BACHELOR’S GRADUTION PROJECT topic deploying a multi layer network using dynamic routing RIP, OSPF, BGP

- The middle layer is the area of routers in the local area network, routed using OSPF or RIP algorithms.. The routing table contains a route to every destination network that a router k

Trang 1

HANOI UNIVERSITY OF SCIENCE AND TECHNOLOGY

SCHOOL OF ELECTRICAL-ELECTRONICS

BACHELOR’S GRADUTION PROJECT

Topic: Deploying a multi-layer network using dynamic

routing RIP, OSPF, BGP

Trang 2

ASSESSMENT OF THE GRADUATION PROJECT

( Use give lecture tablets direction lead )

Name of Instructors:

Name od student: MSSV: 20182930

Name of project:

Select the appropriate score for students to present according to criteria below:

Very poor (1); Poor (2); Pass (3); Good (4); Very good(5)

There is a combination of theory and practice (20)

State the urgency and importance of the topic, issues and hypotheses

1 (including purpose and relevance) as well as the scope of application of the project.

2 Update the most recent research results (domestic/international)

3 Specify and detail the research/problem solving method

4 Have simulation/experimental results and clearly present the obtained results

Ability to analyze and evaluate results (15)

5 A clear work plan including objectives and implementation methods based on the results of theoretical research

in a systematic way

6 The results are presented in a logical and easy to understand manner, all results are analyzed and evaluated satisfactorily.

In the conclusion, the author points out the differences (if any)

7 between the achieved results and the initial goals set out and provides arguments to suggest possible solutions in the future.

Project Report ‘s technical writing skills (10)

The project presents in accordance with the prescribed form with a

8 logical and beautiful structure of chapters (tables, clear images, with titles)has a chapter introduction and chapter conclusion, a list of referencescitations.

Excellent writing skills (standard sentence structure, scientific style,

9 logical and well-founded reasoning, appropriate vocabulary usage, etc.)

Scientific research achievements (5) (choose 1 out of 3 cases)

1 Having a scientific article published or accepted for

publication/winner of 3rd prize at Institute level or higher/scientific

awards (international/domestic) from 3rd prize or higher/ Having a

registered a patent

1 Reported at the Institute council in the conference of scientific

research students but did not win the 3rd prize or higher / Won the

consolation prize in other national and international competitions on b

the subject such as TI contest.

1

Trang 3

TIEU LUAN MOI download : skknchat123@gmail.com moi nhat

Trang 4

Other comments (about students' attitudes and working spirit)

Date: … / … / 20…

Instructor

Trang 5

Before presenting the content of my project report, I would like to express mysincere thanks to Dr Dang Quang Hieu, who has directly guide and providedocuments for me during the project implementation

Due to time constraints and limited knowledge, the report is not avoid makingsome minor errors Therefore, I look forward to receiving comments from teachersand friends to improve the topic

Trang 6

My name is Le Bao Ngoc, student number 2012930, student of Elitech Program’sElectronic Class 01, course 63 The instructor is Dr Dang Quang Hieu I herebydeclare that all the content presented in the project "Deploying a multi-layernetwork system using dynamic routing of RIP, OSPF, BGP" is the result of myresearch The data stated in the project is completely honest, reflecting thesimulation results achieved All information cited is subject to intellectual propertyregulations; The references are clearly listed I take full responsibility for thecontent written in this project

Hanoi, August 2022

The guarantor

Le Bao Ngoc

Trang 7

TABLE OF CONTENTS

LIST OF SIGNS AND ABBREVIATIONS

PROJECT SUMMARY

1 INTRODUCTION

2 OVERVIEW OF STATIC ROUTE, DYNAMIC ROUTE

2.1 Introduce

2.2 Static routing protocol overview

2.2.1 Static routing operation

2.2.2 Noticeable parameters of configuration

2.3 Dynamic Routing Protocol Overview

3 THEORY

3.1 Autonomous System ( AS )

3.2 RIP

3.2.1 Concept

3.2.2 How it works

3.3 OSPF

3.3.1 Concept

3.3.2 How it works

3.4 BGP

3.4.1 Concept

3.4.2 How it works

3.4.3 Order of precedence in BGP

3.5 Multilayer network system

3.5.1 Network Tier 1

3.5.2 Network Tier 2

3.6 Compare routing protocols OSPF, BGP, BGP

3.7 Advantages and disadvantages of routing protocols OSPF, BGP, BGP

3.8 Load Sharing

3.9 Configuration of RIP, OSPF, BGP

3.9.1 Configuration of RIP

3.9.2 Configuration of OSPF

3.9.3 Configuration of BGP

5

Trang 8

4 LAB MODELS OF RIP, OSPF, BGP

4.1 RIP’s lab model

4.1.1 Process of RIP

4.1.2 Compare RIPv1 and RIPv2

4.2 OSPF’s lab model

4.2.1 Introduce

4.2.2 How OSPF works

4.2.3 OSPF packet types

4.3 BGP’s lab model

4.3.1 Introduction to eBGP and iBGP

4.3.2 Data sheets of BGP

5 Emulate a Cisco router on GNS3

5.1 About GN3

5.2 NS3’s Configuration

5.3 Load IOS for router

5.4 Learn the basic router configurations (how to assign IPs to interfaces, check IP parameters)

6 INSTALLATION INSTALLATION

6 .FIrst Network model

6.2 General settings (using the Linux OS commandline)

6.3 Deploying the top-of-the-line network model

6.3.1 IP Configuration

6.3.2 Config of OSPF routers

6.3.3 Operating the OSPF model on the top layer

6.4 Deploy the middle and lower layer network model

6.4.2 Configuration RIP router

6.4.3 RIP operation on middle and lower layer network model

6.5 Deploy top layer network model

6.5.2 Config for OSPF routers

6.5.3 Operating OSPF on top layer’s network model

6.6 Configure BGP for routers

Trang 9

6.6.1 IP Configuration 73

6.6.2 Operate BGP across the network model 74

7 SYSTEM OPERATION 77

7.1 Check connection by ping and tracepath 77

7.2 OSPF responsiveness test with AS 1 79

7.3 Check the responsiveness of RIP with AS 2 80

7.4 Steps to test the system by installing and connecting to DNS servers 81

8 SIMULATION ON C 82

8.1 Simulation RIP 82

8.2 Simulate OSPF _ 88

8.3 Simulation BGP _ 99

9 EXPANDED : AODV AND OSPF COMPARISON 117

9.1 About AODV 117

9.2 Evaluation of AODV and OSPF when operating in WiMAX 118

CONCLUSION 120

REFERENCES 121

7

Trang 10

LIST OF FIGURE

Hinh̀ 1: Overview Of Routing Error! Bookmark not defined.

Hinh ̀ 2: Overview Of Static Route 122

Hinh ̀ 3: EGP and BGP between ASs 133

Hinh ̀ 4: Internal route and external route 144

Hinh ̀ 5: Network Don’t Use Split Horizon hoặc Poison Reverse 222

Hinh ̀ 6: Count to infinity 233

Hinh ̀ 7: OSPF Areas 266

Hinh ̀ 8: OSPF Network with Headquarters (HQ)- area 0 277

Hinh ̀ 9: Topology diagram of RIP concentrate on R1 and R2 311

Hinh ̀ 10: RIPv2 Model 322

Hinh ̀ 11: Large OSPF networks are decentralized and divided into many areas 333

Hinh ̀ 12: One-zone OSPF model 355

Hinh ̀ 13: Multi-zone OSPF model 355

Hinh ̀ 13: iBGP Peering ‘s Model 366

Hinh ̀ 14: eBGP Peering ‘s Model 377

Hinh ̀ 15: BGP routing information reception and filtering 38

Hinh ̀ 16: Example Of AS-path 39

Hinh ̀ 17: Example Of Next-hop 401

Hinh ̀ 18: Example Of Local Preference 411

Hinh ̀ 19: Example Of MED 422

Hinh ̀ 20: Example Of Weight 425

Hinh ̀ 21: Router Config’s Mode 46

Hinh ̀ 22: Static route’s lab model 49

Hinh ̀ 23: Overview Of Network Model 500

Hinh ̀ 24: Top layer‘s network model 510

Hinh ̀ 25: The middle and lower layer network model (1) 511

Hinh ̀ 26: The middle and lower layer network model (2) 521

Hinh ̀ 27: Throughput vs speed of nodes 11817

Hinh ̀ 28: NRL vs speed of nodes 11917

Hinh ̀ 29: PDR vs speed of nodes 1198

Trang 11

Trang 12

LIST OF SIGNS AND ABBREVIATIONS

Letter write Turn off Cluster are from write full enough

IS -IS Intermediate System To Intermediate System

LSA Link-State Advertisement

IETF Internet Engineering Task Force

AODV Ad Hoc On-Demand Distance Vector Routing

Trang 13

PROJECT SUMMARY

In this project, first of all, I thoroughly understand the theory of routing, then

learn the network model, learn its operating structure, then I deploy multilayer

network using RIP dynamic routing , OSPF, BGP on GNS3 to study the

accuracy and feasibility of deploying the system to operate network models

thanks to the above implementation on Linux OS and Windows After

successfully deploying on GNS3, I tried to optimize the model, routes, IP

configuration for each route and the accuracy when doing work when doing

simulation operations on the network model.

After successful training on GNS3, that multilayer network, I also performed

simulation on C with technical requirements for multilayer network From the

technical requirement, I build a test plan and follow it to verify the design The

design has passed the specification when 100% functional coverage has been

achieved In addition, I also build MANET network model, a single network

model with not too complicated configuration and can directly code separate

functions for each leg of the network and from there compare it with the

multi-layer network system that I have I designed in this project and the feasibility of

using this system in practice.

10

Trang 14

1 INTRODUCTION

In the development of today's life, the Internet has become a tool for us to interact,transact, look up information as well as store data Therefore, understanding the

implementation as well as understanding the algorithms on routers that support

routing is extremely important for future engineers Therefore, in the framework of

this exercise, I would like to show you how to set up and operate a 3-tier network:

- The top layer is the area for the routers of the network operator (ISP), in this area, the routing algorithm used is BGP

- The middle layer is the area of routers in the local area network, routed using OSPF or RIP algorithms

- The bottom layer is the personal computers, directly connected to the Routers

Trang 15

The routing table contains a route to every destination network that a router knowshow to access When you configure interfaces, they are listed as directly connectedinterfaces in the routing table You can manually advertise routes to this table tospecify the destination network However, as the network becomes larger andmore complex, manually configuring every route on each router becomesinfeasible Even if you use default routes and central routers to minimize thenumber of routes that individual routers must know, configuring routes manuallyfor network expansion can be time consuming time Entering static routes is alsoerror prone: it's easy to press the wrong key and enter incorrect routes Instead ofconfiguring static routes, you can use dynamic routing protocols, which allowrouters to exchange routing information with other routers in the network Eachrouter can then use this information to build its routing table.

There are two basic types of routing, Static Route and Dynamic Route

Network administrators when choosing a dynamic routing protocol need toconsider factors such as the size of the network system, the bandwidth of thetransmission lines, and the router's capabilities Router type and router version, theprotocols running in the network

2.2 Static routing protocol overview

For static routing, the route information must be entered by the networkadministrator for the router When the network structure has any changes, thenetwork administrator must delete or add routing information for the router Suchpaths are called fixed paths For a large network, the maintenance of the routernetwork as above takes a lot of time As for the small network operator system,there is little change, this job is less laborious Because static routing requires thenetwork administrator to configure all routing information for the router, there is

no flexibility like dynamic routing In large networks, static routing is oftencombined with dynamic routing protocols for some special purpose

RTZ(config)#ip route 172.24.4.0 255.255.255.0 172.16.1.2

Overview Of Static Route

2.2.1 Static routing operation

12

Trang 16

Static routing can be divided into three specific steps:

 First, the network administrator configures fixed routes for the router

 The router installs these routes into the routing table

 Data packets are routed along these fixed paths

2.2.2 Noticeable parameters of configuration

 Destination-network: The network address to go to

 Subnet-mask: Subnet-mask of Destination-network

 Address: The ip address of the port on the router from which the packet will

go on the interface: the port of the router from which the packet will go

2.3 Dynamic Routing Protocol Overview

Routing protocols differ from routed protocols both in function and in mission

Dynamic routing protocol is used to communicate between routers The dynamicrouting protocol allows this router to share routing information it knows with otherrouters From there, the router can build and maintain its routing table

EGP and BGP between ASs

- A number of dynamic routing protocols: RIP, IGRP, EIGRP, OSPF, are used to directuser data A routing protocol will provide sufficient information about the network layer address so that datapackets can be transmitted from one host to another based on that address structure There are two protocolsthat we need to pay attention to in dynamic routing: Internet Protocol (IP), Internetwork Packet Exchange(IPX)

3 THEORY 3.1 Autonomous System ( AS )

A collection of interconnections of several managed IP networks routed by anadministrative entity Each entity consists of many subunits Each of these units

manages and operates the physical network system independently These networks are

Trang 17

then interconnected and routed according to a common design defined by the entity.

As such, this entire internal system can be thought of as an AS autonomous system

In this system, the network connection configuration and diagram can be clearlydefined On the other hand, obviously, one AS will not be able to grasp the connection

diagram of another AS This resulted in separate routing protocols defined for

implementation in and out of the AS, including:

- Internal Gateway Protocols (IGPs): are protocols that allow routers to route within the

AS In this article, we will use 2 IGP protocols, including Routing Information Protocol (RIP) and OpenShortest Path First (OSPF)

- Exterior Gateway Protocols (EGPs): are protocols that route connections between ASs Inthe article, we use EGP protocol is Border Gateway Protocol (BGP)

Internal route and external route

Typical examples of ASs are ISPs Viettel, VNPT, FPT Telecom, are the ASsthat contribute to the creation of the global Internet A normal business networkcan also become an AS in some special cases, but in most of the cases that I havereferred to, a business network, a home network is not necessary become an AS to

be able to connect to the Internet, but these networks only need to subscribe to acertain ISP to be able to access the Internet

An AS needs to be uniquely identified by a value called the Autonomous SystemNumber (ASN)

ASN has 2 formats: 2-byte or 4-byte

 With the 2-byte range, ASNs range from 0 to 65535.

Trang 18

3.2 RIP 3.2.1 Concept

RIP is an IGP routing protocol used for small ASs, not for large and complexnetworks The routing information protocol is a type of distance-vector routing protocol,

which uses a value to measure that is the number of hops (hop count) in the path from the

source to the destination Each hop in the path from source to destination is considered to

have a value of 1 hop count When a router receives a routing update for a packet, it adds

1 to the measurement and updates the routing table

3.2.2 How it works

RIP uses the Distance Vector Algorithms (DVA) routing algorithm This is aninteroperable routing algorithm that computes the shortest path between pairs of nodes

in the network, based on a centralized method known as the Bellman-Ford algorithm

Network nodes exchange information on the basis of destination address, next node,

and shortest path to the destination

3.2.3 Limit

RIP right judge physical one number magic error _ prize muscle department causeout Head first , in transparent time " holddown " time later when the yes specified

information _ line bag replace change , if the router receives Okay updated

information _ Japan are from a smooth router Neighbors other but this information _

give know yes Street arrive network X with pine number determined line good than

the road that router first there then it will ignore , no access Japan this information _

Next follow to be error count enter term Dinh line repeat yes can happen outwhen the board determined line above routers yet ? Okay access Japanese do so

submit festival capacitor slow

3.3 OSPF 3.3.1 Concept

OSPF is a typical IGP link-state routing protocol This is a protocol widelyused in large enterprise networks The OSPF protocol is standardized for routers to

exchange information and build link state databases OSPF operates in only one AS

region, so it is classed as RIP

Each router running the protocol sends its link states to all routers in the area

After a period of exchange, the routers will identify the link state database table (Link

State Database - LSDB) with each other, each router will have a network map of the

whole area From there, each router will run Dijkstra's algorithm to calculate a shortest

path tree (Shortest Path Tree) and based on this tree to build a routing table

Trang 19

When a router runs OSPF, there must be a unique value that identifies therouter in the community of OSPF routers This value is called Router-id Router - The

id on the router running OSPF in the format of an IP address

By default, the OSPF process on each router will automatically elect the routervalue - id is the highest IP address in the active interfaces, giving priority to the

loopback port To change the router - id of the process, you have to restart the router

or remove the OSPF process and reconfigure, then the router - id election process will

be done again with the existing interfaces on the router

Another way to reset the router-id value is to use the “router-id” command tomanually set this value on the router

Router (config) # router ospf 1

Router (config-router) # router-id ABCD

or set via config file with line

ospf router id ABCD

3.4 BGP 3.4.1 Concept

BGP is an important component of the Internet in routing routers betweendifferent ASs It works based on updating a table containing network addresses

(prefix) indicating the linkage between autonomous systems (autonomous systems), a

collection of network systems under the control of an administrator network, usually

an Internet service provider, ISP In addition to using BGP between ASs, BGP can

also be used in large-scale private networks because OSPF is not available Another

reason is to use BGP to support multihome

Most Internet users do not use BGP directly Only Internet service providersuse BGP to exchange routes BGP is one of the most important protocols for ensuring

the connectivity of the Internet

Routers using BGP connect pairwise with each other by establishing a TCPsession over port 179 This connection is maintained by sending keep- alive 19 bytes

every 60 seconds (default)

There are four types of BGP messages: open (opening session), update(notifying or withdrawing a path), notification (notifying error), keep-alive

(maintaining the connection)

3.4.3 Order of precedence in BGP

 Select the explicit path in the previous table (compared to the default path)

 Choose the path with the highest weight (Cisco router only)

16

Trang 20

 Choose the path with the highest local preference.

 Select the route set by the network administrator himself on the router (static route, with

origin attribute INCOMPLETE)

 Choose the path that goes through the least AS (the shortest AS path)

 Select the inner-origin path first (origin = IGP < EGP)

 Choose the path with the lowest near/far priority MED (Multi exit discriminator)

 Select the external path first

 Choose the path with the lowest IGP metric to the next hop

 Choose the path that exists in the table the longest (oldest one)

 Choose the path to the next router with the lowest BGP ID

3.5 Multilayer network system

Just like apps, the internet is also stratified into tiers Tiers on the bottom layertranslate to the tier above The tiers above pay to receive the services of the tiers

below Currently, the model of the Internet is divided into several layers The tier 1

networks will be on the top tier, providing connectivity for the tier 2 networks on the

bottom tier to connect to each other Tier 1 networks are large companies, as listed in

the table below, while tier 2 networks are national network service providers (ISPs), in

Vietnam such as: VNPT, FPT, Viettel, …

3.5.1 Network Tier 1

In the world, although there are many network providers ( Internet ServiceProvider /ISP), however, to be able to become a global network (Tier 1 network),

currently only a few companies can achieve it This is due to limited financial

potential as well as geo-political factors

of AS Ranking cable length

Deutsche Telekom Global

Carrier

first

Trang 21

Tata Communications India 6453 6 700,000 won

17

Trang 22

Verizon Enterprise Solutions America 701 22 805,000 won

3.5.2 Network Tier 2

The tier 2 and tier 3 are companies that use the cable service provided by Tier 1 toprovide services to the users we call ISPs

3.6 Compare routing protocols OSPF, BGP, BGP

routes

-Number of hops

-Autonomous(AS) throughhave passed

-Weigth

-Prefix length

-Connection and its status: -Routes include+ Connect to the network through packetsConnect to another router passed

-IDs of the routers in the - The set filter internalmulti-access network the set ready

- Aggregate routes within a determined linepredefined network (send the set determined line

ASBR(autonomous.systemborder router), send usingABRs

-External route or defaultroute for external traffic(send using ABRs)

neighbor routers

Trang 23

-Interfaces thatreceive a routebroadcast areinaccessible-Passive interfacereceives theupdate but doesnot send it

In multi-access networks,every router sends LSAs tothe DR and back-upDR(BDR) and receivesLSAs from the DR

-ABRs summarize routesinto stub areas

3.7 Advantages and disadvantages of routing protocols OSPF, BGP, BGP

-RIP v2 can communicate convergence -Simple WANswith external network -Metric is based on hop -Connect to external

-If used to connect to ISP, -Do not use for dial-upISP must redistribute connection

routes into BGP

OSPF -Exact routes taking into - Complex configuration -More extensive LAN

account link speed and -Costs can be high and WAN networks

- Convergence happens an EGP without dial-up connections

-As low as RIP if thenetwork is well designed

BGP -ISPs use BGP - Complex configuration -Connect to ISP

-BGP provides tight -The network must also run -Not to be used overcontrol over which routes IGP dial-up connections

Trang 24

Typically, a routing table can only include one best route per destination Even if arouter learns many equally good routes to the same destination, it must choose one.

Other routes cannot be used unless the selected route fails for some reason

However, when configuring routers, implementing lab models of protocols,building network models can also perform load sharing, allowing it to add multipleroutes to the same destination to its routing table it This option allows the router

to use redundant connections for the same remote site

When you enable load sharing, a router can set up to six routes to the same destination

in its active routing table It can learn these routes from any source, meaning you canenter them manually or the router can learn them using a dynamic routing protocol

However, keep in mind that load sharing allows the router to choose the best multipleroutes Routes must have the same metric and administrative distance; otherwise, onlythe route with the lowest value will be selected Because different routing protocolshave different administrative distances, multiple routes will generally be discoveredusing the same dynamic protocol another route In this case, the traffic may not beproperly balanced across multiple connections, but the more sessions the routersupports, the more balanced the traffic will be routing each time it routes a newpacket to the destination network However, configuring the router for shared load inthis way can cause packets to arrive at their destination out of order and generallyunappreciated

3.9 Configuration of RIP, OSPF, BGP 3.9.1 Configuration of RIP

Before sending a RIP route, the Security Router checks the route's source or next hopaddress If the router is sending an update to a source for a particular route, it willsend an abnormal reverse instead of the normal route Poison Reverse is a route with ametric of 16 (which is infinite for RIP) Poison Reverse distinguishes a legitimatebackup route from one that the local router has received from a neighbor Basically,Poison Reverse notifies the neighbor that it cannot access the network in question

through the local router This mechanism is called " Speeding Convergence: Split Horizon, Poison Reverse, and Triggered Update" Neighbors is listed as the next

address that will change the metric for the route The router then changes the metricfor the route in its own table to a new metric plus a new metric Another neighboradvertises a route with a lower metric The router changes the route to list thisneighbor as the next step address and enters the metric new Router does not receiveroute information for the entire length of the invalidation period Router marks theroute for deletion It sends unique updates to the route in two update cycles updatebefore removing the route completely from its routing table RIP update, v1 and v2RIPupdate packets contain different information, depending on whether the RIP version is

1 or 2 A RIP v1 packet includes: one command field - indicates whether the package

is a request or a reply version field (set at 1), an address family

Trang 25

field - set at 2, for bi Note that addresses in IPv4 format have a maximum of 25 entries, each of which includes:

of which includes:

• Destination IP address

• Subnet mask-provides support for variable length subnets

• A metric-number of hops to the destination address from the next hop address

When a router discovers a new or better route to the destination from a RIP v2packet, it enters the route with the next hop IP address specified in the packet Ifthe IP address field of the next hop is all zeros, the router will assume that thesource of the packet is the IP address of the next hop (This assumption providessome backward compatibility with RIP v1) RIP v1 interfaces broadcast theirrouting updates to the entire subnet RIP v2 routers join the pool for the RIP v2multicast address (224.0.0.9) and multilayer updates to this address Therefore, theRIP v1 and v2 interfaces may not receive each other's updates

*Speeding Convergence: Split Horizon, Poison Reverse, and Triggered Update

One shortcoming of RIP is the relatively slow convergence in some networkenvironments The router sends updates every 30 seconds In a large network, a routermay not receive accurate and up-to-date information about a route for several minutes

Another problem with slow convergence is that it can trigger an infinity of networkcongestion when the connection fails For example, examine the network in thediagram below and consider the updates each router receives for Network 1 when therouters run simple RIP without Split Horizon or Poison Reverse

21

Trang 26

Network Don’t Use Split Horizon hoặc Poison Reverse

Router B is directly connected to Network 1, so it advertises a route to it with index

1 Router A and C receive this route from Router B Both host it route to Network 1 with B beingthe next hop address and metric 2 Routers A and C then start advertising this route Router C receives the routefrom Router A It does not change its routing table to indicate that Router A is the next hop , because metric (2)

is higher than the metric advertised by Router A router B Router B also receives the route from Router A There

is nothing in the update that Router B received from Router A indicating that this route eventually passedthrough Router B itself Router B simply rejected the route for the same reason the reason that Router C did: themetric was higher than the route it took Exempt to be network still is fine determined , too submit this nextcustom smooth share Although of course , the question topic bouncing born if structure bamboo contactconclude replace change

See review thing what will happen out when the contact conclude Between The setdetermined line B and Network 1 no Fort public Router B catches headadvertising fox one line arrive Network 1 with a metric of 16 to only out that it arenot can access access okay

Routers A and C receive Okay copy access Japan this from router B and replacechange the metric, but are not before when the they already to send private routers_ of the me give Network 1 with metric is 2 Router A receives determined linefrom router C and router C receive same route from router A

By Because the line Street this have metric short than _ _ line router line B, router

A and C save store the line Street this in board determined line of the them ( extra

22

Trang 27

one enter metric ) Because conclude connect of the main it with Network 1 noFort public , router B accepts take line road

Split Horizon is one solution to the convergence problem Split Horizon specifies

that an interface must not send updates about a route to the interface it received theroute from In other words, routers assume that the router from which theyreceived a route to an original destination is more directly connected and update on

that destination Split Horizon also minimizes the number of packets sent during

Trang 28

Interval Router Default

Because OSPF routers send each other more messages than RIP routers send,OSPF can consume more bandwidth However, OSPF minimizes the number ofpackets a router has to send in a number of ways In a point-to-point network, onlyneighboring routers fully exchange their databases In a multicast network, onlyone router (DR) floods the LSAs Additionally, OSPF interfaces only send theirown link-state updates instead of sending all routes detected by the protocol, likeRIP interfaces do

LSA

OSPF is a link state protocol; Routers send each other LSAs to distributeinformation about their connections to the network and to other routers LSA helpsrouters synchronize their databases All routers in an AS (or region) must use thesame database to generate correct routes OSPF defines several types of LSAs

Some of these LSAs are flooded to all routers or DRs in an area, and some are sent

to routers throughout the AS Interfaces in the stub do not listen for certain LSAs

OSPF defines specific rules for synchronizing databases with minimal trafficbetween routers Any two routers that have an interface on the same network asneighbors are capable of sending LSAs to each other However, not all neighborsestablish full proximity - that is, exchange LSAs OSPF establishes protocols bywhich all routers can synchronize their databases without exchanging LSAs

Point-to-Point Versus Multi-Access Networks

In a point-to-point network, a router only establishes full contiguity with therouters to which it is directly connected Even Frame Relay networks are based onpermanent point-to-point virtual circuits (PVCs) connected through interfaces onFrame Relay

In a multi-access subnet, such as an Ethernet network, a router can become aneighbor to all other routers on the subnet To minimize OSPF packets, routerschoose one DR and one BDR so that all other routers establish full adjacency That

Trang 29

is, routers only send LSAs to DR and BDR Only DR broadcasts LSA If the DRfails to transmit an LSA within a certain period of time, the BDR assumes it hasfailed and accepts as a new DR.

Areas

One of the most important tasks of an OSPF network administrator is to groupsubnets together into areas so that routers don't need to maintain a large andcomplex database in order to smoothly route traffic to their destination itsdestination An area is a group of subnets in an OSPF network, each of which runsits own copy of OSPF and has its own topology database This means that routers

in separate areas do not need to know each other's topology or exchange LSAs As

a result, database synchronization consumes less bandwidth Less powerful routersand routers that mainly route traffic internally no longer have to keep routingtables wider than they really need to be

-Areas must be identified to:

 All areas connect to the network backbone, or zone 0

 A network backbone consists of routers that interface on multi-site networks, or ABRs

Traffic in OSPF networks is divided into three categories:

outside It then injects the external route, or a default route for external traffic,into the OSPF network The ASBR is usually in the network backbone, but itcan also be in the stub connecting to a remote site When a stub area connects

to a remote site, it is called a nonstub area (NSSA)

25

Trang 30

OSPF Areas

Stub Areas and Stub Routers: The stub network is the network in which the

traffic terminates The network receives traffic destined for its servers, but itdoes not forward any traffic to another network A stub area is an extension ofthe idea of a stub network

Backbone(Area 0): Network's backbone, or area 0, linking all stub areas As

discussed above, it includes ABRs Through exchanges with other ABRs in thebackbone, all ABRs keep a topological database for the entire network Theygenerate route summaries for each non-backbone area They then send theseroute summaries to each other and to the internal routers they serve Obviously,the ABR has to handle more routes than the primitive router andcorrespondingly requires more power

NSSA: NSSA is an area that resembles a stub area in many ways It connects to

the network backbone and usually does not redirect traffic to other areas

However, a router in NSSA also connects to a remote site or an ISP through theASBR Normally, OSPF will not allow external routes to be delivered into thestub However, internal routers within an NSSA may receive LSAs specificallydefined for external routes

Route Computation

Routers use the information they receive from the LSA to assemble the AS'stopology (or, if configured, region) database This database includes:

- Routers belong to separate AS or area

- Networks belonging to a separate AS or area

- Connections belong to a separate AS or area

- Cost for links belonging to a separate AS or area

Trang 31

OSPF Configuration Concerns-Role of each router:

OSPF Network with Headquarters (HQ)- area 0

3.9.3 Configuration of BGP

BGP is an external protocol: it allows different autonomous systems to exchangeroutes BGP is the protocol most ISPs use, and it was designed to allow diverse,sometimes competing organizations to communicate:

27

Trang 32

-BGP can filter both the routes it receives and those it sends by bit length, thusminimizing the number of routes exchanged.

-BGP uses policies to determine the best routes instead of the number of hops perhop, as RIP does, or link states, as OSPF does Autonomous systems can set theirown policies

-The router only communicates with manually configured neighbors

-Configure different policies to exchange routes with different neighborsBGP runs under External BGP (eBGP), which is the protocol used to communicatebetween two autonomous systems, and Internal BGP (iBGP), which is the protocolthat the AS uses to synchronize its own routing tables

Enable BGP

To enable BGP, you must set the local AS number, then enter the context of BGPconfiguration

Local Network Promotion

Specify the local networks that remote sites can access Users should onlyadvertise networks originating from their AS

Set up the router's ID

The BGP interface identifies itself with its neighbors by its router ID Usually this

ID is the IP address of the logical interface connecting to each neighbor It can also

be the address of the repeater interface used as the update source

Configuration for BGP Neighborhood

BGP differs from many routing protocols because it does not allow a router toautomatically search for peers from which to obtain routes You must configure aseparate BGP neighbor for each router with which the local router cancommunicate For each neighbor, you can configure a policy to specify the routesthat the BGP interface sends to and accepts from the neighbor

Set up ID for BGP Neighbor

BGP identifies a peer router by its IP address You set the neighbor's ID whencreating the policy for it

Distinguish Local and Remote AS

The router includes the local AS number in the BGP routes it receives from yourrouter and advertises it to another peer Typically, ISPs forbid their routers toadvertise routes using your AS on the path to outside neighbors The local ASshould be the same number, assigned to you by your ISP, that you configuredwhen BGP was enabled

Trang 33

Load Balancing

Multi-tier BGP routers connect to more than one ISP or more than one ISP router

Such a router can legitimately forward external traffic through more than oneconnection Load balancing ensures that one connection is not used to the exclusion ofanother There are many ways to balance loads, some of which are quite complex andbeyond the scope of this configuration guide This section will only give you somegeneral tips on ways you can try to distribute external traffic via:

-Multiple connections to the same neighbor on the same router-Connect with multiple neighbors on the same router

-Connect to multiple neighbors on multiple routers

Load Balancing on the connections of different neighbors

- Balancing outgoing traffic: In this situation, the BGP route selection algorithm

automatically balances outbound traffic

- Inbound Traffic Balancing: Manually balance incoming traffic by letting the router

advertise certain networks to one neighbor and others to the other neighbor

Prefix configuration example

Router A in AS 1 connects to the Internet It uses a default route for regular Internettraffic, but needs routes to private networks at a remote VPN site Each site in theVPN uses addresses in the 10.1.0.0/16 range To minimize the number of routesrouters have to learn, the organization decided that each site should advertise therange your subnet as a 20-bit network For example, the local site uses subnets in the10.1.0.0/20 range, Site 2 uses the subnets in the 10.1.16.0/20 range, etc

Configure route maps

The route map applied to outgoing data determines how the router advertisesroutes to its neighbors You can configure this route map to perform tasks like:

Define routes the router can advertise to:

• network address or prefix length

• AS that traffic must go through

Create route map entries

You can apply a route map for each neighbor for outgoing data and a map forincoming data You can configure multiple policies in a single route map bycreating entries with the same name but different sequence numbers

29

Trang 34

Configuration for community list

To create a community list, switch to the global configuration mode context Fromthis mode context, you can select one or more clearly defined communityrelationships You can also enter a value for a privately defined community

Configuration for AS path list

You can use the AS path list to select routes for a policy according to the values inthe route's AS field

Define routes that routers can advertise

You can control whether the BGP interface advertises the route to the neighbor byroute:

-Network address-Prefix length-AS path-Community-Metric

Filter incoming routes

We can control the routes that the local router advertises to a neighbor, we can alsocontrol the routes the router accepts from a neighbor You can filter incomingroutes by:

-Destination network address and prefix length-Community

- RIP is the address classless routing protocol RIPv2 has more features as follows:

 There is a verification mechanism between routers when updating to secure the routing table

30

Trang 35

 Support for VLSM(Varibale Length Subnet Masking) masks of different lengths).

- RIP avoids infinite count loop routing by limiting the maximum number of hops allowedfrom the sender to the receiver, the maximum number of hops per path is 15 For the paths the routerreceives from the throughput update of the neighboring router, the router will increase the hop count by 1because the router considers itself a hop in the path If, after increasing the hop index to 1, this index isgreater than 15, the router will consider the destination network not corresponding to this path, so it cannot

be reached In addition, RIP has similar characteristics to other routing protocols : RIP also has a horizonand holddown to avoid updating incorrect routing information

Topology diagram of RIP concentrate on R1 and R2

4.1.2 Compare RIPv1 and RIPv2

Routing by address class Address classless routing

Do not send information about subnet Send information about subnet mask inmask in routing information routing information

Does not support VLSM Therefore all VLSM support Networks in an IPv2networks in a RIPv1 system must have system can have different subnet mask

No mechanism to verify routing There is a mechanism to verify routing

31

Send broadcasts at 255.255.255.255 Sending multicast to address 224.0.0.9

should be more efficient

Trang 36

RIPv2 Model

4.2 OSPF’s lab model 4.2.1 Introduce

OSPF is a link-state routing protocol implemented based on open standards OSPF

is described in many standards of IETF (Internet Engineering Task Force), Openstandard here means OSPF is completely open to the public, no read rights

- Compared with RIPv1 and v2, OSPF is a better IGP internal routing protocol because of itsscalability RIP is limited to 15 hops, converges slowly, and sometimes chooses a slow path because whendeciding to choose, it does not consider other important factors such as bandwidth OSPF overcomes thedisadvantages of RIP and it is a powerful, scalable routing protocol that is suitable for modern networks.OSPF can be configured as a single-zone to use small networks

32

Trang 37

Large OSPF networks are decentralized and divided into many areas

- Large OSPF networks need to use a hierarchical design and divide into multiple zones Theseregions are all connected to the same partition 0, also known as the backbone area This design allows forcontrol over routing updates Such partitioning reduces the load of routing operations, speeds upconvergence, limits the variability of the network to each region, and increases operational efficiency

The following are the features of OSPF:

 It is a link-state routing protocol

 Use the SPF algorithm to calculate the best path

 Update only when the network structure changes

4.2.2 How OSPF works

OSPF collects link state information from neighboring routers Each OSPF routeradvertises the status of its links and forwards the information it receives to allother neighbors

The router processes the information received to build a database of link state in anarea All routers in the same OSPF zone will have this same database Therefore,all routers will have the same information about the state of the links and theneighbors of the other routers Each router applies the SPF algorithm and itsdatabase to calculate the best path for the destination network The SPF algorithmcalculates the cost of the link bandwidth The path with the lowest cost is selectedfor inclusion in the routing table

33

Trang 38

- Each router keeps a list of intimate neighbors, this list is called the intimate neighbordatabase Neighbors that are called intimate are neighbors with which the router has established abidirectional relationship A router can have many neighbors, but not all neighbors have an intimaterelationship For each router the intimate neighbor list will be different.

To reduce the amount of routing information exchanged with many neighboringrouters in the same network, OSPF routers elect a representative router calledDesignate router (DR) and a redundant proxy router called Designated backup

BDR) as the central point for routing information

4.2.3 OSPF packet types

OSPF has 5 types of packets: Hello, Database Description, Link State Request,Link State Update, and Link State Acknowledge

Router IDArea ID

Authentication Data

OSPF’s Packet

- Hello: Hello packet is used to detect and exchange information about neighboring routers

- Database Description: This packet is used to select which router will be authorized to

exchange information first (master/slave)

Link State Request: This packet is used to specify the type of LSA to use during the exchange of DBD packets

- Link State Update: This packet is used to send LSA packets to the adjacent router

requesting this packet when it receives the Request message

- Link State Acknowledge: This packet is used to signal that the Update packet has been received

34

Trang 39

One-zone OSPF model

Multi-zone OSPF model

4.3 BGP’s lab model 4.3.1 Introduction to eBGP and iBGP

Although BGP is designed to run routing between ASs, to run each of these protocols

we still have to configure each specific router In an AS, a suitable number of routerswill be selected to run BGP These routers will shake hands and shake hands withother AS routers to build a network of routers running BGP routing

The handshake operation between routers running BGP is called BGP peering Inthis operation, two routers running BGP together will send each other BGP packets

to build a neighbor relationship; Once the neighbor relationship is successfullybuilt, the two routers can start exchanging routing information with each other

The BGP routing protocol uses TCP as a transport method BGP packets will beencapsulated into TCP segments for exchange between the two routers Therefore, inorder to build a BGP peering between two routers, first, a TCP connection must beestablished between these two routers, the router that initiates the TCP connection willuse a random port greater than or equal to 1024 and the router receives

35

Trang 40

Receiving a connection must open port 179 for TCP We say that BGP runs on TCP,using port 179 Administrators need to understand to properly configure data controldevices (eg Firewall), control devices need to let through flows TCP with port 179.

Another point worth noting when BGP uses TCP for transmission is that TCP doesnot support sending data in groups, so the establishment of neighbor relationship

between two routers completely uses unicast method That is, the administrator It

is mandatory to explicitly declare the IP address of each neighbor that the router isconsidering to establish peering Neighbors in BGP must in principle be declaredmanually BGP does not support automatic multicast neighbor setup as withinternal routing protocols

A BGP neighbor relationship (or BGP peering) can be established between routersbelonging to the same AS or between routers located on two different ASs:

 The first case is called iBGP peering (internal BGP)

 The latter case is called eBGP peering (external BGP).

iBGP Peering ‘s Model

Định dạng
Số trang	125
Dung lượng	4,37 MB