Forum of Layer 2 Layer 3 technologies

PPP vs HDLC and L3 switch vs Router

2009-11-11T22:28:00.005+05:30

Difference between PPP and HDLC :
=================================
1 . Authentication :
PPP supports the authentication and HDLC does not support the authentication

2 . Fragmentation Handling :
PPP handles the fragmented packets in better way . Jitter can be controlled more effectively by PPP

3 . Compression Support :
PPP supports the compression and HDLC does not support the compression

4 . Standard :
PPP is OSI standard protocol and HDLC is cisco proprietary protocol

5 . Deployment :
PPP is used in async dial up networks . eg PPPoE , PPPoA . So it is used in LAN
HDLC is used in point to point serial links. It is used in WAN

Difference between L3 switch and Router :
=========================================
1 . Switching speed :
L3 Switch does the hardware switching using ASIC . So it is faster .
Router does the software switching using microprocessors and device drivers. So it is comparatively slower than switch in LAN

2 . Physical component Support :
L3 switch has only ethernet ports and supports only ethernet traffic .
Router has the interfaces of many carrier protocols such as ATM , Ethernet , POS . It needs to support encapsulation from one carrier to another carrier

3 . Deployment :
L3 switch is deployed in LAN . Router is deployed in WAN .

4 . Broadcast :
Router creates multiple collision and broadcast domains .
L3 Switch create one broadcast and multiple collision domain based on the configuration .

Note :
======
Please share your thoughts on the topic and make this blog as forum of technical topics . Please let me know the topics you want to know . It would help me to post the same here

Telecom Testing Openings

2009-11-06T22:21:00.004+05:30

Telecom Testing Openings :

Requirement : EMS/NMS testing .
Domain Knowledge : L2/L3 , switches , routers preferable
Experience : 3+ years
Job Location : Chennai

This is an urgent requirement

Requirement : L2/L3 testing
Domain knowledge : L2/L3 protocols with TCL , EXPECT
Experience : 3+ years
Job Location : Bangalore
Please send your resume to mkvsenthilkumar@gmail.com

virtual circuit connectivity verification

2009-10-18T01:53:00.003+05:30

Virtual circuit Connection Verification (VCCV) :
==================================================
VCCV provides connection verification services such as ping ,mpls ping regardless of underlying protocol such as MPLS ,IP tunnel. A network operator may use this to test the liveliness of the network.
Ping and other IP messages are encapsulated using the PWE3 encapsulation .These messages are referred to as VCCV messages.VCCV messages are exchanged after negotiation between PEs.

MPLS as PSN :
-------------
VCCV creates control channel between PWE3 PES to exchange the IP monitoring tools. For more details of this , please refer PWE3 architecture.Packets sent across this channel are IP Packets ,allowing maximum flexibility.
When control word is present on VC , it is possible to indicate the control channel by setting the control channel header . this is referred as inband MPLS VCCV as the control channel would be in band.
When the control header is not in use , use of MPLS router alert label to indicate the IP control channel is also proposed.

IP probe traffic :
-------------------
Both ICMP and LDP ping are used for connectivity verification.
ICMP ping :
------------
When ICMP Packets are used , the source address is source address of LDP session and the destination address is destination address of LDP session.
The identifier and sequence number of fields of ICMP echo request and reply are used to track what VCs are tested.These fields are only intrepretted by sending PEs.

MPLS ping packet :
------------------
The MPLS ping packet must contain the sub TLV 8 of PW circuits.This sub TLV contains VC id of circuit to be verified.
L2 circuit ID TLV for MPLS ping :
TLV contains the following fields
1 . Source and destination address of LDP session
2 . VC ID
3 . Encapsulation type
4 . PWID
FEC 128 VCID , FEC 129 attachment
5. PW parameters – interface parameters
L2TPV3 as PSN :
-----------------
When L2TPV3 is used as the underlying PSN , VCCV mechanism is needed . The L2TPV3 control connection employs the keep alive mechanism but it is not sufficient for data plane.
L2TPV3 messages are encoded in L2TPV3 session packet.VCCV mechanism is needed for verifying the session state at egress router.
Bidirectional forwarding detection protocol :
=============================================
BFD is a detection protocol is a detection protocol designed to provide fast forwarding path failure detection times for all media types,encapsulation ,topologies and routing protocols.Apart from this , BFD provides consistent failure detection method for network administrators.

L2TP

2009-10-15T18:28:00.003+05:30

L2TP is a protocol that is used to tunnel PPP over public network using IP.L2TP allows the encapsulation of any layer 3 protocol in its packets .The reason is tunnelling is done at layer 2 irrespective of layer 3 protocol.
How L2TP provides security

Like GRE , L2TP depends IPSec or any application layer mechanism to provide the type of security.

Devices in L2TP session :
=========================
PC , L2TP access concentrator (LAC),L2TP network server (LNS)
The PC establishes a connection to a server known as LAC using dial up , POTS and DSL .The LAC initiates L2TP session to LNS .Typically authentication,authorization and accounting of the end user are done on the LNS itself using AAA server or local database.

In running L2TP over IP backbone , UDP is used as carrier of all L2TP traffic which includes the control traffic of session between LAC and LNS.
The initiator of tunnel (LAC) uses UDP port 1701.

Types of L2TP tunnels :
========================
1 . Compulsory tunnelling – client is completely unaware of presence of L2TP connection
L2TP unaware client----------------------- LAC -------------------------LNS
|-----------------PPP-Data-----------------|------------L2TP-Data------------|
2 . voluntary tunnelling - client is aware of L2TP . After the PPP session with LAC , the client sends the L2TP traffic encapsulated in PPP to LNS through LAC. Here client plays the role of LAC.
L2TP Aware Client--------------LAC unware L2TP---------------------------LNS
|-------------PPP-L2TP-Data--------->|--------------L2TP-Data------------>|

Note :
In voluntary tunnelling ,LAC is unware of L2TP .
In compulsory tunnelling , SP must take care of maintenance of LAC devices known as Network Access server (NAS). Compulsory tunnelling hides the details of VPN connectivity from clients and
Effectively transfers the management control over the tunnels from clients to ISP

Two different messages used by L2TP :
=====================================
1 . control messages -L2TP passes control and data messages over separate control and data tunnels.
2 . Data messages – are used to encapsulate the PPP frames that are sent over L2TP tunnels.
L2TP uses registered UDP data prot 1701 . The initiator selects available port no as source port and 1701 as destination port. The initiator sends L2TP packets to establish session . In reply , the destination port is same as source port of packets from initiator.

In CISCO IOS , both source and destination port are set to 1701.
Note : Layer 2 Forwarding (L2F) protocol and L2TP use the same UDP port no. The version field in the header is used to discriminate between L2F and L2TP.
L2F uses 1 and L2TP uses 2.

Fragmentation and GRE tunnels

2009-10-08T15:36:00.005+05:30

Fragmentation and GRE tunnels :
===================================
What is Tunnel :
A tunnel is logical interface that provides a way to encapsulate passenger packet inside a transport protocol.
Tunneling has three memory components .
1 . Passenger protocol (Apple talk , CLNS,IP or IPX,DECNET)
2 . Carrier protocol – one of the encapsulation protocols
GRE ,IP in IP tunnels
3. Transport protocol – The protocol used to carry the encapsulated protocol
Original Packet :
IP-TCP-Telnet
Tunnel Encapsulated packet :
IP-GRE-Original packet
Where IP is transport protocol , GRE is the encapsulation protocol ,
IP is passenger protocol

Example :

Where IP and DECNET are passenger protocols and GRE is carrier protocol.
Why tunnelling is required here ?
Two non IP discontiguous networks are separated by IP network.
The administrator may not want to connect them together by configuring DECNET in the IP network and may not want to permit DECNET routing which affects the performance of the IP network.
So DECNET data is encapsulated into IP at one end and treated as IP packet in IP network.
At the other end , decapsulation is done to get DECNET packets.
Advantages :
1 . Endpoints are using private addresses and backbone does not support the routing of these addresses
2 . Allow VPN across internet
3 .Encapsulating multiple protocols over single protocol backbone
4 . Encrypt traffic over the backbone or internet
Consideration regarding tunnel interfaces :
1 . There are security and topology issues when tunnelling packets. Tunnels can by pass access control list and firewall.if packets are tunnelled through firewall , it is bypassed the firewall for whatever passenger protocol inside ... So it is recommended to enforce the firewall functionality at the end points of tunnels to enforce the any policy on passenger protocols
2 . Tunnels might create problems with the transport protocols that have limited
timers (for eg DECNET) and increase the latency
1 .tunnels across different speed environments introduce packet reordering
problem .Some passenger protocols function poorly in mixed media networks
2 . Routing protocol may prefer tunnels over real link because tunnel might
deceptively appear to be a one hop link with lowest cost and it can be avoided using proper routing configurations
3 . Problems with recursive route can be avoided by configuring proper static routes to tunnels.A recursive route is when the best path to tunnel destination is through Tunnel itself.this situation will cause the tunnel interface to bounce up and down .
Reasons for recursive routing :
1 . Misconfiguration that leads the tunnel to route through same tunnel itself
2 . temporary instability caused by the routing flapping elsewhere in the
network
The router as a PMTUD participant at the endpoint of the tunnel :
The router has two different roles to play when it is at the end point of tunnel.
1 . For PMTUD processing , the router needs to check the DF bit and packet size
of the original data packet and take necessary actions
2 . After the router has encapsulated the packet , the router is acting more like a host with respect to PMTU and in regards to tunnel ip packet
In first role , it checks DF bit , the size of packet . Based on these , it does the fragmentation or dropping
There are 2 ways to do the encapsulation in tunnelling. They are
1 . Encapsulate and fragment
2 . Fragment and Encapsulate
Please note that By default the routers do not do PMTD on the GRE tunnel
packets .It needs to be enabled using the command “tunnel path-mtu-
discovery”
Let us see how fragmentation handled for GRE tunnels

1 . The sender sends 1500 bytes packets (20 bytes IP header + 1480 TCP payload)
2 . Since MTU of the GRE tunnel is 1476 , the packets are divided into 1476 and 44 .
Now they become the 1500 and 68 which include 24 bytes of GRE header.
3 . GRE IP packet s are forwarded to the remote end
4 . Remote end router removes the GRE header and forwards them to host
5 . Reassembly is done at the destination

Fragmentation , MSS , PMTUD

2009-10-04T11:54:00.006+05:30

Fragmentation and TCP MSS , MPTUD :

TCP Maximum segment size and TCP Path MTU Discovery play important role in fragmentation.

Let us see first about TCP MSS

TCP maximum segment size defines maximum amount of data receiver is willing to accept in single TCP/IP stream .

How MSS is calculated for TCP/IP stream :

MSS calculation is done based on buffer size in both server and client side

During TCP/IP connection setup , The SYN segment contains MSS option. If machine does not want to mention MSS , a default of 536 bytes is assumed.

The default value of 536 is derived from 576 minus 40 bytes of header which includes 20 bytes of each ip and tcp header.

Please note that MTU of dial up connection is 576

Diagram :

1 . Host A sends MSS of 16 k

2 S 2 . Server receives SYN and sets send MSS of host A to 16k

3 S 3 . Server sends MSS of 8k

4 H4. Host A receives and sets send MSS of server to 8k.

Here MSS is based on minimum buffer size and MTU.

Sender compares both MSS and MTU and will take lowest value as MSS . MSS is MTU-40 as mentioned earlier.The hosts will compare the MSS size received against MTU and choose lowest.

Here let us take Host A .

Host A has MSS of 16 k and MTU of 1500 . Now it chooses 1500.

When it receives packet from server with MSS of 4422 , again it will compare with MSS and choose lowest.

Please note that TCP MSS takes care of fragmentation at both ends but not at intermediate networks as it does not take care of lower MTU interfaces . So to avoid that , Path MTU discovery was introduced.

Path MTU discovery :

PMTUD is used to dynamically determine the path MTU along the path from source to destination.

If PMTUD is enabled , all TCP/IP packets from the host has DF bit set.

Please note that PMTUD is supported only on TCP and the ip tcp path-mtu-discovery is used to enable PMTUD for TCP connections.

PMTUD is done independently of both directions of TCP flow.In some cases , the trigger of

PMTUD Will lower the MSS of one side and other side keeps the original send MSS because it

never sent an

IP datagram large enough to trigger PMTUD

PMTUD uses the ICMP error message which includes next hop MTU . This next hop MTU

determines MSS of TCP/IP machines.

Scenario 1:

Here client uses default MSS and server uses MSS of 1500. So here Server triggers for PMTUD.

Scenario 2 :

Here packets from client are routed via routers And B . packets from server are routed via Routers C and D.

So there is no need of PMTUD trigger from Client side since it will never receive ICMP error message “Destination unreachable” with code indicating “Fragmentation needed and DF bit set “

Problems with PMTUD

3 things than can break PMTUD and two things are uncommon.

1 A 1 . router can drop packets and does not send ICMP error message (uncommon)

2 A 2 . router can drop and send icmp error message but sender ignores the message(uncommon)

3 A 3 . router can drop and send icmp error message but blocked by firewall or router with acl (Common)

So acl configuration given below needs to be configured.

Access-list 101 permit icmp any any unreachable

Access-list 101 permit icmp any any time-exceeded

Access-list 101 deny icmp any any

Access-list 101 permit ip any any

There are some other techniques that can be used to help alleviate the ICMP blocking .

1 . Clear the DF bit on the router and allow fragmentation .. But fragmentation will trigger some issues

2 . Manipulate the TCP MSS option value using the interface command ip tcp adjust-mss <500-1460>

Fragmentation issues :

1 . Needs more CPU and memory for fragmentation and reassembly

2 . Dropping of single fragment enables retransmission of entire datagram and it leads to fragmentation again

3 . Firewalls may have problem in processing the fragmented packets as firewall uses L4 to L7 details.

If the IP fragments are out of order , a firewall may block the non initial fragments because they do not carry the information that would match the filter.This causes reassembly failed.If firewall is configured to allow non initial packets , attack through non initial packets could occur.

Common network topologies that need PMTUD :

1 . Token ring – where MTU is larger that MTU of Ethernet

2 . PPPoE (ofter used with ADSL) needs 8 bytes for its header . This reduces effective MTU of Ethernet to (1500-8) 1492 bytes

Tunnels like GRE , IPSec and L2TP needs space for their headers and trailer .. so it obviously reduces the effective MTU of the interfaces

What is Tunnel :

A tunnel is logical interface that provides a way to encapsulate passenger packet inside a transport protocol.

IP Fragmentation Reassembly

2009-09-28T20:37:00.006+05:30

IP Fragmentation :

If the MTU of egress network is less than MTU of packet and DF bit is set to 0,

then fragmentation is needed .

Fields involved in fragmentation :

1 . Fragment id

2 . offset

3 . Flag

Fields changed when fragmentation :

Apart from above 3 , following fields are changing in fragmentation

1 . Header length and total length

2 . Header checksum

3 . Options

A Packet Fragmentation Example

If a 2,366 byte packet enters an Ethernet network with a default MTU size,

it must be fragmented into two packets.

The first packet will:

· Be 1,500 bytes in length. 20 bytes will be the IP header, 24 bytes will be the TCP header, and 1,456 bytes will be data.

· Have the DF bit equal to 0 to mean "May Fragment" and the MF bit equal to 1 to mean "More Fragments."

· Have a Fragmentation Offset of 0.

The second packet will:

· Be 910 bytes in length. 20 bytes will be the IP header, 24 bytes will be the TCP header, and 866 bytes will be data.

· Have the DF bit equal to 0 to mean "May Fragment" and the MF bit equal to 0 to mean "Last Fragment."

· Have a Fragmentation Offset of 182 (Note: 182 is 1456 divided by 8

Please note that fragmented packets can also get fragmented.

Reassembly :

Reassembly is done at the destination . The reason is fragmented packets can traverse

independently .

Packet over SONET/SDH

2009-08-29T13:46:00.001+05:30

Packet Over SONET/SDH :
Many of us use the POS interfaces in the routers .
So here we will have one small technical summary about POS(packet over SONET/SDH)
POS also known as PPP over SONET/SDH . This is scheme which uses PPP encapsulation to map IP datagrams into the SONET/SDH payload .

Why SONET/SDH payload and PPP are used for enacapsulation :
Since SONET/SDH is point to point circuit , PPP is well suited here .

POS layers :
There are three pos layers. They are
1 . Bottom layer : mapping into SONET/SDH
2 . Mid layer - Framing of PPP with HDLC .
3 . Top layer – IP encapsulation into PPP
Here I rememember Sridhar G , my mentor in HCL . I created the stream for POS interface and traffic was not success . At that time , he told me to select the HDLC in Agilent RT.
For Ethernet , We use some other encapsulation in RT. Please share that info if u know.

Operation of POS is :

When transmitting : IP -> PPP->FCS generation->Byte stuffing->scrambling-SONET/SDH framing
When receiving : SONET/SDH deframing->descrambling->de stuffing->FCS detection->PPP->IP

Bit stuffing :

In PPP , packet begins and ends with 01111110 .when ever sender sees five consecutives 1s , it automatically inserts the 0 before transmitting . This is stuffing
When ever receiver five consecutive 1s followed by 0 , it destuffs the 0

The main components :

Flag – delimiter of the Frame
Address – Address in PPP packet
Protocol – identifies the network layer protocol
FCS – frame check sequence
Control – Control information in PPP

How Ping and Trace route work

2009-08-28T20:41:00.000+05:30

Trace Route Process

A . Traceroute sends out 3 ICMP echo packets to the named host, but with a TTL of 1; then with a TTL of 2; then with a TTL of 3 and so on. Traceroute will then get 'TTL expired in transit' message back from routers until the desination host computer finally is reached and it responds with the standard ICMP 'echo reply' packet.

Please note that TTL increment happens till destination is reachable or TTL reaches its maximum value.

Trace route is mainly used in troubleshooting of the networks

Possible ICMP error messages in trace route :

H :- Host unreachable. The router has no route to the target system.

N :- Network unreachable.

P :- Protocol unreachable.

S :- Source route failed. You tried to use source routing, but the router is configured to block source-routed packets.

F :- Fragmentation needed. This indicates that the router is misconfigured.

X :- Communication administratively prohibited.

Trace route explained with snippet:

1 51 ms 59 ms 49 ms 10.176.119.1

2 66 ms 50 ms 38 ms 172.31.242.57

3 54 ms 69 ms 60 ms 172.31.78.130

Here value in ms are round trip time of 3 ICMP echo request and ip address is next hop address

The Ping Process

A. The source host generates an ICMP protocol data unit.

B. The ICMP PDU is encapsulated in an IP datagram, with the source and destination IP addresses in the IP header. At this point the datagram is most properly referred to as an ICMP ECHO datagram, but we will call it an IP datagram from here on since that's what it looks like to the networks it is sent over.

C. The source host notes the local time on it's clock as it transmits the IP datagram towards the destination. Each host that receives the IP datagram checks the destination address to see if it matches their own address or is the all hosts address (all 1's in the host field of the IP address).

D. If the destination IP address in the IP datagram does not match the local host's address, the IP datagram is forwarded to the network where the IP address resides.

E. The destination host receives the IP datagram, finds a match between itself and the destination address in the IP datagram.

F. The destination host notes the ICMP ECHO information in the IP datagram, performs any necessary work then destroys the original IP/ICMP ECHO datagram.
G. The destination host creates an ICMP ECHO REPLY, encapsulates it in an IP datagram placing it's own IP address in the source IP address field, and the original sender's IP address in the destination field of the IP datagram.

H. The new IP datagram is routed back to the originator of the PING. The host receives it, notes the time on the clock and finally prints PING output information, including the elapsed time
The process above is repeated until all requested ICMP ECHO packets have been sent and their responses have been received or the default 2-second timeout expired. The default 2-second timout is local to the host initiating the PING and is NOT the Time-To-Live value in the datagram.

OSI Layers

2009-08-28T19:32:00.000+05:30

In the early 1980s, the International Standards Organization (ISO) recognized the need for a standard network model. This would help vendors to create interpretable network devices. The Open Systems Interconnection (OSI) reference model, released in 1984, addressed this need.

The OSI model describes how information makes its way from application programs through a network medium to another application program in another computer. It divides this one big problem into seven smaller problems.

Each of these seven problems is reasonably self-contained and therefore more easily solved without excessive reliance on external information. Each problem is addressed by one of the seven layers of the OSI model.

Layers - Functions - Devices

The application layer

The application layer of the OSI model is the layer that is closest to the user. Instead of providing services to other OSI layers, it provides services to application programs outside the scope of the OSI model. It's services are often part of the application process. Main functions are:-

• identifies and establishes the availability of the intended communication partner.
• synchronizes the sending and receiving applications.
• establishes agreement on procedures for error recovery and control of data integrity.
• determines whether sufficient resources for the intended communications exist.

Devices:-

• Browsers
• Search engines
• E-mail programs
• Newsgroup and chat programs
• Transaction services
• Audio/video conferencing

• Telnet

• SNMP

The presentation layer

It ensures that information sent by the application layer of one system will be readable by the application layer of another system. It provides a common format for transmitting data across various systems, so that data can be understood, regardless of the types of machines involved.
The presentation layer concerns itself not only with the format and representation of actual user data, but also with data structure used by programs. Therefore, the presentation layer negotiates data transfer syntax for the application layer.
Devices:-

• Encryption

• EBCDIC and ASCII

• GIF & JPEG

The Session Layer

The main function of the OSI model's session layer is to control "sessions", which are logical connections between network devices. A session consists of a dialog, or data communications conversation, between two presentation entities. Dialogs can be

• simplex (one-way)
• half-duplex (alternate)
• full-duplex (bi-directional)
Simplex conversations are rare on networks. Half-duplex conversations require a good deal of session layer control, because the start and end of each transmission need to be monitored.
Most networks are of course capable of full-duplex transmission, but in fact many conversations are in practice half-duplex.
Devices:-

Some examples of session layer protocols and interfaces are:

• Network File System (NFS)
• Concurrent database access
• X-Windows System

• Remote Procedure Call (RPC)

• SQL

• NetBIOS Names
• AppleTalk Session Protocol (ASP)
• Digital Network Architecture

The Transport Layer

You can think of the transport layer of the OSI model as a boundary between the upper and lower protocols. The transport layer provides a data transport service that shields the upper layers from transport implementation issues such as the reliability of a connection.
The transport layer provides mechanisms for:-

• multiplexing upper layer applications
• the establishment, maintenance, and orderly termination of virtual circuits
• information flow control
• transport fault detection and recovery

Devices:-

• TCP, UDP, SPX and Sliding Windows.

Multiplexing & De-multiplexing

The transport layer uses a technique called multiplexing to segment and reassemble data from several upper layer applications onto the same transport layer data stream.
When data is being sent, the source machine includes extra bits with the data that encode the message type, originating application, and protocols used.
The destination machine de-multiplexes the data stream, and reassembles the data so that it can be passed up to the destination peer application.

The transport layer data stream provides end-to-end transport services.

It constitutes a logical connection between the end points of an internetwork, that is, the originating host and the destination host.
Before data transfer can begin, both the sending and receiving applications inform their respective operating systems that a connection is going to be initiated.
In essence, one machine places a call that must be accepted by the other.
Protocol software modules in the two operating systems communicate by sending messages across the network to verify that the transfer is authorized and that both sides are ready.
After all the synchronization has occurred, a connection is said to be established and data transfer can begin.

Sequencing - Acknowledgements - Flow Control (Windowing)

During a transfer using TCP, the two machines continue to communicate with their protocol software to verify that data is received correctly. Once data transfer is in progress, congestion can occur for two reasons.
First, the sending device might be able to generate traffic faster than the network can transfer it.
Second, if multiple devices need to send data through the same gateway, or to the same destination, the gateway or destination may experience congestion.
When datagrams arrive too quickly for a device to process, it temporarily stores them in memory and the process being called as buffering. If the datagrams are part of a small burst, this buffering solves the problem.
However, if the traffic continues to arrive at this rate, the device eventually exhausts its memory and must discard additional datagrams that arrive. Instead of losing data, the transport function can issue a "not ready" indicator to the sender. This acts like a stop sign and signals the sender to discontinue sending segment traffic to its peer.
After the receiving device has processed sufficient segments to free space in its buffers, the receiver sends a ready transport indicator - which is like a go signal. When it receives this indicator, the sender can resume segment transmission.

The transport layer may provide a reliable service regardless of the quality of the underlying network. One technique that is used to guarantee reliable delivery is called "positive acknowledgement with retransmission".
This requires the receiver to issue an acknowledgement message to the sender when it receives data. The sending device keeps a record of each packet it sends and it waits for an acknowledgement before sending another packet. The sender also starts a timer when it sends a packet. It retransmits the packet if the timer expires before an acknowledgement is received.

Acknowledging every data segment, however, has its drawbacks. If the sender has to wait for an acknowledgement of each data segment, the throughput will be very low.
A technique called "windowing" is used to increase the throughput. Time is available after the sender finishes transmitting the data segment, but before the sender finishes processing any received acknowledgement. This is used for transmitting more data. The number of data elements the sender is allowed to have outstanding is known as the "window".
For example, with a window size of three the sender can transmit three data segments before expecting an acknowledgement.
In reality, the acknowledgements and data segments will intermix as they communicate across the network. This is known as "piggyback acknowledgement".

The Network Layer

Layer three of the OSI model is the network layer.

• The network layer sends packets from source network to destination network.

• It provides consistent end-to-end packet delivery services to its user, the transport layer.
In wide area networking a substantial geographic distance and many networks can separate two end systems that wish to communicate. Between the two end systems the data may have to be passed through a series of widely distributed intermediary nodes. These intermediary nodes are normally routers.
Routers are special stations on a network, capable of making complex routing decisions.
• The network layer is the domain of routing.
Routing protocols select optimal paths through the series of interconnected networks.
Network layer protocols then move information along these paths.

• One of the functions of the network layer is "path determination".
Path determination enables the router to evaluate all available paths to a destination and determine which to use. It can also establish the preferred way to handle a packet.
After the router determines which path to use it can proceed with switching the packet.
It takes the packet it has accepted on one interface and forwards it to another interface or port that reflects the best path to the packet's destination.

Devices:-

• IP, IPX, Routers, Routing Protocols (RIP, IGRP, OSPF, BGP etc), ARP, RARP, ICMP.

The Data-Link Layer

Layer two of the OSI reference model is the data-link layer. This layer is responsible for providing reliable transit of data across a physical link. The data-link layer is concerned with

• physical addressing; Bridges, Transparent Bridges, Layer 2 Switches
• network topology; CDP
• line discipline (how end systems will use the network link)
• error notification
• ordered delivery of frames
• flow control

• Frame Relay, PPP, SDLC, X.25, 802.3, 802.3, 802.5/Token Ring, FDDI.

At the data-link layer, the bits that come up from the physical layer are formed into data frames, using any of a variety of data-link protocols. Frames consist of fields, containing bits.
The data-link layer is subdivided into two sub layers:

• the logical link control (LLC) sub layer
• the media access control (MAC) sub layer

The LLC sub layer provides support for

• connections between applications running on a LAN
• flow control to the upper layer by means of ready/not ready codes
• sequence control bits.

The LLC sub layer rests on top of other media access protocols to provide interface flexibility.
Because the LLC sub layer operates independently of specific media access protocols, upper layer protocols, for example IP at the network layer, can operate autonomously without concern as to the specific type of LAN media. The LLC sub layer can depend on lower layers to provide access to the media. It provides Service Access Points (SAP's) and flow control. This layer puts 1's & 0's into a logical frame.

The MAC sub layer provides orderly access to the LAN medium. For multiple stations to share the same medium and still uniquely identify each other, the MAC sub layer defines a hardware, or data-link address called the "MAC address". The MAC address is unique for each LAN interface. On most LAN interface cards the MAC address is burned into ROM.
The ROM MAC address is sometimes known as the burned-in address (BIA).

The MAC address is a 48-bit address expressed as 12 hexadecimal digits written in three groups of four digits. The first six hexadecimal digits (the first 24 bits) represent a vendor code known as the organizationally unique identifier (OUI). To ensure vendor uniqueness, the IEEE administers OUIs. The last six hexadecimal digits are administered by the vendor and often represent the interface serial number.

Process of Finding Hosts on the Same Network Segment- ARP

Before a frame is exchanged with a device on the same LAN, the sending device needs to have a MAC address it can use as a destination address.
The sending device may use an address resolution protocol (such as TCP/IP's address resolution protocol (ARP)) to discover the destination's MAC address. In other protocols the MAC address can be determined directly from the network address.
For example, assume that host Y and host Z are on the same LAN. Host Y broadcasts an ARP request onto the LAN looking for host Z. Because it is a broadcast message all devices on the LAN, including host Z, process the request. However, host Z is the only device to respond and it does so with its MAC address. Host Y receives host Z's reply and stores the MAC address in local memory. This is often called an "ARP cache". The next time host Y needs to communicate with host Z it recalls host Z's stored MAC address.

Process of Finding Hosts on the Different Network Segment- ARP + Router

Let's look at how host Y communicates with host X on a different LAN, which it can access via router A.
As before host Y broadcasts its ARP request. Router A, along with all the other devices on the LAN, processes the request. It knows that host X will not see the request because it is on another LAN, and that any packets destined for host X will have to be relayed. So instead, router A provides its own MAC address to host Y as a "proxy" reply to the ARP request. Host Y receives the router's response and saves the MAC address in its ARP cache memory. The next time host Y needs to communicate with host X, it recalls the stored MAC address of router A.

The Physical Layer

Layer one of the OSI model is the physical layer. The physical layer is concerned with the interface to the transmission medium. At the physical layer, data is transmitted onto the medium (e.g. coaxial cable or optical fiber) as a stream of bits.
So, the physical layer is concerned, not with networking protocols, but with the transmission media on the network.
The physical layer defines the electrical, mechanical, procedural, and functional specifications for activating, maintaining, and deactivating the physical link between end systems. This layer puts 1's & 0's onto the wire.
Characteristics specified by the physical layer include

• voltage levels
• timing of voltage changes
• physical data rates
• maximum transmission distances
• physical connectors

Devices:-

• Hubs, FDDI Hardware, Fast Ethernet, Token Ring Hardware.

Example of Layered Transmission

Let's look at the transport layer in TCP/IP as an example.
The transport layer must use the services of the network layer in order to communicate to the peer TCP function on another system. Each lower layer in turn takes upper layer information as part of the PDUs it exchanges with its peer layer.

Each lower layer adds whatever headers and trailers it requires to perform its functions. This is called "data encapsulation".
The transport layer's segments become part of the network layer's "packets" exchanged between IP peers. Network layer packets are also known as "datagrams".
The network layer adds to the start of the PDU, a header to the data that identifies the source and destination logical addresses. These addresses help network devices send the packets across the network along a chosen path.
The Host-to-network layer takes the IP packet and adds a header to form a "frame". The header contains information required to complete the data-link functions. For example, the frame header contains a physical address which allows the network device to communicate over its interface to the next directly connected network device on the link.
Ultimately, these frames must be converted into electrical pulses as the data is finally transmitted by the physical layer protocol across the wire or other physical medium used by the network.