Here are some issues regarding Lab Five: 1. We don't need to snoop any traffic which does not involve the snooping pc in any of the experiments. So, we do not have to worry about the hubs. We can use any hub, the new ones and the old one, for any network. 2. Before starting the experiment, we must disable "checksum-offloading". Checksum-offloading is actually an optimization in network interface cards. When it is enabled, the OS does not calculate the checksum and handover the packet to the interface card. The interface card calculates it just after sending the frame to the wire. Thus, CPU cycles are saved. But ethereal captures sending packets from OS's network stack. So, it gives error message that checksum is incorrect. These error messages will unnecessarily distract and confuse you. In our lab, all the eth1 interfaces (i.e. built-in cards in motherboards) have this functionality. We have to disable this by the following command: ethtool -K eth1 rx off tx off Do it both in pc1 and pc2. You have to issue this command after configuring the interfaces with ifconfig command. Note that if you reboot your pc this option will be reset to enabled as Linux will opt for optimization automatically. Whenever, you reboot or cold boot your pc, you have to do this. We don't have to worry about this issue for eth0 interfaces as they don't have this capability. 3. In experiment 1 while sending UDP packets by ttcp by the following command: ttcp -ts -l1024 -n10 -p4444 -u 10.0.5.22 sometimes packets do not go in the first time. You may find in ethereal that only 2 or 3 UDP packets of 4 bytes payload were sent. In this case, you should immediately issue the command again. Remember to put pc2 in listening mode again before repeating the command in pc1. Whenever this phenomemon occurs, retrying immediately will solve the problem. 4. In the network configuration for experiments 5 to 8, use interface Serial0/0 in the routers. We don't have any Serial1/0 (which is used in the book) interface in our routers. 5. For drawing graphs for experiments 6 to 8 follow the instructions of previous TA provided to you by Dr. Kamil Sarac. All the tools are available in pc1. 6. In experiment 8, use clock rate 9600. In the book, the clock rate is 1000000, but as the note in the book suggests, the rate is optimized for a 10Mbps Ethernet link between router 1 and 2. But all our routers have 100Mbps interface, so setting that speed will not give you adequate time to perform the experiment. After trying with several values, I have come up with the value 9600 which gives sufficient time to perform the experiment. You can lower the clock rate if you need more time but setting it higher may cause you problem. You can see all the available rates by the following commands: configure terminal interface Serial0/0 clock rate ? As soon as you type the "?", all the available rates will be shown to you. 9600 should work fine. Good luck in your experiments! ==================================================================== Issues Related to Lab 6 As we have only one true hub, experiment 3(a) and 3(b) will produce same result as it needs to flood the hubs which requires true hub behavior. There is one printing mistake in 7(a), the netmask should be 255.255.255.0 instead of 255.555.255.0. ==================================================================== Issues Related to Lab 7 1. In fig 7.1 the ip address of Ethernet0/0 of Router2 will be 10.0.1.1, it is a printing mistake. 2. For the first part of the lab, in private network 2, use the big white hub. 3. In part 2, if DHCP server does not reply to PC3's dhcp request, make sure that /vat/lib/dhcp/dhclient.leases exists. If it does not, do the following a. Create an empty file using touch command: touch /vat/lib/dhcp/dhclient.leases b. Restart DHCP server process 4. In step 3 of part 3, before following instructions of (b) run the dhclient on eth1 in PC2 to get ip. Otherwise, you cannot set the default gateway. 5. In step 4 of part 3, add the line "options routers 10.0.1.1" for subnet 10.0.1.0 in DHCP config file. 2. Use "debug ip udp" to display all ip udp traffic on a cisco router (may need to be in IOS config mode). This will help you test whether the router is correctly forwarding DHCP broadcasts. 3. On PC2 in 7.2d.2, make sure you start wireshark on BOTH eth0 and eth1! The book just says "start wireshark". 4. Most importantly, copy /etc/dhcpd.conf verbatim (word-for-word) as described in 7.2b.intro BEFORE making the changes needed for the exercise noted in 7.2b.1. If you do what's natural (i.e. blank the /etc/dhcpd.conf file and retype it from scratch) you will forget to define subnet 10.0.2.0, and this mistake will come back to haunt you later: in part 7.2d omission of the subnet 10.0.2.0 entry will cause the DHCP server to completely ignore eth1 and not send any DHCP Offers in response to DHCP discovers from PC3 (basically, all DHCP Discovers from PC3 will be discarded and the experiment will fail). However this will not affect the outcome of 7.3 because the lab configuration changes slightly. ==================================================================== Issues Related to Lab 8 A. You don't have to bring the downloaded files. They are already in "/root " folder of each pc. 2. Use the white hub for network 10.0.1.0 3. For ethereal display, you may use the filter "icmp || arp || dns" to limit the displayed number of packets. Otherwise there will be too many packets displayed. 4. You must enable domain-lookup in the routers for part 6c. Though, the book says that the routers should have domain-lookup enabled by default, I did not find it in any of the routers of Set 1. You can issue the command right after setting the name-server. The command is: "ip domain-lookup" B. The initial part of the lab requires you to install name server configurations for each PC. According to Professor Sarac's notes, you are not required to download the tar.gz file, these are provided a different way. 1. In the root directory of each PC (not VM) you will find a directory "internetlab-rh90-20030604". Open this directory. 2. Select all items in this directory and copy them. 3. In each PC vmware, open the directory "root" and create a new directory "swinstall" inside. 4. Paste (or SCP) the contents of "internetlab-rh90-20030604" inside of "swinstall". 5. CD into the hosts/ directory and run the auto_conf script. 6. After it completes say NO to the reboot. 7. CD up and into ../labtools/ and run the auto_conf script there. 8. Say YES to reboot. 9. Repeat 4 through 7 on each PC Vmware. 10. After the reboots, you can follow the configuration installation as outlined in the beginning of Lab 8, because there will now be a /root/namedpackage/ directory with the proper files and installation scripts. * Wait until after the VMware reboots before you configure eth0 or eth1.