (Open System Interconnection) 7 layer model, which came into picture not before 1984. Hence there was only D2D 4 layer model till that time.
Also, in 1970, DIX (Digital, Intel and Xerox) together developed a standard called Ethernet II, which was used by ARPAnet, hence also called ARPA Ethernet.

1980s, was the time when transition from Centralized processing to Distributed processing took place. Terminals became intelligent with there own processing components and traditional hosts then evolved as file servers. For eg. Novell Netware, which was very popular at that time. It was the time when actual LAN and WAN definition came into picture.

Also, in 1980s, IEEE took 802 project.
802.1 was the original standard whose responsibility was High-level internetworking protocols and management.
802.2 is the standard for LLC (Logical Link Control), which was created to do Upper layer 7-3 abstraction from Layer 2 Data link. This abstraction was necessary otherwise we wouldn’t be able to use multiple NIC (Network Interface Card) in a machine and bind multiple protocols to a single NIC.
802.3 is the standard for Ethernet
802.4 is the standard for Token bus (Burroughs, Honeywell etc.)
802.5 is the standard for Token ring (IBM)
802.6 is the standard for MAN (Metropolitan Area Networks)
802.7 is the standard for Broadband LAN using Coaxial Cable
802.8 is the standard for Fiber Optic.
802.9 is the standard for Integrated Services LAN (ISDN)
802.10 is the standard for Interoperable LAN Security
802.11 is the standard for Wi-Fi

Today there are different kind of Networks, namely:-

PAN - Personal Area Network is a computer network used for communication among computer devices close to one person

LAN - Local Area Network is a computer network covering a small physical area, like a home, office, or small group of buildings, such as a school, or an airport.

CAN - Campus Area Network is a computer network made up of an interconnection of local area networks (LANs) within a limited geographical area. It can be considered one form of a metropolitan area network, specific to an academic setting.

MAN - Metropolitan Area Network is a network that connects two or more local area networks or campus area networks together but does not extend beyond the boundaries of the immediate town/city.

WAN - Wide Area Network is a computer network that covers a broad area (i.e. any network whose communications links cross metropolitan, regional, or national boundaries.

GAN - Global Area Networks specification is in development by several groups, and there is no common definition. In general, however, a GAN is a model for supporting mobile communications across an arbitrary number of wireless LANs, satellite coverage areas, etc.

SAN - Storage Are Network is a network designed to attach computer storage devices. It is an architecture to attach remote computer storage devices (such as disk arrays, tape libraries, and optical jukeboxes) to servers in such a way that the devices appear as locally attached to the operating system.

CDN - Content Delivery Network, a system of computers on the Internet that deliver content transparently to end users. It facilitates copying pages of a web site to different geographical areas so that they can be accessed quickly by the closest server.

Internetworking - It involves connecting two or more distinct computer networks or network segments via a common routing technology. It has three variants:-

    Extranet – It’s for partner or affiliate companies to access organization’s private LAN, though not has full access.

    Intranet – It is the internal network of an organization, using the Internet Protocol and IP-based tools such as web browsers and file transfer applications, that is under the control of a single administrative entity.

    Internet - It consists of a worldwide interconnection of governmental, academic, public, and private networks based upon the networking technologies of the Internet Protocol Suite. It is the successor of the Advanced Research Projects Agency Network (ARPANET). The Internet is also the communications backbone underlying the World Wide Web (WWW).


VPN –Virtual Private Networks establishes a virtual circuit enabling remote access of organization’s private LAN to employees through Internet.