Hardware Network Taps

June 19, 2008

The last year or so I have been using analyzing network traffic using either a mirrored port or a passive network tap.  The passive network tap is a powerful and very cheap tool; however, it requires that you run a sniffer on two different physical ports.  Each of the ports on the passive tap carries traffic one direction in the communication stream.  So… to make the sniff really useful you need to merge the two packet captures together.  Doing this once and a while is fine; however, doing this constantly is a real hassle and tedious to say the least.

If you are curious here is the pin out to make a passive Ethernet tap;

Passive Ethernet Tap

If you are interested in making your own tap Google for “passive ethernet tap”  Most parts can be found in a home depot or similar homes good store.  If not you can definitely order them on line.  I usually would suggest building a few and leaving them in key parts of your lab or home network topology.  You do not need to leave the sniffer connected for them to function.

Thankfully I recently received a hardware based network tap, which allows me to sniff using one physical interface and only one capture file.  Life has just gotten so much easier in the old home lab environment.

Cat 5K

June 19, 2008

One of my major projects at work is starting to ramp up again, while another is slowly being transitioned off to my peers.  One of the challenges we faced this year was that we would retain some older Catalyst 6506′s running hybrid mode.

A few weeks ago I added two Catalyst 5000′s to my home lab.  However, right now they are living in our data center to assist in pre-cutover testing.  While not completely and apple to apple comparison, the Catalyst 5000 with an RSM module in hybrid mod is similar to the Catalyst 6506 in hybrid mode.

Besides that some progress was made testing new Cisco 2611′s that will be replacing the 2610 in the home lab.  Over the next few weeks I will be swapping out my all ten of my 2610′s for ten new to me 2611′s.  All of these routers, with the exception of one or two 2610′s are running at max DRAM and FLASH.  The 26xx platform is really only useful when you max out its memory so you can run a 12.3 IOS with 802.1Q trunking support.

Panduit D Rings

June 6, 2008

Managed to find some extremely inexpensive second new in package Panduit D rings.  Got them for a whole $1/ring!  Using them now on the backside of the rack rails to route power and console cabling.  Should keep a bit of that cable pile away from the front of the rack.

When setting up my home lab I generally view things like power and console cables as items that can be pretty static and unchanging.

Spanning Tree Protocol

June 6, 2008

Put some time into the lab the last two days.  Primarily playing around with some STP scenarios.  Mostly using four of my Cisco 2950 switch’s to simulate specific desired topologies.  Want to know exactly how the network would converge regardless of the failed component.  Was fairly easily, hardest part was probably doing up the diagram and patching in the cross over cables.

Either way good to get time in on the lab.  Wanted to play around with MST a bit; however, I did not have the time to do up those configurations.  Hopefully I can revist that topic a bit later.

Cisco 2600 Memory

May 27, 2008

Received in what will hopefully be the second to last order of memory for my Cisco 2600 series routers.  The 2600 routers need two MEM2600-32D modules to bring them up to their maximum DRAM values.  This recent round of memory upgrades will bring me up to a total of eight 2610 routers, which are running all at their maximum memory configurations.

Later this week hopefully I should have two or three 2611 routers arriving.  I’m unsure what level of memory these new routers will arrive with; however, all things considered I paid very little for them, that adding a few memory upgrades still keeps them within a reasonable price.

Depending on the condition of the 2611 that arrive in I may be swapping out 2610 routers.   I save all my old memory and flash in static bags, in a parts bin when I upgrade devices.  If I pull out a few 2610 I may depending on the eBay prices post a few up for sale to off set the recent lab purchases.

Hardware Testing

May 24, 2008

Had a bunch of hardware arrive the past few days.  Primarily various NM and WIC modules for Cisco gear.  I’m getting close to the point where I don’t need anymore modules.  The time I had the last evening or two was spent testing.

Generally when I get gear in I will tag it with a little paper tag and put it into a test router.  The tag lets me know where I got the part from.  So far I have not had any failed parts or needed to send them back to sellers for refunds.  The first basic test I do is of course, does it show up in the router?  Then I’ll configure the interfaces with static addresses.  If I have a lot of devices to test that day I will configure some quick dynamic routing protocols and start pings between all the interfaces.  The goal is to get a fairly full mesh setup going, with packets originating and arriving at each interface.  Once the packets are sent I will check the interfaces for errors.  No errors, then I assume it is fairly good hardware.

Cable Management

May 22, 2008

Lucked out, some items that I thought would be in transit longer came in the mail today early.  So far I have decided to go with Panduit parts, which I have always been a fan of.  Unfortunately I do not get to use them as often as I would like because they are unfortuantely more expensive than their competitors.  Fortunately for my home lab I have managed to source what I have needed from people selling off extras.

2U Panduit Wire Management

Now all I need is my verrtical cablemanagement pieces and a few more 1U and 0U pieces.  I found what appears to be some nice 0U items at a network surplus web store.  Hopefully they will arrive tomorrow or early next week.  Stinks that Monday is a holiday!  Screws up shipping.  However, I’m looking forward to our family cookout.

Cisco Technical Documents

May 22, 2008

There are many reasons why I enjoy working with Cisco products, in fact probably to many to even cover in a blog format. However, one of the top reasons why I like Cisco is the quality of their technical documentation. To often, for example with Nortel you get just a bit of documentation frosting thrown in with the expensive product you may have just purchased. With Cisco I have always felt that you could easily read for hours on end. The hours spent reading would be at times more valuable than similar hours put in reading competitors documentation or printed books.

Here are a few short PDF documents, which I tend to keep handy or re-read;

Lately I have been going back to those documents due an on going latency and general network degradation of service investigation at work. At work we have a very mixed environment in regards to both age of equipment and also the vendors who provide the equipment. Wishing some vendors would step up the quality of their products and provide simple things like the ability to run a debug ip packet command in IOS.

Used Network Hardware

May 22, 2008

The last few weeks I have been anxiously working to get my home lab to a certain point. I want to use the finalize but I know that is unrealistic statement to make. Networking labs are like art to some extent in the context that they really are never complete.

Most of my energy lately has been focused on the actual racking aspects of my gear. I ditched two relay racks, which I had gotten used several months ago. What replaced them was one brand spanking new Chatsworth standard black 19″ relay rack. The primary reason for replacing the previous racks was that the previous owner(s) had done a good job of stripping out some of the pre-threaded holes. If I was racking up 4U servers, or only sparsely populating the rack, I may have been able to utilize it longer. Given that reason for replacing the rack, I talked myself into buying nearly 300 12-24 threaded screws for the new rack(s). So far my only disappointment in the rack is that the advertised rack U space markings were done in a teal color, which is barely readable on the black coated rack.

One thing to note is that black racks and screws are coated, whether it is painted, powder coated or anodized, metal does not come out of the forge black. I found it worth while screwing in the screws once prior to mounting the gear, which caused the coating inside the threaded hole and on the screw itself to wear away with out the weight / positioning of gear causing the screws to incorrectly thread and strip out the holes. Read the rest of this entry »


Follow

Get every new post delivered to your Inbox.