Network jitter, or jitter, is the name given to a variation in the time delay between packet arrival. This concept is better explained using the illustration below.

In the above illustration, the gaps between each of the packets represent the time it takes for each packet to reach the destination. Jitter is shown by the uneven gaps between packets two and three, as well as three and four.

In a perfect world, every packet should arrive at a set amount of time after the preceding packet. In reality, there are many factors to take into consideration when jitter is experienced. Sometimes the cause of jitter is beyond your control, since issues may arise outside your network.

The effect Jitter has on network applications can vary. It is unlikely for a user surfing the internet to report a problem that is a result of jitter. Other real time services, such as VOIP, can experience serious problems related to jitter. Lucky for us, many vendors (including Cisco) build provisions into their routers that can compensate for jitter.

On Cisco routers, the playout delay buffer (PDB) is the mechanism that is used to compensate for jitter. The PDB stores the incoming packets and then sends them to the next destination as a steady stream. This buffering process is similar to that used with other real-time protocols such as those used for audio and video. The ultimate goal of the buffer is to negate any jitter by relaying the packets as a steady stream. An illustration of the packets before and after the playout delay buffer does its job is shown below.

The buffer can only compensate for packets that are delayed within a specified range. If packets start to arrive outside of the working range of the buffer, those packets are dropped. With VOIP a dropped packet can mean the loss of some of the audio, which can make part of a conversation seem choppy.

In a Cisco router, the playout delay buffer sends a steady stream of packets to the digital signal processors. The main job of the DSP is to convert the audio from digital to analog. A secondary function of the DSP is to compensate for missing packets. If a packet is missing, the DSP can make an educated guess as to the contents of the missing packet and insert that missing piece into the audio stream. The result is that the end user never hears a difference. The DSP can only compensate for a finite amount of dropped packets before the end users start to notice an effect on call quality.

Jitter can present some interesting problems, particularly when you are dealing with real-time services. Cisco has built some provisions into their routers to help counter some of the effects of jitter.

Network administrators are like software developers. That is documentation is usually an afterthought, and sometimes it is not a thought at all! We have all faced this problem. This is the situation that I walked into this past summer. The key here is not to look at the situation as a problem, but as a challenge.  Our largest building was our biggest challenge. The cabling was not done in a structured manner, and in some cases I was baffled how some of the “redundant links” did not cause network loops since STP was not activated.

The biggest challenge in this building was troubleshooting link problems between the switch closets. Dropping the link from one cabinet to another would cause half the building to go down in some cases. Without proper network documentation, we were left in the dark. With the way the cabling between the switches was laid out, it would be a copious amount of work document this by hand. Enter Solarwinds with their LANsurveyor product.

LANsurveyor can be used to create network maps with a minimal amount of information. In my case, I was able to create a network map for each of our seven buildings using nothing more then the IP address range for that building and our SNMP community strings. Even if your community strings are not standardized, changing them on your devices would be considerably less work than creating a manual network map.

This convenience does come at a price. The version I use sells for about $2000 and is a standalone piece of software. There is a $500 version that integrates with Visio, but I have not used that so I cannot comment. You do have the option of downloading a trial before you purchase the software. If $2000 is not in your personal budget, you may be able to sell the software to your boss if you calculate all of the man hours it would take to map your network by hand!

Below are some screen shots of LANsurveyor, one of which shows the chaotic network map for the building I mentioned above.

LANsurveyor network discovery settings:

Sample Network Map (I zoomed out so you can see how difficult this would have been to document by other means):

Note: This is not an advertisement for Solarwinds, and thus I am not profiting from this post! If you have used other software to solve a similar problem, please share your experiences in the comments section.

Hyper-V has a few requirements to get everything up and running. First, you need to install Windows Server 2008 x64 (64-Bit). Other versions will not satisfy the Hyper-V requirements. The second requirement is probably going to be the biggest problem for people trying to run older hardware.

Hyper-V requires a processor that supports hardware assisted virtualization. AMD calls this AMD-V, while Intel calls it VT. Some older processors may support this with an update to the motherboard BIOS, others will not. I found that my 3-4 year old Opteron 165 did not support hardware assisted Virtualization. Some research will be required to see if your specific hardware is supported.

Unfortunately all of this caused many things to fall by the wayside. This site and my CCNA study where two major things that I fell behind with. The past few weeks I have really stepped up my CCNA study. The down side of that is that I have had to review many old topics and my books are now a version behind the current CCNA tests.

I am going to try to update this site more often. One change I am going to make is the renaming of the “VMware Whitebox” section. I have not given up on that, but I have decided to play with some ofther virtualization options. That means I will be setting up and playing with Windows Sever 2008 with Hyper-V! I am hoping to get this setup soon and share my setup here.

Note: All commands should be issued in global configuration mode unless otherwise noted.

The first step is to enable the web (http) server on the router:

ip http server

Next, you need to tell the web server where to find the HTML files used by the GUI:

ip http path flash:

Finally, you need to tell the web server which form of authentication to use:

ip http authentication {aaa | enable | local | tacacs}

Below is a quick run down of the authentication methods.

• aaa - Use aaa login service.
• enable - Uses the enable password that is set on the router (This is the default authentication method).
• local - Uses a local username and password that is set on the router using the username command.
• tacacs - Uses a TACACS server.

Before you can access the CME GUI, you need to set an initial username and password for the administrator. The following commands will allow you to do this.

Enter telephony service configuration mode:

telephony-service

Set the GUI administrator username and password:

web admin system name username {password string | secret {0 | 5} string}

With the last command it is suggested that you use the secret 5 option, since it will encrypt the password using an MD5 hash.

Once you have completed all of these steps you should be able to access the CME GUI from a web broswer by navigating to http://xxx.xxx.xxx.xxx/ccme.html (relace xxx.xxx.xxx.xxx with the IP address of your router). You will be prompted for a username and password, which should be the one you just set with the web admin command.

Note: I had some problems with the GUI when viewing it with Firefox (2.0.0.14). It’s suggested that you used Internet Explorer.

This only covers a very small portion of CME configuration. I suggest that you consult the Cisco Unified Communications Manager Express Administrator’s Guide. It is very detailed and worth the time spent reading though it.

delete flash:/file.bin

After doing that I was surprised to see that the files were only marked for deletion in flash, which meant they were also still occupying the storage space I needed for other files. In order to permanently delete the files, the following command needs to be issued:

squeeze flash:

After trying to run the squeeze command, I received an error saying no squeeze logs could be found and the operation failed. After some research I discovered that you need to run squeeze on the flash memory while it is empty (I later found you can do this with just the IOS image in flash). Here is the process I followed.

1. Make sure you have a backup copy of the IOS image you want on the router.
2. Format flash (Do not reboot the router after this)

   erase flash

3. Reload the IOS image into flash.

   copy tftp://xxx.xxx.xxx.xxx/ios_image.bin flash:/

4. Run squeeze on flash

   squeeze flash:

After all of this is completed, you will be able to mark files for deletion using the delete command and then permanently delete them with the squeeze command.

Right now, I have only done a very basic configuration of CME. I do not have any IP Phones connected or anything. There will be further posts that will cover more advanced configuration topics.

There are a few things that you need in order to run CME. They are as follows:

1. An IOS version that has CME support (Use the Cisco Feature Navigator to see if your IOS supports CME).
2. CME basic files. This is related to the IOS version you have. The Cisco download page will tell you which CME basic file you should download for your version of IOS.
3. CME GUI files

Before proceeding, consult the CIsco CME Matrix. You can look at the version of CME you wish to run and the IOS version you wish to run it on. The matrix will tell you if it is possible to run that version of CME with the version of IOS you have. If you click on the specifications link for each version you will find the the minimum required DRAM and Flash needed to run CME.

The first step is to upgrade your IOS, if you need to. After that, you are ready to start uploading the CME files to Flash. This can be a bit tedious since there are so many files. After some suggestions from a member on Networking-Forums, I have found a process that saves you a bit of time.

1. Extract both the CME basic and CME GUI archives into a single folder on your computer.
2. Repackage the extracted files (except .tar files) into a new tar file (7-Zip is free and can create tar files in Windows). It is very important not to package other .tar files in this new archive, since tar files need to be individually uploaded and extracted on the router.
3. Use the following command to upload and extract each tar files you need (I usually start with the tar file I just created):

   archive tar /xtract tftp://xxx.xxx.xxx.xxx/file.tar flash:/

You should repeat the last step for each tar file you need. Some files you may not need, such as those for phones you will not be using, or one for the ACD feature.

That should be the basics of getting the needed files onto your router. In the second part I will detail how to enable the GUI interface and setup authentication for the interface.

I started getting the following message:

*Mar  1 00:16:35.203: %LINEPROTO-5-UPDOWN: Line protocol on Interface Serial0, changed state to down

Which a few seconds later would be followed by this message:

*Mar  1 00:16:35.203: %LINEPROTO-5-UPDOWN: Line protocol on Interface Serial0, changed state to up

I went through everything I could think of, and even cleared each routers config file multiple times. Baffled, and a bit frustrated, I walked away for a bit. When I came back, it dawned on me that I never checked the cables I was using (I had swamped them, but I did not check them).

Sure enough, a few pins pins on the serial cables that connected to my third 2513 router were bent. I examined the serial ports on that router and discovered that there were pins jammed in the connector. I can honestly say this was not caused by me, since I only have one set of Serial cables, and all of the pins on those were intact (but bent at this point).

So my problems are solved, for now. I will attempt to remove the pins form the router, but if that fails it may get replaced with another 2500 or more likely a 2600XM router.

ip host 2513-01 2001 172.21.1.1

The “2513-01″ is the host name of the router. The “2001″ tells the terminal server what RJ-45 connector is attached to the devices console port. In this case, the device is attached to connector one. If it was attached to connector 11 the port number would be 2011, and so on. Each device will have the same IP address associated with it. This IP address is associated with a looback interface using the following commands:

interface Loopback1
    ip address 172.21.1.1 255.0.0.0

That is a simple rundown of what you need to get a terminal server setup. Below is a link to a the Cisco page that I used to learn how to setup the server. That page will go into more detail as to the parameters that are used.

Configuring a Terminal/Comm Server

Recently, I have been playing with Dynamips, which is an emulator that allows you to run real Cisco IOS images. Experience with Linux made me feel right at home with the configuration and operation of the virtual Cisco devices in Dynamips. However, getting the hang of the configuration and inter-workings, could be confusing for some (including myself at points).

While browsing the Dynagen web site this morning, I came across a program called GNS3. This is a GUI interface that simplifies the use of Dynamips. Since it is a GUI, it allows you to see the topology of the network you are creating, which I find helpful.

The emulator in general is very handy when experimenting with certain things without purchasing additional hardware. I know many CCIE’s find Dynamips useful, since they have access to all of the downloads from Cisco. In addition, Dynamips can be bonded to a physical interface, allowing you to connect it to a physical lab or an outside network.

Below are links to the software I have talked about. Keep in mind, all of this software is free.

• Dynagen - Text-based front end for Dynamips router emulator
• GNS3 - GUI front end of Dynamips

The Day the Routers Died

I have nothing against exercises, it’s just that I wanted a more in-depth experience. While exercises focused only on the topic in the chapter, I was looking more for a book of labs that built on the topics covered in each previous lab.

After looking around, I came to the conclusion that the Cisco Networking Academy Lab Companion books were the best choice. There are two books based around the Cisco Networking Academy courses. The CCNA 1 and 2 Lab Companion starts with the very basics, including cabling (which is something that is valuable if you are unfamiliar with cabling topics). The CCNA 3 and 4 Lab companion goes into more depth, including advanced configuration of CIsco routers and switches.

If you are looking for some labs to go along with your CCNA study, take a look at these books.

CCNA 1 and 2 Lab Companion

CCNA 3 and 4 Lab Companion

Cisco calls this process password recovery, as opposed to a password reset. The goal is to maintain the current configuration on the device, in this case a switch, while allowing you to reset the PrivilegeEXEC and Telnet passwords to something you can remember. In my case, I skipped the last few steps in the Cisco process since I was not concerned with saving the configuration on the switch.

This is a good page to bookmark, since you will likely encounter a situation where you purchase a used switch that will have a password set by someone else.

Cisco Switch Password Recovery

A terminal server is a device that uses a breakout cable of sorts to connect to the console ports of many other devices. What you then do is use one console cable to connect to the terminal server, and initiate Telnet sessions to the other connected devices. In the world of Cisco labs, this terminal server is often a 2511 or 2509 router. These router models have asynchronous interfaces (the 2511 has two), each of which can usually support a terminal connection to eight devices with the proper cable.

Here is a quick list of the hardware that would be required:

• Terminal Server: Cisco Router (2511 or 2509)
• Octal cables
• Console cable and USB to DB9 adapter to console into the router

There are a few things to note. The 2511 and 2509 routers are not your only options. You can use a 2800 series router with the proper network module, however this will cost you considerably more. In addition, the 2511 has two asynchronous ports allowing you to connect up to 16 devices (using two octal cables). If you have more then eight devices, or you plan to expand you lab, the 2511 might be the router for you.

Prices for this stuff will vary. If you’re shopping on E-bay it seems your best bet is to try and scout out a package deal that comes with the router as well as the octal cable(s). Around $200 seems to be the running price for everything. If you are patient, you can find a better deal.

Once I get my terminal server and configure it, I will create a how-to document on configuring one for a lab setting. For now, this following Cisco article will provide the basics.

Configuring a Terminal/Comm Server

• 3640 Router (mentioned earlier)
• WS-C3550-24PWR-SMI switch (With three GBIC modules)
• WS-C3524-XL-A switch
  

The 3640 router arrived today, and it’s in great condition for the price I paid. The three foot power cables I ordered from Mono Price also arrived today. They cleaned up the cabling in my rack nicely, since I don’t have six feet of extra power cables coiled on the side of the rack.

Once the switches arrive and I confirm they are working, the next step will be to purchase some fiber to connect the switches. My goal is also to take pictures of everything after I get the two switches mounted and cabled.

The router is going to come with a NM-2E2W network module (two Ethernet ports) and the DRAM and FLASH maxed out (64MB/16MB). This should be good to get me started. later on down the road I will probably add a serial module to connect this route to my 2513’s.

Next up will probably be a 3550 switch. I am trying to get one with the enhanced image to play around with some of the Layer-3 features.

• 3 - Cisco 2513 Routers
• 1 - Cisco 2924 Switch
• 3 - AUI to Ethernet Transceivers (For the AUI ports on the routers)
• 3 - DCE to DTE Cables (Used for serial connections between routers)
• 1 - Skeletek 16U Rack
• 1 - Aneutronics 20 Outlet Rack mount Power Strip
• 1 - USB to DB9 Dongle
  

Most of this stuff was sourced from E-Bay. If you are patient, you can find some good deals.

The linked USB to DB9 dongle is probably the best I have used. I went through a couple different brands of varying prices, and most of them ended up crapping out on me. The Amazon price is a bit more then the price I paid, but shopping around will pay off. I also find it useful to have a six to ten foot long USB extension as well. This gives you more freedom to move around if you need to.

Eventually I will expand this equipment for more advanced labs and certifications. Some planned upgrades include:

• 16MB DRAM and 16MB Flash on all 2513 routers (Maxed out)
• One or two 3500 series switches
• A few better routers (2600 and 3600)
• VMware whitebox (for servers)

I will take some pictures of my the lab and post them later.

Currently, I am preparing to finish up the chapter on VLANs. That leaves about six to seven chapters left in this book. Once I am finished with that current book, I will move on to the Cisco press books I have for CCNA study.

Below are the books I currently have in my collection:

• CCNA Study Guide (Currently Reading)
• CCNA Offical Exam Certification Library

1. Save money on the hardware (no enterprise class server to purchase).
2. Save money on my power bill.
3. Keep noise to a minimum, or the level you would expect from a typical desktop.

Why did I choose virtualization, and why did I choose VMware ESX server?

The choice to go with virtualization was a no-brainier for me. In a lab setting there will probably be a need to run many different servers. These servers can range from a typical Windows XP build all the way to a Linux or Solaris based server. Having one or two physical servers running two different operating systems just did not give me the flexibility I needed.

With virtualization, I can build as many virtual servers as I want on one physical server. In this case the limit is only the amount of hard drive space the host has available. Keep in mind, I am not talking about having 10-20 virtual servers running at the same time. Your ability to do this without negatively effecting performance will depend on many factors, including the host hardware.

Rebuilding servers is also easier with virtualization. For example, I can build a Debian based virtual server and copy that image to another location off the host hard drive. If in the process of experimenting with different configurations, I completely hose that Debian server, rebuilding it is as simple as coping that original image file back to the host hard drive. This greatly minimizes the amount of time spent rebuilding and reconfiguring servers.

Why VMware ESX? This product is used by many companies of varying sizes (including the company I currently work for). In the process of playing around, I expect to pickup basic VMware administration skills. Adding another skill to you resume, or impressing your boss with you knowledge of VMware can’t hurt!

Recently the Dell poweredge was sold, and I’m still waiting to unload the Sun Netra X1. My plan is to replace the servers with a VMware ESX server whitebox.

The term whitebox refers to a computer or server that has hardware that may not be listed on the official VMware hardware list. Building a VMware whitebox is not only cheaper, but it is also more practical for a home setting where you don’t want a huge server keeping you up at night (let alone driving your power bill through the roof).

Based on the official VMware hardware list, we can find out what network chipsets, SAS chipsets, etc are supported by the VMware kernel. This sounds like an easy task, but this information is sometimes difficult to find. On top of that, you have to be very careful about the hardware you choose, since the VMware kernel is notoriously picky about the hardware it will recognize

My ultimate goal is to build a whitebox for under $1,000 that has the ability to run 7-8 virtual servers with ease.