Networking Theory (Part 1)

Introduction
Overview of the basic concepts of networking Also discusses essential topics of networking theory

What is a Network?
A network is a collection of devices that share a common communication protocol and a common communication medium. Devices - computers, printers, telephones, televisions, coke machines, etc.

What is a Network?
Computing-centric model - services and devices bound to individual machines Network-centric model - services and devices are distributed across a network Network and software standards (e.g. Jini) exist to allow devices and hardware talk to each other over networks and to allow instant plug-and-play functionality

What is a Network?

Besides devices providing services, there are also devices that keep the network going, for example,
Network

cards - to allow a computer to talk to a network. E.g. ethernet card. Routers - machines that direct data to the next "hop" in the network Hubs - allow multiple computers to access a network Gateways - connect one network to another. E.g. a LAN to the Internet.

How do Networks Communicate?

Networks consist of connections between computers and devices. Connections:

Wires

and cables - use electricity for transmitting data Wireless - use infrared / radio Fiber-optic cables - use light

How do Networks Communicate?

Connections carry data (bits - 0's and 1's) between one point (node) in the network and another. For data to be successfully delivered to individual nodes, these nodes must be clearly identifiable.

Addressing
Each node in a network is typically represented by an address. The manufacturer of the network interface card (NIC) is responsible for ensuring that no two card addresses are alike, and chooses a suitable addressing scheme. Each card will have this address stored permanently, so that it remains fixed.

Addressing
There are many addressing schemes available. E.g. Ethernet network cards are assigned a unique 48-bit number. This physical address is referred to by many names, such as:

Hardware

address Ethernet address Media Access Control (MAC) address NIC address

Addressing
Often, machines are known by more than one type of address. E.g. a network server may have a physical Ethernet address as well as an Internet Protocol (IP) address, or it may have more than one network card. For inter-network communications, the IP address is used.

Data Transmission Using Packets

Sending individual bits of data from node to node is not very cost effective.
Overhead

involved - e.g address of destination node.

Most networks group data into packets.

Data Transmission Using Packets

A packet consists of a header and data segment.

Header fields Data 1101000111010100001

The header contains:

Addressing

information (e.g sender & recipient) Checksums to ensure packet has not been corrupted Other info needed for transmission across network

Data Transmission Using Packets

To transmit data, a direct connection is usually not available. So packets are sent to their destination nodes via intermediary nodes in the network. Due to network conditions (such as congestion or network failures), packets may take arbitrary routes, and sometimes may be lost or arrive out of sequence.

Data Transmission Using Packets

Packet transmission and transmission of raw bits are low-level processes. Most network programming deals with high-level transmission of data.

Communication Across Layers

The concept of layers was introduced to acknowledge and address the complexity of networking theory. The most popular approach to network layering is the Open Systems Interconnection (OSI) model created by the International Standards Organization (ISO)

Communication Across Layers

The OSI model groups network operations into seven layers.

Communication Across Layers

Each layer is responsible for some form of communication task, but each task is narrowly defined and usually relies on the services of one or more layers beneath it. Generally, programmers work with one layer at a time; details of the layers below are hidden from view.

Layer 1 - Physical Layer

This layer is network communication at its most basic level. At this level, networking hardware transmit sequence of bits between two nodes. Java programmers do not work at this layer - it is the domain of hardware driver developers and electrical engineers. No real attempt is made to ensure errorfree data transmission

Layer 2 - Data Link Layer

This layer is responsible for providing a more reliable transfer of data, and for grouping data together into frames. Frames are similar to data packets but are blocks of data specific to a single type of hardware architecture. Frames have checksums to detect errors in transmission. Corrupted frames are discarded so that they will not be passed to higher layers.

Layer 3 - Network Layer

The network layer deals with data packets which are sent across the network. Communication at this level is still very low-level; network programmers are rarely required to write software services for this layer.

Layer 4 - Transport Layer

This layer is concerned with controlling how data is transmitted. It deals with issues such as automatic error detection and correction, and flow control (limiting the amount of data sent to prevent overload).

Layer 5 - Session Layer

The purpose of this layer is to facilitate application-to-application data exchange, and the establishment and termination of communication sessions. Connection-oriented communication can increase network delays and bandwidth consumption. Some applications choose to use a connectionless form of communication.

Layer 6 - Presentation Layer

This layer deals with data representation and data conversion.

Different

machines use different types of data representation (e.g. 8-bit integers on one system and 16-bit integers on another). Data compression Data encryption

Layer 7 - Application Layer

This layer is where the vast majority of programmers write code. Protocols for this layer dictate the semantics of how requests for services are made (e.g requesting a file). In Java, almost all network software written will be for this layer.

Advantages of Layering
Helps simplify networking protocols. Protocols can be designed for interoperability

Software

that uses Layer n can talk to software running on another machine that supports Layer n, regardless details of the lower layers.
Example: a network layer protocol can work with an Ethernet network and a token ring network.

Networking Theory (part 2)

Internet Architecture
The Internet is a worldwide collection of smaller networks that share a common suite of communication protocols (TCP/IP). It is an open system, built on common network, transport and application layer protocols, while granting the flexibility to connect a variety of computers, devices and operating systems to it.

Design of the Internet

The Internet is the result of many decades of innovation and experimentation. The TCP/IP protocols have been carefully designed, tested and improved over the years.

Design of the Internet

Major design goals:

Resource

sharing between networks Hardware and software independence Reliability and robustness

Fault tolerant protocols - data could be rerouted depending on the state of the network

"Good"

protocols that are efficient and simple.

TCP/IP Protocol Suite

Major protocols:
Internet

Protocol (IP) Internet Control Message Protocol (ICMP) Transmission Control Protocol (TCP) User Datagram Protocol (UDP)

Internet Protocol (IP)

IP is a Layer 3 protocol (network layer) It is used to transmit data packets over the Internet It is the most widely used networking protocol in the world. IP acts as a bridge between networks of different types

Internet Protocol (IP)

IP is a packet-switching network protocol. Information is exchanged between two hosts in the form of IP packets (IP datagrams). Each datagram is treated as a discrete unit - there are no "connections" between machines at the network layer. Connection services are provided by the higher-level protocols at the transport layer.

Internet Protocol (IP)

The IP datagram consists of a header and the actual data being sent. The header contains essential information for controlling how it will be delivered.

IPV4 datagram format

Internet Protocol (IP)

Although each machine has its own physical address, each host machine under the Internet Protocol must be assigned a unique IP address. The IP address is a four-byte (32-bit) address. Example: 192.168.1.5 The IP address is not bound to a particular physical machine. Network programming in Java does not require the use of the physical address; only the IP address is used.

Internet Protocol (IP)

Humans do not find IP addresses easy to remember. An addressing scheme is also used which allows the use of textual names (hostnames) instead of numerical values. Example: java.sun.com

Internet Control Message Protocol (ICMP)

The Internet Protocol provides absolutely no guarantee of datagram delivery. The Internet Control Message Protocol (ICMP) is a mechanism for error-control. It is used in conjunction with the Internet Protocol to report errors when and if they occur.

Internet Control Message Protocol (ICMP)

The relationship between IP and ICMP is a strong one.

E.g:

IP uses ICMP if it needs to notify another host of an error. ICMP requires IP to send the error message.

Note that a host cannot rely solely on ICMP to guarantee delivery as there is no guarantee that ICMP messages will be sent or that they will reach their intended destination.

Internet Control Message Protocol (ICMP)

Five error messages are defined:

Destination

Unreachable

If a gateway is unable to pass a datagram on to its destination, this message is sent back to the original host.

Parameter

Problem

This message is sent to the sending host if a gateway is unable to process the header parameters of an IP datagram.

Internet Control Message Protocol (ICMP)

Redirect

If a shorter path, or alternate route, is available, a gateway may send this message to the router that passed on a datagram

Source

Quench

This message may be sent in an attempt to reduce the number of incoming datagrams when a router, gateway or host becomes overloaded.

Time

Exceeded

Whenever the TTL value of a datagram reaches zero is discarded. This message may be sent if this event occurs.

Internet Control Message Protocol (ICMP)

ICMP supports several informational messages such as:

Echo

Request/Echo Reply

Used to determine whether a host is alive and can be reached.

Address

Mask Request/Address Mask Reply

Provides the functionality to determine the address mask which controls which bits of an IP address correspond to a host, and which bits determine the network/subnet portion.

Transmission Control Protocol (TCP)

TCP is a Layer 4 protocol (transport layer) that provides guaranteed delivery and ordering of bytes. TCP uses IP to send TCP segments, which contain additional information that allows it to order packets and resend them if they go astray.

Transmission Control Protocol (TCP)

TCP uses communication ports to distinguish one application or service from another. A host machine can have many applications connected to one or more ports. Although TCP provides a simpler programming interface, it may reduce network performance.

User Datagram Protocol (UDP)

UDP is a Layer 4 protocol (transport layer) that applications can use to send packets of data across the Internet (as opposed to TCP, which sends a sequence of bytes). UDP also supports communication ports. UDP does not guarantee delivery packets. It also does not guarantee that they will arrive in the right order. Although unreliable, UDP offers faster communication.

Internet Application Protocols

Network programmers are more interested in the protocols at the application layer. Examples:

Protocols

for accessing and sending email Protocols for transferring files Protocols for reading Web pages

Telnet
A

service that allows users to open a remoteterminal session to a specific machine. Uses TCP port 23.

File Transfer Protocol (FTP)

Allows

file transfers Uses TCP port 21 (to control sessions) and TCP port 20 (for the actual transfer).

Post Office Protocol version 3 (POP3)

Used

to access e-mail Allows users to read mail offline. Uses TCP port 110.

Internet Message Access Protocol (IMAP)

Less

popular than POP3 as it requires continual connection to the mail server. Message are stored on a server and not on the user's system. Uses port 143.

Simple Mail Transfer Protocol (SMTP)

Allows

messages to be delivered over the Internet. Uses port 25.

HyperText Transfer Protocol (HTTP)

One

of the most popular protocols in use on the Internet; it made the World Wide Web possible. Java provides good HTTP support. Uses TCP port 80.

TCP/IP Protocol Suite Layers

Although there are seven OSI network layers, not all are used in Internet programming. The layers beneath the network layer are encapsulated from the network programmer.

TCP/IP Protocol Suite Layers

HTTP SMTP FTP TCP POP3

UDP
Internet Protocol

ICMP

TCP/IP Stack

TCP/IP Protocol Suite Layers

POP3 command: TCP segment: RETR1 Application layer

TCP header

RETR1

Transport layer

IP datagram: IP header TCP header Modem frame:

RETR1 Network layer

Data link layer RETR1

Physical layer

Modem header IP header TCP header Carrier wave:

Security Issues: Firewalls and Proxy Servers

Firewall: while being an excellent tool for n/w admin often is a developers worst enemy It blocks direct UDP and TCP access, making application that uses these protocols practically unusable Need to adapt software to proxy requests using protocol such as HTTP Direct UDP/TCP communication is simpler and offers better performance Using proxy server and HTTP can add delays

Proxy Server

It is a machine that acts as a proxy for application protocols The server accepts incoming connections from machines within a local n/w and makes requests on their behalf to machines connected to the Internet Advantages:
Direct

access to internal machines is never established

Proxy Server
Only popular protocols such as HTTP is permitted access while newer application such as games or RealAudio is not permitted Most proxy servers also log networking events to allow n/w admin to track unusual communications and their origin