Hi Guys I have post up our Progress Report that we have worked on today. Please take a look at it when you have time. Thanks for your effort to complete our report today.
Tevita
Thursday 21 June 2012
Final Progress Report
PROGRESS REPORT 2012
Comparing the Convergence Time
Of
Routing Protocols
By: Egious
Tapera
Cyprien Nzohabona
Tevita Vaka
Abstract
It is sometimes a daunting task for non-experienced networking personnel to choose the best routing protocol to use on their networks. Why is it important to choose a routing protocol? Routing protocols can help in deriving the best performances of offered services from our networks in terms of availability and quality of service (QoS).There's need to understand the local network environment and the traffic which will be involved since most organisations are implementing converged networks. There are several routing protocols exist nowadays but in this research will try explore and choosing the best routing protocol between RIPv2, OSPF and EIGRP on the basis of convergence in a network and the choices will be made through experimenting and research. It must also be borne in mind that there has to be some tradeoffs in any choice one makes.This research will try to make it easier in selecting the best routing protocol bearing in mind the short falls of each protocol and the type traffic involved. We will test the protocols in various network designs or scenarios to reveal the performances of the routing protocols. The performance of these routing protocols are analyzed using laboratory network and some simulation tools to enable us to generate realistic data , at the same time collecting realistic measurements. These measurements and data will then be used to provide recommendations.
It is sometimes a daunting task for non-experienced networking personnel to choose the best routing protocol to use on their networks. Why is it important to choose a routing protocol? Routing protocols can help in deriving the best performances of offered services from our networks in terms of availability and quality of service (QoS).There's need to understand the local network environment and the traffic which will be involved since most organisations are implementing converged networks. There are several routing protocols exist nowadays but in this research will try explore and choosing the best routing protocol between RIPv2, OSPF and EIGRP on the basis of convergence in a network and the choices will be made through experimenting and research. It must also be borne in mind that there has to be some tradeoffs in any choice one makes.This research will try to make it easier in selecting the best routing protocol bearing in mind the short falls of each protocol and the type traffic involved. We will test the protocols in various network designs or scenarios to reveal the performances of the routing protocols. The performance of these routing protocols are analyzed using laboratory network and some simulation tools to enable us to generate realistic data , at the same time collecting realistic measurements. These measurements and data will then be used to provide recommendations.
Acknowledgement
We want to
thank Herve' Carpentier our supervisor for his ongoing encouragement and
guidance on technical content during the project and also every members of our
group for all the contribution and hard work that they put in for the progress
of this research.
CONTENT PAGES
ABSTRACT....................................................................................................................
4
ACKNOWLEDGEMENT..................................................................................................
4
INTRODUCTION...........................................................................................................
5
PROCEDURE AND METHODOLOGY..............................................................................
6
BACKGROUND of ROUTING PROTOCOLS......................................................................7
3.1
RIP v2.........................................................................................................7
3.2
EIGRP.........................................................................................................8
3.3
OSPF..........................................................................................................9
EXPERIMENTS/SAMPLE RESULTS................................................................................11
4.1 RIP v2 TEST
RESULTS..................................................................................12
4.2 EIGRP TEST
RESULTS..................................................................................14
4.3 OSPF TEST
RESULTS...................................................................................16
FURTHER WORK
NEEDED...........................................................................................16
CONCLUSION.............................................................................................................20
REFERENCES..............................................................................................................21
1. Introduction
Routing
protocols are very important in our networks since this is the mechanism for
routers or devices in autonomous systems to share routes and update each other
of which routes are operational and those that are not. It simply provides the
forum for devices to share the intelligence of routes and provide network
resilience where alternate routes exist. There are two different
categories of routing protocols which are Interior Gateway Protocols (RIPv2,
EIGRP and OSPF) and Exterior Gateway Protocols (BGP). These routing
protocols use different methods for the selection of best routes to use and
which routes can be used when unexpected failures do occur within the network.
The mechanisms that are used by these routing protocols make them different and
unique.
The goal of
this project to investigate the behaviour of routing convergence time of the
selected protocols and compare them. We are going to carry out an in-depth study
of them so as to be able to compare the protocols and be able to advise which
one is best to use. We are going to assist IT professionals, without
experience, choose the right protocol by comparison and evaluation of the
routing protocols based on performance metrics such as network convergence,
network convergence activity, CPU utilization and bandwidth
utilization. We are going to concentrate this research to the working
of the Interior Gateway Protocols namely RIPv2, EIGRP and OSPF
2. Procedure
and Methodology
Our research
has mainly focused on giving advice on which protocols to use dependent on the
convergence time. This advice does not come easily in an environment where
there are many routing protocols and which have variable, rather different
mechanics of operation. The other aspect to complicate the already
difficult task is that no organisational networks are the same hence are built
for different purposes which is why it becomes crucial to give advice or
recommend certain routing protocols. In order to achieve this we have
looked at designing different test network topologies which we will call
scenarios, and these have been used in the laboratory to test the performance of
the three routing protocols.
The approach taken was that once each scenario was set up each routing protocol was configured one at a time and ensured that there was end to end connectivity between end devices. We then set and varied certain parameters which we 'assumed', from routing protocol theory, would affect the rate of convergence. It was mandatory to run all the three protocols one at time on the same scenario and captured results for analysis.
Our tests were targeted on some critical issues like the available bandwidth, load on interfaces, technology /hardware/equipment type, size of network, and memory size. We believe these have a bearing on the network convergence times.
From the details so far it can be seen that this research is mainly qualitative, we are looking at what makes one routing protocol superior than others, and hence most of our outcomes would be raw data which we can then process. So we have used commands to extract the data software to capture it, which would be the end result of some processes and processing within individual devices in the autonomous system. Convergence time we looking at is the time it takes the devices in network to notice a failure within it , work out an alternative route , relay the information within the network and resume to transmit data using the new route. This is the time we are considering as convergence in our project, when devices have the same information on how to route data to its intended destination.
During the tests we simulated faults on network by either disconnecting or shutting down fast ethernet or serial links.
The comparative analysis has been done in the same network with proposed protocols and performance has been evaluated on the basis of some parameters aimed to figure out the effects of routing protocols.
The approach taken was that once each scenario was set up each routing protocol was configured one at a time and ensured that there was end to end connectivity between end devices. We then set and varied certain parameters which we 'assumed', from routing protocol theory, would affect the rate of convergence. It was mandatory to run all the three protocols one at time on the same scenario and captured results for analysis.
Our tests were targeted on some critical issues like the available bandwidth, load on interfaces, technology /hardware/equipment type, size of network, and memory size. We believe these have a bearing on the network convergence times.
From the details so far it can be seen that this research is mainly qualitative, we are looking at what makes one routing protocol superior than others, and hence most of our outcomes would be raw data which we can then process. So we have used commands to extract the data software to capture it, which would be the end result of some processes and processing within individual devices in the autonomous system. Convergence time we looking at is the time it takes the devices in network to notice a failure within it , work out an alternative route , relay the information within the network and resume to transmit data using the new route. This is the time we are considering as convergence in our project, when devices have the same information on how to route data to its intended destination.
During the tests we simulated faults on network by either disconnecting or shutting down fast ethernet or serial links.
The comparative analysis has been done in the same network with proposed protocols and performance has been evaluated on the basis of some parameters aimed to figure out the effects of routing protocols.
3. Background of Routing Protocols
We cannot give good reason for the adaptation of these protocols without understanding the complete autonomous systems. It is therefore, of paramount importance to have an understanding of the routing protocols this research is concentrating on namely RIPv2, EIGRP and OSPF. This background will help up us predict the behaviours of the protocols and assist us in the comparisons and analysis of the convergence times.
3.1 RIPv2
Routing Information Protocol version 2
(RIPv2) is an improved version of its earlier version and both of them are a
distance vector protocols. This means that they make use of direction and
distance to a required destination within the network. It is based on the
Bellman-Ford algorithm and it only uses a single routing metric, which is the
hop count, to measure the distance between the source and destination network.
Each router tells its neighbour what it knows, that is, the networks directly
connected to it and the neighbour passes the information to its neighbour. It
is generally referred to as routing by "gossip". The hops are
incremented each time is passes a node or router to ascertain the number of
hops to reach that particular destination. 15 hops is the maximum number of
hops which is allowed between source and destination network above
that the network is unreachable.
To maintain the network within the autonomous system RIPv2 makes use of timers
for the devices to inform each other of routes still operational and those that
have failed. Those that are no longer operational are removed from the
routing table. The default timers are listed bellow these can be changed to meet
certain operation standards as explained later in the research.
The default settings of the four timers are:
- 30s (update timer), Initiates the sending of the routing table
- 180s (invalid timer), is set each time a routing table is received and failure to receive will start this timer and its expiry means the route is unusable.
- 180s (hold-down timer) is initiated when an invalid timer expires and it keeps route for the duration of the timer in case route comes back.
- 240s (flush timer) Route will be removed after the expiry of the flush time
It is
important to understand these mechanics for the effective investigation of how
these affect convergence time and how to make the best of them. We will explore
these factors in our research.
3.2 EIGRP
Enhanced Interior Gateway Routing Protocol (EIGRP) is a Cisco proprietary routing protocol which is mainly supported on Cisco devices. When a network has a mixture of devices other routing protocols can be used or a mixture depending on the network setup. Since the devices we using are all Cisco devices we will be investigating the performance of EIGRP routing protocol for the comparisons.
EIGRP is also known as a Hybrid routing protocol since it combines the best features of Distance -Vector and Link-State routing protocols. It uses Diffusing Update Algorithm (DUAL) for the calculation on the best routes to use within the autonomous system. It uses bandwidth and delay for the calculation of its metric and uses hello packets to monitor and maintain network connections. This protocol stores its data in three main tables namely:-
Enhanced Interior Gateway Routing Protocol (EIGRP) is a Cisco proprietary routing protocol which is mainly supported on Cisco devices. When a network has a mixture of devices other routing protocols can be used or a mixture depending on the network setup. Since the devices we using are all Cisco devices we will be investigating the performance of EIGRP routing protocol for the comparisons.
EIGRP is also known as a Hybrid routing protocol since it combines the best features of Distance -Vector and Link-State routing protocols. It uses Diffusing Update Algorithm (DUAL) for the calculation on the best routes to use within the autonomous system. It uses bandwidth and delay for the calculation of its metric and uses hello packets to monitor and maintain network connections. This protocol stores its data in three main tables namely:-
- Interface table- stores all the interfaces that have been configured to process EIGRP
- Neighbour table- keeps track of all live neighbors and any new neighbors
- Topology table- keeps track of all advertised routes within and without the autonomous system, routes with low metrics are then offered to the IP routing table.
Through the use of DUAL EIGRP keeps backup routes which offer
alternative path to destination networks. This gives it an edge when a link
fails if a feasible successor exists the protocol has no need to recalculate
but offers that as a usable route ,it saves time. We will explore more to see
some of these routing protocol features at work in our experiments
Link-State algorithm is when routers do not send every router its routing table but sends information about the links they have established or lost to other routers.
Link-State algorithm is when routers do not send every router its routing table but sends information about the links they have established or lost to other routers.
- Non-Periodic - on change only
- Partial - only relevant changes
- Bounded - sent to effecting neighbors
3.3 OSPF
Open Shortest Path First (OSPF) is a link
state routing protocol that uses a Shortest Path First (SPF), which
calculates the least cost, in combination with the Djikstras algorithm to
calculate shortest best routes to destinations. OSPF uses cost as its metric
which means in its calculation it selects the least cost path to a destination
node. OSPF is a link state routing protocol that is used to distribute
information within a single Autonomous System. Link state routing protocol
constructs a tree structure of the network which fully describes all possible
routes together with their costs. This means that each router has full
knowledge of the whole network in its autonomous system and selects its routes
it will use to reach various networks within its domain.
All routers need to have the following for neighbor adjacency to form
- Area ID
- Same Subnet
- Authentication (if used)
- Hello Interval timer
- Dead Interval timer
- Area type (Stub, NSSA)
- Router ID must be unique
4. EXPERIMENT AND SAMPLE RESULTS
- We have used the
same Topology on all the 3 routing Protocols as we needed to compare them after
we get the final results.
EQUIPMENTS USED
- 6 Cisco Routers - 3 Layer 3 Switches
- 3 Layer 2
Switches - 6 PC (Host
connected to different Vlans)
4.1 RIP v2 TEST RESULTS
This protocol uses
hop count and the routes with less hops
are selected and used for transmitting data. Hence can choose some suboptimal
routes to destinations
|
R1#sh ip route
C 172.16.12.0 is directly
connected, FastEthernet0/0
R 172.16.15.0 [120/2] via 172.16.12.2,
00:00:22, FastEthernet0/0
R 172.16.10.0 [120/2] via 172.16.12.2,
00:00:22, FastEthernet0/0
10.0.0.0/24 is subnetted,
6 subnets
R 10.1.10.0 [120/1] via 172.16.12.2,
00:00:22, FastEthernet0/0
|
The highlighted
areas show the hops to the relevant networks.
|
DLS0# ping 172.30.6.2 repeat 200
Type escape sequence to abort.
Sending 200, 100-byte ICMP Echos to 172.30.6.2, timeout is 2 seconds: Packet sent with a source address of 192.168.101.2 !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! !!!!!!!!!!!!!!!!!!!!!!!!!U!.!U!.!U!.!U!.!U!.!U!.!U!.!U!.!U!.!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! Success rate is 91 percent (182/200), round-trip min/avg/max = 17/166/219 ms |
We had 182 successful pings and 18 pings packet are lost; it can be translated into
thirty-six seconds’ traffic lost because the timeout of each ping packet
is two seconds. So this can be attributed to 36 seconds for the network
to converge and begin re-transmitting the pings again.
The above is output is from Wire shark
The yellow
highlighted area reveals the time taken from the moment of link failure to when
the devices re-start sending data/packets.
This can be the time taken for the router to converged.
4.2 EIGRP TEST RESULTS
We noted a
difference in behaviour when we had the following outputs configured on our routers,
we were expecting EIGRP to choose the
route from R7àR6 àR5 because it had higher bandwidth than R7àR6àR4. On
convergence the lesser bandwidth route was still being selected, possibly
because it had less delay and hops. Even increasing the bandwidth to 3 times of
the link, EIGRP kept using the shorter route. It revealed to us that the difference
should be really high if the route has an additional hop. More has to be
investigated on this issue to determine the exact cause.
R5 to R6
R5 output
|
R5#sh int
s0/2/0
Serial0/2/0
is up, line protocol is up
Hardware is GT96K Serial
Internet address is 172.16.1.1/30
MTU 1500 bytes, BW 5000 Kbit, DLY 20000 usec,
reliability 255/255, txload 1/255, rxload
1/255
<output
omitted>
Available Bandwidth 3750kilobits/sec
5 minute
input rate 0 bits/sec, 0 packets/sec
|
R4 to R6
R4 OUTPUT
|
Serial0/0/1 is up, line protocol is up
Hardware is GT96K Serial
Internet address is 172.16.2.1/30
MTU 1500 bytes, BW 1544Kbit, DLY 20000 usec,
reliability 255/255, txload 1/255, rxload
1/255
<output omitted>
Available Bandwidth 1158kilobits/sec
5 minute input rate 0 bits/sec, 0
packets/sec
|
|
10000 pings DLS2 to Host connected to ALS1
|
10000 pings DLS2 to Host connected to ALS1
(port-channel disabled)
|
|
Time:
* First Packet: 2012-06-16 21:54:10
*Last Packet: 2012-06-16 21:54:48
* Elapsed: 00:00:38
Traffic captured:
* Packets: 20032
* Between first and last packet:
38.007 sec
* Avg packets/sec: 527.059
|
Time:
* First Packet: 2012-06-16 22:05:58
* Last Packet: 2012-06-16 22:22:29
* Elapsed: 00:16:30
Traffic captured:
* Packets: 20910
* Between first and last packet:
990.154
* Avg packets/sec: 21.118
|
Wire Shark Statistics Summary Brief
From these
statistics, we can notice that the ether channel link is much faster than the
redundant link. It is able to process 527 packets per second when the redundant
link need the same time to process only 21 packets.
Note that the
redundant link here is R7 to R6 to R4 to R2 to R1 to DLS1
4.3 OSPF TEST RESULTS
IMPACT OF BANDWIDTH ON OSPF
·
DLS1 -->
R1 --> R2-->
R4--> R6 -->
R7 --> DLS2 -->
PC2 (172.30.20.20)
|
Link
|
Bandwidth (kb/s)
|
|
R4 à R6
|
1544
|
|
R4 à R5
|
1544
|
|
R5 à R6
|
128
|
-the above output shows the route from PC1 to PC2 (with Etherchannel
Disable), and because the link from R4 à R6 have about the same bandwidth as from R4 à R5 OSPF will
choose R4 à R6 because of its shorter path then the other.
·
DLS1 -->R1 -->
R2 -->
R4 -->
R5 -->
R6 -->
R7 -->DLS2 -->
PC2 (172.30.20.20)
|
Link
|
Bandwidth (kb/s)
|
|
R4 à R6
|
64
|
|
R4 à R5
|
1544
|
|
R5 à R6
|
128
|
- In this second
output we made a change to the bandwidth
from from R4 à R6 to 64kb/s, at this point, the traffic will
go through the longer path or more hop count because of its higher bandwidth
|
DLS2#ping 192.168.10.10 repeat 500
Type escape sequence to abort.
Sending 500, 100-byte ICMP Echos to
192.168.10.10, timeout is 2 seconds:
!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
!!!!!!!!!!!!!!!!!!!!! <output
ommited>!!!!!!!!!!!!!!!!!!!!!!!!
!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
Success rate is 100 percent (500/500), round-trip
min/avg/max = 109/112/135 ms
|
IMPACT OF LINK FAILURE ON OSPF CONVERGENCE
·
DLS2 -->
R7 --> R6 -->
R4 --> R2 -->
R1 --> DLS1 -->
PC1 (192.168.10.10)
-
First part of the test we send 500 ping from PC2 à PC1 through the path above with every link
enables, as it shows a 100% success 500/500
|
DLS2#ping 192.168.10.10 repeat 500 (R6 S0/2/1 disabled)
Type escape sequence to abort.
Sending 500, 100-byte ICMP Echos to
192.168.10.10, timeout is 2 seconds:
!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
!!!!!! <output
ommited>!!!!!!!!!!!!..U..U..!!!!!!!!!!!!!!!!!!!
!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
Success rate is 98 percent (494/500), round-trip
min/avg/max = 109/133/160 ms
|
·
DLS2 à
R7 à
R6 à
R5 à
R4 à
R2 à
R1 à
DLS1 à
PC1 (192.168.10.10)
- Second part of the test we disconnect the link from R6 à R4, so OSPF
will recalculate the new routes and send the packets through. In the above
output shows “!” for every successful packet delivered and “...” for packets
failed. It shows the final output of 494/500 so it missed only six packets.
-
6 (packet lost) x 2 (timeout) = 12 seconds
(convergence time)
|
DLS2#ping 192.168.10.10 repeat 2000
Type escape sequence to abort.
Sending 2000, 100-byte ICMP Echos to
192.168.10.10, timeout is 2 seconds:
!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
<output ommited>
!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
Success rate is 100 percent (2000/2000),
round-trip min/avg/max = 1/2/9 ms
|
·
DLS2 à
DLS1 à
PC1 (192.168.10.10)
|
DLS2#ping 192.168.10.10
repeat 2000 (Port Channel disabled)
Sending 2000, 100-byte ICMP Echos to 192.168.10.10, timeout is 2
seconds:
!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
<Output Ommited>
!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
*Mar 1 02:50:46.961:
%LINEPROTO-5-UPDOWN: Line protocol on Interface FastEthernet0/4, changed
state to down
*Mar 1 02:50:47.003:
%LINEPROTO-5-UPDOWN: Line protocol on Interface FastEthernet0/3, changed
state to down
<Output Ommited>
*Mar 1 02:50:47.011:
%LINEPROTO-5-UPDOWN: Line protocol on Interface Port-channel2, changed
state to down
!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! U.U.!!!!!!!!!!!
Success rate is 99 percent (1996/2000), round-trip min/avg/max = 1/44/143
ms
|
·
DLS2 -->
R7 --> R6 -->
R4 --> R2 -->
R1 --> DLS1 -->
PC1 (192.168.10.10)
-
the last test we disable the etherchannel again to see if there will be a difference
as the first test using the Fast Ethernet link with 500 ping. The result shows
not much of a difference as it lost only 4 packets out of 2000.
-
4(packet lost) x 2 (timeout) = 8 seconds
(convergence time)
IMPACT OF COST ON OSPF
|
2000 pings with R6àR5 shutdown;.
R6 à R4 band = 64 kbps
|
2000 ping with R6à R4 Cost set to
1000
|
2000 ping with R6àR4 Cost = 1000;
R5-R6 Cost = 500; R4-R5 Cost = 500
|
|
Time:
* First Packet: 20:14:21
*Last Packet: 20:19:18
* Elapsed: 00:04:57
Traffic captured:
* Packets:
3865
* Between first and last packet: 297.022 sec
* Avg packets/sec: 13.013
|
Time:
* First Packet: 21:26:50
*Last Packet: 21:31:56
* Elapsed: 00:05:05
Traffic captured:
* Packets: 4249
* Between first and last packet: 305.468 sec
* Avg packets/sec: 13.910
|
Time:
* First Packet: 21:53:19
*Last Packet: 21:58:53
* Elapsed: 00:05:34
Traffic captured:
* Packets: 4270
* Between first and last packet: 334.080 sec
* Avg packets/sec: 12.781
|
Wire
Shark Statistics Summary Brief
5. Further Work Needed
From the
experience we have found out that we have some additional work which needs to
be carried out in our research project
-
We
need to do more testing before make a final decision eg: run every test for
about 10 times and get the average out of it as a final test result. Since our
project is more qualitative in nature and if we are to draw conclusions from
the gathered raw data.
-
Another
point we have noted is that our first topologies were too basic for the
comparison tests we are carrying out, there’s need to challenge or drive the
protocols to their limits. There might be need to have other scenarios with
several nodes.
-
We
had a lot of results but were rendered useless due to the failure t o link them
to specific scenario and tests . So
there’s need to be specific on all captured results , specifically naming them
as for them to be valid.
-
We
tried to use the TGN router for our traffic generation, but met some issues ,
these have to be resolved. We are also looking at traffic generation software.
-
Lastly
, all pending issues on our original proposal have to be resolved before the
completion of the project.
6. Conclusion
Routing tables across the entire network should
converged in minimum time in order to avoid excess traffic loss and that is the
main interest of this project. By the experiment that we did so far, we
investigated routing convergence under difference situation like disconnecting
or disabling a link and also playing around with the Bandwidth to see the
effect on convergence time. In concluded that OSPF and EIGRP have converged
faster than RIP by the results that we pull out from the test that we have.
Although there will be more testing later on for
more result and to finalise our final conclusion for our research.
7. References
1.
Implementing
Cisco IP Routing (ROUTE), First Edition June 2010, Cisco Press
Diane Teare
2.
CCIE
Professional Development; Routing TCP/IP volume 1, 11th Edition
1998, Cisco Press
Jeff Doyle
3.
Networking
Explained, 2nd Edition 2002, Digital Press
William M Hancock PH.D
Michael A. Gallo PH.D
4.
Implementing
Cisco Switched Networks (SWITCH), First Edition June 2010, Cisco Press
Richard Froom, CCIE No.5102
Balaji Sivasubramanian
Erum Frahim, CCIE No.7549
5.
Troubleshooting
and Maintaining Cisco IP Networks (TSHOOT), First Edition March 2010, Cisco
Press
Amir Ranjbar, CCIE No.8669