Sunteți pe pagina 1din 4

11. TCP/IP protocol architecture is used on the Internet.

It consists of the
following layers://Application Layer that includes FTP(transferring files),
HTTP(accessing web pages), SMTP(for email service), MIME(for facilitating
file transfer such as image, video etc), SNMP(simple network management
protocol for managing the internet), TELNET(for accessing a remote
computer), BGP(for updating the routing table fro external routers),
DNS(getting ip addresses from domain name)//Host-to-Host (Transport
Layer): TCP, UDP//Internet Layer: IPv4, IPv6, ICMP(Internet control message
protocol, responsible for status control and report), OSPF(opens shortest
path first responsible for updating routing tables interior routing table),
RSVP(resource reservation protocol providing quality of service)//(Not a
part of TCP)Network Access Layer: Ethernet, Frame relay, ATM, Wi-FI, PPP,
SLIP//Network Access Layer is split into Data link layer and Physical
Layer//Protocols such as ICMP, OSPF, RSVP directly carried by IP//ARP(used
to discover mac address from Ip address)/RARP between Internet Layer and
Network Access Layer. Application Layer Handles communications
between applications and processes. Host-To-Host Layer//End to end
transfer of data//Flow and error control between process
communications//Connection between processes
Internet Layer//Routing of data packets. Network Access Layer//Media
Access Control//Physical transmission of signals//Error control and flow
control on a link by link basis. TCP//Destination Port Number//Sequence
Number//Acknowledgment Number//Window Size for Flow
Control//Checksum. IP//Destination Address//Options//Service
Request//Payload Type//Lifetime of packet. Ethernet Frame: (802.3 Header,
802.2 header, IP header, TCP header, Application layer header, Application
layer payload, 802.3 trailer). This frame is transmitted over Ethernet by
physical layer. IP version 4: IP is the widely used network layer protocol that
is responsible for routing packet from source to destination. IP provides a
datagram (connectionless) service, i.e., a best effort service that does not
guarantee the delivery of packets. IP does not provide any error or flow
control hence we need TCP protocol. IP supports various options including
source routing, route recording, and timestamp recording.IP supports
fragmentation and reassembly of packets to meet the maximum packet size
requirements on various networks. IPv4 Packet Header: (Version (4 or 6),
Header Length (header length could vary depending on the options), Diffserv
(DiffServ is Differential Services, a class-based mechanism to provide QoS on
IP networks. Traffic is classified into classes. Routers then handle various
class traffic differently) and ECN (ECN is Explicit Congestion Notification that
allows for notification of congestion. When a router detects congestion, it
sets the ECN bit to inform the source to slow down), Total Length (length of
the entire IP packet (including data) in octets), Identification (used for
identifying packets especially useful during fragmentation and reassembly),
flags (more flag, dont fragment flag and reserved), fragment offset (in (64
bits multiple) to indicate how the fragments can be joined to get the original
datagram) , time to live(Life time of a packet. Prevent packets from looping
endlessly. Typically represents hop count), protocol (code to identify the
protocol of the payload (e.g. TCP =6)), header checksum (16 bit checksum for
error detection in the header), source address, destination address, options
(security, source routing, route recording, time stamping) and payload.). Max
length of IP packet = 65,535 Octets or bytes
Five classes of IP addresses (now becoming obsolete given Classless Inter
Domain Routing (CIDR)). Class A (0 +7N + 24 H)(1-126). Class B (10 +14N +
16H)(128-191). Class C (110 +21N + 8H)(192-223). Class D (11110 + 19) // for
multicasting. Class E (111110 + 18) // Reserved. Subnet and subnet masks:
Subnets partition large organizational network address spaces into smaller
groups, with each subnet (group of addresses) having the same network
prefix. This leads to better utilization of addresses, as well as single entry in
the routing table. In the previous classful scheme, host portion of address
partitioned into subnet number and host number, with subnet masks used to
interpret the IP addresses- identifying host bits and subnet bits.Currently
with CIDR, a variable length subnet masking (VLSM) scheme used and
denoted by /N notation. This scheme allows for efficient use of IPv4
addresses, and aggregation of prefixes to enable fewer entries in the routing
tables.

Subnet # = decimal of 00010001 = 17 Host# = decimal of 00001101 = 8 + 4 +


1= 13. Subnet no is the decimal equivalent of ip address no corresponding to
1s in the second last octet of subnet masks and host no equivalent of all 0s
in subnet masks. Route Aggregation in CIDR: Instead of assigning address
blocks to end-user organizations, assign larger blocks of addresses to large
network service providers. Organizations acquire addresses from these
network service providers. Routers in the rest of the world only maintain one
entry to the ISPs larger address block. Service providers handle routing
within their networks. Internal networks hidden from outside world.Internet
Control Message Protocol (ICMP): Protocol used on the Internet for control,
as well as for status reporting. ICMP Messages are carried by IP packets.
ICMP Header (Type (8 bits) identifying the type of ICMP message, Code (8
bits) codes that provide additional information, Checksum(16 bits) error
control for header + data, Parameters (32 bits)). ICMP Message Types
(Partial List): Destination unreachable, Time exceeded, Parameter problem,
Source quench, Redirect, Echo, Echo Reply, Timestamp, Timestamp reply,
Address mask request, Address mask reply. Address Resolution Protocol
(ARP): Protocol used to find the physical/MAC address of a host given its IP
address. Source broadcasts an ARP request containing the IP address.
Destination matching the IP address replies with an ARP response containing
its Physical/MAC address. Inverse ARP (InARP): Protocol used to find the IP
address of a computer, given its layer 2 address (such as IP address of a
computer for a given DLCI number in Frame Relay). Uses ARP message
format, but different codes. Autonomous Systems (AS) A network that is
managed by a single entity. AS has two major classes of routing protocols.
Interior router protocol to update the routing tables of interior routers
within AS Open Shortest Path First (OSPF). Exterior router protocol to update
the routing table of border/edge routers. Border Gateway Protocol (BGP)
Each Autonomous system has an AS number that is used in BGP. A multihomed AS maintains connection to multiple AS. Border Gateway Protocol
(BGP): Widely used exterior routing protocol. BGP peers establish a
connection using TCP. Keep Alive message sent periodically to keep the
connection alive. Update messages that advertise feasible routes to
reachable destinations or withdraw routes exchanged among peer routers to
update routing tables. Organizations that are served by multiple ISPs, it is
necessary to run BGP. Open Shortest Path First (OSPF) Protocol: Interior
Router Protocol. Uses link state, where every node receives connectivity
information from other nodes. Each node independently computes the
shortest path using Dijkstras algorithm. The link cost metric is user
configurable and could include delay, data rate, hops, or other factors. Once
the shortest paths are computed, a node updates it routing table. Cisco uses
a proprietary interior routing protocol called: Enhanced Internal Gateway
Routing Protocol (EIGRP).
IP Version6 (IPNG): 128 bits address length supports a larger address space
in comparison to 32bit v4 addresses, which are getting exhausted.IPv6 host
can be configured automatically on being connected to a IPv6 network. On
being connected, a host sends out a multicast request. Routers then respond
to the request by sending a router advertisement packet with network layer
configuration parameters. Multicasting capability (multiple copies of a packet
is made as it travels over the network). Jumbo packets as large as 4GB
allowed. Network layer security provided by IPsec, which is part of IPv6 suite.
IPsec can also be used for IPv4. Optional Headers used. IPv6 Packet Header:
(version 6, traffic class indicated priority, flow label (used for QoS. Packets
with the same flow labels are provided the same QoS), payload length
(length of the payload), next header (points to the next (i.e., options or
higher layer) header), hop limit (lifetime of the packets in hop count), source
address, destination address (Address has two parts: 64 bit network prefix;
64 bit host part (which can be generated automatically from the MAC
address)), options header 1, options header 2, payload (Hop-by-Hop Options
Header (Jumbo Payload, Router Alert-used in RSVP), Destination Options
Header, Routing Header (source routing), Fragment header, Authentication,
Encapsulating security payload) IPv6 Host Configuration:In IPv4, DHCP was
used to configure host. In IPv6, uniqueness of MAC addresses can be used to
let hosts configure themselves-64-bit host address obtained using EUI-64,
Extended unique identifier 64, Extends 48-bit MAC address. Adds ff-ee in
the middle of the MAC address, e.g, 00-10-cd-23-67-b8 00-10-cd-ff-ee-2367-b8.

Binary IP address 10000000 10101011 00010001 00001101 and subnet mask


11111111 11111111 11111111 00000000

One additional step Convert global uniqueness bit to local uniqueness bit 0010-cd-ff-ee-23-67-b8 02-10-cd-ff-ee-23-67-b8. IP Security (IPSec): A suite
of Layer 3 protocols that provide security to IP packets. Consists of
Authentication Header (AH) to verify that the data came from the right
source and was not modified. Encapsulating Security Payload (ESP) for
encrypting data. Internet Key Exchange (IKE) that allows communicators to
negotiate on choice of the authentication method, encryption method, and
key parameters used for secure communications. Transitioning from IPv4 to
IPv6: Dual Stack: Both IPv4 and IPv6 protocol stacks are present. Upgraded
nodes use IPv4 while communicating with IPv4 nodes only. Map IPv4 address
to IPv6 address. Tunneling mechanism whereby IPv6 packet sent inside an
IPv4 packet in networks supporting IPv4 only.

checksum. Socket Programming: IP addresses and port numbers together


allow a computer to send and receive data over the Internet.
Combination called a socket, Serves as the endpoint of communication on
a computer network. Communication typically involves two components.
Server-Listens for any incoming connection requests and responds to them.
Client-Connects to a known listening server
Voice Over IP (VoIP): Sending voice using IP packets. Can support multiple
phone calls over the same access line. Encryption and authentication of
phone call is possible. Problems with jitter, delay, and packet loss. Prebuffering in play-out buffers can address jitter problem. Challenges in
crossing firewalls. Skype uses a proprietary protocol to pass through
firewalls. Session Initiation Protocol (SIP) widely used as a signaling protocol
for VoIP along with H.323 (for videoconferencing over IP). SIP is independent
of transport layer protocol. Need an analog phone adaptor to connect analog
phone to data line. Vonage, and Comcast are major service providers of VoIP.
Real-time Transport Protocol (RTP): Defines a packet format for delivering
audio or video over the Internet. RTP typically carried by UDP. RTP does not
provide any QoS guarantees. Real Time Control Protocol (RTCP) used to
monitor QoS. Provides mechanisms such as timestamp and control for real
time traffic. Support for gateways such as audio/video Bridge. CISCO Router:
Operating System of CISCO Routers is IOS. Can configure router using the
serial port. IOS has command line interface (two modes)-Exec Mode- Read
Only Access and Privileged Exec Mode -> Make Configuration Changes. Use
enable to configure. Each router interfaces would have to be configured. This
also includes access list configuration. Default Route has to be configured.
Common Cisco Routers 2800 series and 6500 series. In some routers VPN
capabilities are available at the router interface. Steps for Designing
Networks: Traffic Estimation, Size link capacities, Select network devices with
the right throughput and interfaces. If certain needed functionality (such as
VPN) missing from interfaces, then add additional hardware/software. Note,
never put a switch as an edge device.

12. The need for a Transport layer: The network layer (IP) sends packets of
data to their correct destinations with best effort. Packets may get lost in
transit, therefore, network layer is not adequate for most applications to
work correctly. The maximum packet size in IP is 65,536 bytes, what happens
if the application wants to send a file of size 5,000,000 bytes? Need to break
files into smaller blocks at the sender end and re-assemble these blocks into
the file at the receiving end. Port addressing-Distinguish between
applications on computers. Transport Layer Services: Providing end-to-end
reliable delivery of data by taking data from the source application and
delivering it to the destination application. Multiplexing application layer
data over a single transport connection. Providing end-to-end flow control.
Recovering from crash. Ensuring ordered delivery of data. Providing
mechanisms to address applications. Segmenting application data. Transport
Layer Protocol Types: Connection oriented TCP (Reliable transfer of data,
Flow control, ordered delivery, High Overhead) Connectionless UDP (Best
Effort Service, Reliable delivery not guaranteed, No flow control or
sequencing of packets, Low Overhead Service) TCP Functions Reliability:
TCP recovers from network damage to dat. Basic mechanism (Receiver sends
a positive acknowledgment (ACK) if all goes well. If the ACK is not received
within a timeout interval, the sender retransmits the data). Datagrams may
also get duplicated (Sequence numbers help identify these duplicates).
Multiplexing: TCP supports multiplexing by carrying multiple application
layer messages to the same destination application over a single TCP
connection, which occurs over a specific TCP port. A network address and
port address together is called a socket. A pair of sockets uniquely identifies
each connection. Standard Ports: 80-web (http), 25: email (smtp), 443: SSL
(https), 445: microsoft-ds (smb), 20-ftp data, 21-ftp, 23-telnet to access a
remote computer, 53-domain name service DNS. TCP Sliding Window
Control: Flow control is the regulation by the receiver of the amount of data
the sender may send. Creates a sliding window at the sender. Packets that
have received permission for transmission are within the window. Window
slides as receiver acknowledges packets or modifies window size. TCP
Functions Connection Establishment: Before communication starts, sender
and receiver negotiate a set of sequence numbers to use in TCP, Called 3-way
handshake. Sequence numbers are core part of TCP. It is not a good idea to
reuse the same sequence numbers in succession. Creates problems in
detecting duplicates. Transmission Control Protocol (TCP) Header: source
port, destination port, sequence no (sequence number of the first byte
carried in the data), Ack no (valid if ACK flag is set. It is the sequence number
of the next byte that the receiver expects next. ACK 1079 implies that all data
till byte number 1078 have been received correctly), data offset (Size of the
header, indicates the starting position of data), reserved (Not used. Should
be set to 0), flags (URG, ACK, PSH flag when set pushes data in the buffer,
RST set to reset connection, SYN used to set 3 way handshake, FIN no more
data from sender), window (indicates the number of bytes that can be
received by the receiving side. Used in flow control), checksum (used for
error checking of header and data), urgent pointer (points to urgent data
when the URG flag is set), options (optional). User Datagram Protocol (UDP):
Defined in RFC 768 (1980). Many applications do not need TCP, e.g when
applications send very small amounts of data (DNS), Prefer speed to
reliability (voice). In these cases, if we can avoid TCP, we eliminate the
overhead of keeping track of sequence numbers, window sizes etc. it is a low
overhead protocol. UDP Header: Source port, destination port, Length,

7. LAN: Distinguishing Characteristics-Spans a limited geographical area


comprising a few square miles. Typically data transmission at high rate.
Private ownership. Benefits: Hardware sharing, Software sharing, Data
sharing, Access to other networks. LAN Architecture: IEEE 802 Reference
Model: (Physical, Media Access control, Logical ink control). In OSI model
data link layer is broken into two parts, Media Access control and logical ink
control. Medium Access Control (MAC)-At the transmitters end, data is
assembled into frames, with the addition of flag, address, control and error
detection fields.MAC protocol is then used to transmit frames. At the
receivers end, frame is disaggregated and error detection is performed.
Logical Link Control (LLC)-Provide an interface to higher layer, Perform flow
control and error control (note MAC only does error detection), various
services offered including acknowledged and unacknowledged service. MAC
Frame: MAC header, LLC header, IP header, TCP header, Application layer
payload, MAC Trailer. Ethernet Frame: Flag, Delimiter, Dest Address, source
address, length, dest sap, source sap, LLC control, payload, CRC. LAN
Topologies Bus and Tree: Broadcast topologies, Can support full duplex
transmission, Media access control method needed to share bandwidth,
Signal absorbed by terminators at the end points. Ring Topology: Typically
point-to-point link connected by a repeater to station. Token passing based
media access mechanism used. Star Topology: Each station connected to a
central node. Central node could be a hub or a switch. Station addition or
removal does not disrupt the operations of the network. Hubs: Only one
station could transmit at a time. Ethernet hubs are broadcast devices where
every port receives the transmission and can lead to collision (Physical star
logical bus topology). Token Ring hubs could circulate a token inside the hub
(Physical star logical ring topology). Switches: Temporary connection
established to provide point-to-point connectivity between ports. More than
one station could transmit in parallel thereby increasing throughput. No
collision due to buffering. The throughput of an Ethernet switch is greater
than the throughput of an Ethernet hub. IEEE 802.3 100-Mbps (Fast
Ethernet) 100Base-TX (100 Mb/s, Baseband Signaling (MLT3))-2 Cat. 5 UTP2 STP. 100Base-FX (100 Mb/s, Baseband Signaling)-2 strands of multimode

fiber-Encoding: 4B5B/NRZI. Gigabit Ethernet (IEEE 802.3-2005) - Data rate =


1 Gigabit/s, Frames are longer, Frame bursting allows multiple frames to be
sent by a station, 1000BaseT UTP (Cat 5, 6), 100 m, 1000Base-SX multimode fiber, 500 m, 1000Base-LX - single mode fiber, 2km. 10 Gigabit/s
Ethernet (10GbE): Uses Switches and full duplex links, 10GBASE-SR (short
range, uses multi-mode fiber, up to 300 m), 10GBASE-LRM (distances up to
220 m, on FDDI grade multi-mode fiber), 10GBASE-LR (Long Range, uses
single mode fiber, can span up to 25 km), 10GBASE-ER (Extended Range, can
span up to 40Km, use single mode fiber). Functions of NIC: Connects
computers to Media, Creating Layer 2 Frames, Media Access Control,
Transmitting and Receiving frame, Providing a unique Address to the
computer (i.e., Layer 2 address) Cables and Connectors: UTP - RJ-45, Thin
Coaxial - BNC (or T) connector, Thick Coaxial -AUI(or DIX) connector, USB, ST
Connector for fiber, SC Connector for fiber, FC Connector for fiber (primarily
for single mode fiber) Network Operating Systems: Provide Interface
between LAN hardware and application software. Two components-Client
Network Operating System on Client Machines (Redirector Function and
Support for Network transport protocol). Server Network Operating System
on File Server (Directory Service, File Service, Printer Service, Networking and
Connectivity Service, Network Management and Administration Service)

approach. Virtual circuit approach is less flexible in handling congestion,


node or link failures on the designated path, while the datagram approach is
more responsive to congestion and failures. Frame Relay: it is a fast packet
switching technology for wide area network. Fast Packet Switching
Technology based on variable length packets that is used in WANs. Since
multiplexing and switching of packets done at layer two, packets in frame
relay are known as frames. It is a connection-oriented service, where
permanent virtual circuits are established between the end points. Switched
virtual circuits are also available that are established on the fly to support
voice over frame relay. Frame relay protocol is highly streamlined with
reduced processing at nodes and increased throughput (typically < 45 MB/s).
For example, there is no hop-by-hop error or flow control. Any lost/discarded
frames are detected at the end points by higher layer protocols and then
error recovery is initiated. FRAD (Frame Relay Assembler/Dissembler): FRAD
Responsible for assembling and disassembling frames. Sometimes includes
error detection. is a box, usually close to the user, that provides the interface
between the user and the frame relay network. Sometimes included in a
router. Frame Relay Access: DDS: 56 & 64 Kbps, Fractional T-1: 128, 256 &
384 kbps, DS-1: 1.544 Mbps, DS-3: 45 Mbps Virtual Circuits (VC): Permanent
VC (Predefined VC, Identified by DLCI (Data link connection identifier)).
Switched VC (Setup and torn down as needed, Identified by DLCI, Primarily
used for Voice over Frame Relay and ATM/ Frame Relay internetworking).
DLCI: Frame headers in frame relay networks do not use absolute addresses
of computers but instead use virtual circuit numbers. Parameters needed to
enforce traffic in order to manage congestion, and to specify a service level
agreement (SLA) between the customer and the service provider. Committed
Information Rate (CIR): Parameter that specifies the data rate during normal
operations. It is defined for each VC. Frames that are transmitted in excess of
the CIR are marked Discard Eligible (DE) and are discarded if congestion is
encountered. CIR is less than the access rate. Maximum Rate: Frames that
are transmitted at rates higher than maximum rate are discarded.
Committed Burst Size (Bc)-Maximum amount of data that a user is allowed
to send during a predefined time interval. Excess Burst Size (Be)-Maximum
amount of data exceeding Bc, that a user could send during a predefined
time interval. Frames marked discard eligible are discarded only if there is
congestion in the network. Congestion Management: Forward Explicit
Congestion Notification (FECN) - A bit on the frame header is set to indicate
that the congestion should be reduced in the direction of the frame flow.
Receiving station should intimate the source to slow down. Backward Explicit
Congestion Notification (BECN) A bit on the frame header is set to indicate
that the congestion should be reduced in the direction opposite to the
direction of the frame flow. Receiving station then slows down. Discard
Eligible (DE): Frames that are marked DE are dropped by an intermediate
node if congestion is encountered. Traffic policing based on max rate.
Multiprotocol Label Switching (MPLS)-1: Regarded as layer 2.5 packet
switching technology based on connection-oriented approach. Routes traffic
based on labels, rather than network address, and therefore avoids lookups
in a routing table. Labels contain virtual links (path) information. MPLS can
switch all kinds of network access technology payload such as T1, ATM,
Frame Relay, DSL, SONET and Ethernet frames. An MPLS header is added to a
packet. It contains one or more labels. MPLS packets are switched based on
label lookup. Label Edge Routers (LER) are the entry and exit points of MPLS
network. A label-switched path (LSP) is setup by a signaling protocol based
on criteria in Forwarding Equivalence Class (FEC) MPLS-2: A label-switched
path (LSP)is setup starting at the Label Edge Router is setup by a signaling
protocol based on criteria in Forwarding Equivalence Class (FEC) (group of
packets that need to be forwarded the same way (based on QoS, destination,
VPN membership) belong to the same FEC and bound to the same MPLS
label). When a packet enters a MPLS cloud via LER, labels are pushed, while
labels are popped out of a packet exiting MPLS cloud. The label added is
based on the forward equivalence class (FEC). Label is checked against the
index in a table to obtain the next hop, and a new label replaces old label. No
routing decisions made after LER. At any MPLS router the top most label is
examined and accordingly the following operations are performed - Label
push i.e., encapsulation, label pop (de encapsulation) and label swap
(replacing one label with another).

8. Circuit Switching: it is the dominant switching technology in phones


networks while packet switching is dominant in data nw. A path setup
between source and destination and bandwidth dedicated on this path for
communication. No other user can transfer if a dedicated path is set up.
Three phases: Call setup, Communication, Call disconnect. When a call is
disconnected bandwidth is returned to other user to use. Switches have the
ability to setup the path between source and destination. Circuit Switching
Features: Inefficient utilization of bandwidth, Set up time involved in setting
up the circuit, Transfer is transparent the network does not affect the
traffic being sent, Calls can be blocked, if a circuit cannot be set up, No
queuing delays. Fixed Alternate Routing: Order in which routes are tried are
fixed: 2, 1, 3. Dynamic Alternate Routing: Order in which routes are tried can
change during the hours of the day depending upon traffic and congestion.
Routing tables are updated periodically because in case of node or link
failure, if Traffic pattern changes over the different time zones of the day, in
case if a new node is added in the network so to have the updated
information of all the interconnected node in the network. Advantages of
Packet Switching Over Circuit Switching: Transmission efficiency (Bandwidth
is not dedicated on a call-to-call basis, and hence more efficient utilization of
bandwidth), Flow Control (Network can provide buffers to prevent a fast
transmitter from drowning a slow receiver), Non-blocking nature- packets
are accepted even when network is busy, Routing can be done based on
priorities. Types of Packet Switching-1 Datagram Approach (connectionless)
used by IP on the internet: Every packet is routed independently of other
packets. Packets belonging to same message can take different routes.
Packets can arrive out of sequence. Routing decisions made at each
intermediate node. No error or flow control provided, hence packets could
be lost. Thus higher layers at the destination need to take care of error
control, flow control and sequencing. Dijkstra's shortest Path algorithm to
compute the minimum hop routes routing and congestion control
mechanism, queuing system will be applied to calculate the link and flow
control. Virtual Circuit Approach (connection-oriented): A path is set with no
bandwidth dedicated (virtual circuit) before transmitting any packets. A call
request packet is sent as a datagram from source to destination, while a call
accept packet is sent back. These packets help configure intermediate nodes
and set up the virtual circuit. Packet header contains a virtual circuit number
instead of a host address. Routing decisions are not made at intermediate
nodes. Sequencing of packets also provided. Still a store and forward
technology with packets incurring queuing delays. Error control and flow
control provided by VC. Virtual Circuit (VC) Approach vs Circuit Switching:
Bandwidth not dedicated in VCs while it is in circuit switching, Store-andforward mechanism leading to queuing delays in VCs, which is not true in
circuit switching, Transparency is Circuit Switching, where the network does
not affect the traffic in a circuit. Datagram vs Virtual Circuit: Virtual circuit
approach involves setting up a path, which is avoided in datagram approach.
Thus datagram approach is better suited for short transactions done on the
fly. Virtual circuit approach can provide sequencing, flow control and error
control, while datagram does not provide any of these services. In virtual
circuits, packets can be forwarded more rapidly at intermediate nodes since
no routing decisions are made, while this is not the case with the datagram

9. Wireless Personal Area Networks (WPAN): Wireless network to connect


devices in close proximity to each other, such as connecting PDA to mobile
phones, wireless mouse, keyboard and other peripherals to computer.
Range: 1/3 m 10m Ability of devices to locate each other and connect
automatically. Ability to lock out other devices selectively to prevent

interference. Standards: Bluetooth 2.1(IEEE 802.15) 1Mb/s (original rate).


Point-to-point and point-to-multipoint, fixed or mobile devices. Uses radio
waves ISM band 2.4GHz to 2.4635GHZ. Uses frequency hopping spread
spectrum, with hopping over 79 channels, with each channel having a
bandwidth of 1 MHz. Enhanced Data Rate 2.1 Mb/s (optional feature).
Bluetooth 4.0 24Mb/s, Uses 40 channels with 2Mhz bandwidth spacing.
IrDA: Uses infrared light, point-to-point, range: 1 m, data rate: 2.4kb/s16Mb/s. Does not travel across walls, low power due to health reasons.
Wireless USB: Used in printers, digital cameras, game controllers, portable
media players, and storage devices. Based on Ultra-Wide Band platform
operating in 3.1GHZ-10.6GHz range, and can pass through obstacles in the
transmission. Data rate between 53 Mb/s to 480 Mb/s. Wireless Local Area
Networks(WLANs): Wireless data networks with a range of 50-100 meters.
Set up for mobility, ad hoc networks, or when dictated by geographical
considerations. Supports two modes-Ad hoc mode allows peer-to-peer
transmission, Infrastructure mode requires communication via an access
point. Access points in LAN are fixed in contrast to PAN. Use unlicensed
Industrial Scientific Medical Use Band (max transmission power limited by
FCC to 1 watt). Interference from devices such as microwave oven. IEEE
802.11a, IEEE 802.11b, IEEE 802.11g. IEEE 802.11n, IEEE802.11ac,
IEEE802.11ad (WiFi). Wireless Access Point (WAP) versus Wireless Routers:
WAP enables connection of wireless devices to a wireless network. WAP has
one Ethernet port, and acts as a wireless hub (in Gateway mode). Wireless
Routers can route packets; provide private IP addresses to computers
connecting to it ; provide security by hiding private IP addresses. One WAP
for every 2500 sq feet, with a maximum of 15 users. Could use directional or
omnidirectional antennas to create hotspots based on the room layout to
avoid overlap of cells. Wireless Metropolitan Area Networks (MAN):
Network that covers a larger area ranging from few square miles to a city,
compared to Wi-Fi. Application: mobile broadband connectivity, wireless
alternative to local loop and cable. Wireless Local Loop (WLL) (IEEE
802.16)(WiMAX)- 30-40Mb/s, Supports QoS. Wireless local loop connectivity
for voice, data and video. Line of sight requirement for 802.16 (use 10-66
GHz band). No line of sight requirement for IEEE 802.16a (use 2-11GHZ
band). IEEE 802.16d uses orthogonal frequency division multiplexing with
256 sub carriers. IEEE 802.16e uses scalable orthogonal frequency division
multiple access. IEEE 802.16m data rates of 100Mb/s mobile and 1 Gb/s
fixed. Access points typically connected using point-to-point microwave links.
Could be handled Cellular technology- UMTS, GPRS, CDMA2000, LTE, etc.
Mobitex packet switched, narrow band (12.5KHz channels) wireless
technology that supports only data. Fourth Generation (4G) Cellular
Network: Peak requirements 100Mb/s for high mobility communication;
and 1Gb/s for low mobility communication. Provide comprehensive all IP
based mobile broadband communication. Dynamically share and use
network resources to support an increasing number of simultaneous users.
Key technologies for 4G (Open Wireless Architecture (OWA), supporting
multiple wireless air interfaces, in an open architecture platform, Long Term
Evolution (LTE)- Based on GSM/EDGE and UMTS/HSPA, IMT compliant
versions of LTE and WiMax under development- LTE Advanced and Wireless
MAN-Advanced)

stream and send it to the internet router. A filter built in the DSLAM unit can
separate voice and send it to phone switch. Combines voice and data signal
and sends it on the DSL line on the opposite direction. Located close to
customer or at the central office of Phone Company). ADSL channel config:
FDM is used to create sub channels bands between upstream and
downstream. In this case, a single bit stream is split into multiple parallel bit
streams and each portion is carried in a separate frequency band. Echo
cancellation is a signal processing technique that allows transmission of
digital signals in both directions on a single transmission line simultaneously.
In essence, a transmitter must subtract the echo of its own transmission
from the incoming signal to recover the signal sent by the other side. When
echo cancellation is used, the entire frequency band for the upstream
channel overlaps the lower portion of the downstream channel. This has two
advantages compared to the use of distinct frequency bands for upstream
and downstream: the higher the frequency, the greater the attenuation, and
echo cancellation design is more flexible for changing upstream capacity. The
disadvantage of the use of echo cancellation is the need for echo cancellation
logic on both ends of the line. Discrete Multitoned Modulation (DMT): Used
to create sub channels in the downstream frequency band. Use multiple
carrier signals at different frequencies to create 4KHZ subchannels. Based on
orthogonal frequency division multiplexing. Discrete multitone (DMT) uses
multiple carrier signals at different frequencies, sending some of the bits on
each channel. The available transmission band (upstream or downstream) is
divided into a number of 4-kHz subchannels. On initialization, the DMT
modem sends out test signals on each subchannel to determine the signalto-noise ratio (SNR). The modem then assigns more bits to channels with
better signal transmission qualities and less bits to channels with poorer
signal transmission qualities. Stallings DCC8e Figure 8.18 illustrates this
process. Each subchannel can carry a data rate of from 0 to 60 kbps. The
figure shows a typical situation in which there is increasing attenuation and
hence decreasing signal-to-noise ratio at higher frequencies. As a result, the
higher-frequency subchannels carry less of the load. Present ADSL/DMT
designs employ 256 downstream subchannels. In theory, with each 4-kHz
subchannel carrying 60 kbps, it would be possible to transmit at a rate of
15.36 Mbps. In practice, transmission impairments prevent attainment of
this data rate. Current implementations operate at from 1.5 to 9 Mbps,
depending on line distance and quality. Access Technology: Cable Modems:
Access service provided by cable companies. Residual bandwidth on either
side of TV signals used for data service without disturbing TV signal.
Symmetrical or asymmetrical. In case of asymmetrical, upstream data rate
can be as high as 10 Mbps, while downstream can be as high as 30 Mbps.
Downstream digital data placed on 6 Mhz TV channels. Shared Medium, data
rate determined by the number of active users attached to Access Node.
Data over Cable Service Interface Specifications (DOCSIS) 3.0 is a standard
that defines the interface requirements for operations and communication
support for data over cable. Uses TDMA for media access. Uses QAM for
modulating carrier signal. Cable Modem Network Components: Cable
Modem provides access to data signal over Cable TV system. Connects to
computer via an Ethernet interface. QAM, QPSK used for upstream data.
Headend Transmitter transmits TV signals and data signals on the cable.
Cable Modem Termination System (CMTS) offers similar functionality as
DSLAM, i.e., aggregating data traffic from multiple channels and sending it to
the router for transmission over Internet. Splitter separates TV signal from
data signal. At the head end there are servers for accounting and logging,
DHCP, and assigning IP address. What are some benefits of partitioning the
48-bit MAC addresses into 2 parts, with one part identifying the
manufacturer of the NIC? We can track the manufacturer and the location
details from the MAC address. Each combination of MAC address along with
manufacture ID is unique and hence we can avoid the duplicating issues. As
manufacturer information is globally embedded and maintained in MAC itself
it helps to reduce the overhead of costs. It avoids conflicts between interface
address and physical address as OUI is unique for each and every
manufacturer.

10. Regulation: Regulation is needed when competition cannot be sustained


in the market place, when there is a social need, or when some coordination
is essential for enforcing standards and technical compatibility. Access
Technology: Digital Subscriber Line (DSL): Access technology that utilizes
unused bandwidth on the regular telephone copper access lines for
computer data. The lower portion of the bandwidth is used for POTS, the
remaining bandwidth left 25000HZ and above is then used for computer
data. DSL Features: Based on digital transmission, Upload and download data
rates are different in Asymmetric DSL (ADSL), Max date rate is 24 Mbit/s
downstream and 1 Mbit/s upstream based on the distance from DSLAM
ADSL2Plus (ITU G.992.5) DSL Network Components: DSL Modem (Connects a
computer to DSL line, Computers commonly connected to DSL modem using
Ethernet card, occasionally connected using USB, Uses 25KHz and higher
frequency band) DSL Router (Connects a group of computers to DSL line)
Splitter (Separates the POTS signal from the data signal). DSL Access
Multiplexer (DSLAM) (Terminates DSL circuits and aggregates the data

S-ar putea să vă placă și