My home lab setup

I’ve commented about my home lab setup.  I’m hyperlinking to the photos because of the size of them. It was messing up the wordpress page.

one picture
second picture

Yes, it is in my kitchen/dining area.  No, my family isn’t really thrilled about it (well, the 16 month old loves to play with it).


7 Responses

  1. Wow! must be very loud in there.. best of luck attaining your cert, seem you have all necessary to get it… very nice lab…
    I’m actually studying to re-obtain my ccna which exp in ’04…

  2. @ Rod,

    I don’t keep the rack running 24/7. It just isn’t cost effective to do so. The power bill would be outrageous. I generally run the rack all day saturday, and maybe two nights a week if I get a chance to work on labs. I have noticed the lights dim when I fire the rack up, though.

    Good luck with the re-cert. I let my CCNA expire the first time in 2004. I re-certified the CCNA in 2006. I should hopefully have BSCI out of the way next week.

  3. We have similar hardware (I’ve got a 7206VXR, mostly 1841’s but the same 2800s and a few 2621’s) and one thing that really helps manage my time and keep the power bill sane is my ACP9211 PDU’s. They have serial on the front so I connected both to my NM-32A via hoods then the network port to a smaller non-lab switch so they are on the same branch as my netscreen, cable modem and desktop.

    This setup lets me power up the rack all at once, selectively or just the gear for one specific lab. To me this is an aspect just as important as getting your first 2509/2511 – a real blessing. Best part is skipping the loading time by turning it on from work so by the time I get home its ready for me to dive in.

  4. @ Ash

    I have access to a handful of spare 7200’s that have been retired at work. I just can’t see adding more crap to the lab at this point. It’s really unnecessary.

    I have looked into addressable PDUs for the purpose of remote access and power consumption. Between cost and time that I have available for lab work at home, I can’t justify it just yet. At this point, it’s still easier for me to just power the gear up on days where I have a few solid hours to dedicate to lab time. Unfortunately, that hasn’t been all that much lately. Maybe when I’m able to get out of my apartment and dedicate a room/office to my studies…

    Honestly, I’m seriously considering selling off a few of those 2600’s and a few of the switches to fund a decent laptop or mac mini to work dynamips labs for the rest of my routing based studies. That would also allow me easier access to lab time on the fly.

  5. The 7200 series doesn’t really get interesting until you have an NPE400 or better – the reason I use mine is because it lets me reduce the amount of gear I need to have on hand while remaining as flexible as possible. I use a PA-MC-8TE1 card with some modified CAT5 cables to simulate T1’s to the rest of my routers, the quad FE interfaces for QoS/MPLS from the core out and the PA-GE to connect my HS router pod to the switch pod (added bonus of being able to import work configs and model/troubleshoot at home design/deployment issues).

    I use one of my 2811’s as the frame switch feeding four 1841’s with a split switch lab of a 3560G, 3550, 2960G and 2950T. The 2801 is primarily for my voice lab but gets reused as needed. The only in service 2600 is for the NM-32A so I can SSH to it from work through my netscreen. I had 10 2600’s at one point but that platform just can’t keep up, I liquidated those for the 1800’s and I’d encourage you to do the same.

    Dynamips is nice for quick no-frills concentration but frustrating for larger labs, at least for me. I use a 4U white box that runs vmware to support the lab and generate traffic, it was the cheapest way to go while letting me keep everything in a 20U rack on wheels (Skeletek).

    Shoot me an email sometime if you are still looking for PDU’s, I might be able to help.

    Nice blog, keep it up.

  6. @ Ash,

    That’s a lot of great info. Thanks for the heads up. What problems were you having with the 2600’s not keeping up? The majority of my 2600’s are XM’s (granted, I don’t have them maxed out with flash/ram just yet), but they’ve been great for the NP studies, so far. Honestly, my biggest investment at this point has been roughly $70 for the serial cables and $50 for the rack, itself. I’ve been pretty fortunate.

    I was doing something similar with a handful of WIC-2MFT-T1 and WIC-1DSU-T1 cards and the T1 x-over cables before I picked up all of the serial interfaces, last year. Are you running CCME on one of those 2800’s for the voice lab, or are you running a virtual machine with full blown CCM? That’s one area where I’m really lacking and have been wanting to start playing.

  7. The 2600’s are fine if the sole focus is basic routing but when you throw in mpls/qos and security they positively creak. From what I’ve seen the refresh towards 18/28/3800’s has been faster than the 25 to 2/3600 was. It was more worthwhile for me to have current generation hardware so I’m not just familiar with the basics but the expanded features industry regulations has pretty much forced a rapid adoption of.

    Where I work there are twin OC48’s, 7206’s focus ptp T1’s (many for voice) to remote sites with FR as secondary paths and gigE trunks to other campuses. Modifying this deployment model for home (downscaling) makes it easy to run parallel VP and NP labs along with trying on the DP hat to optimize where I can. Remotely accessing the rack helps break bad habits learned from hands on deployment where if something goes wrong thats okay – you are right there to fix it. The two ccie’s I work with encourage me to do as much as I can without touching anything which mimics typical real world restrictions which they say will help on the IE lab.

    I have a CCM vm image and CCME on a 2801 with two 7940G’s. Usually my vmware server has two vm’s running, one generating traffic, the other collecting metrics as a challenge to do a change with no or minimal measurable impact. Granted, when I’m diving into something I’m shaky on I care less about the numbers but I can see after its all done what it looked like from the outside or from an SLA standpoint. Could be overkill but its helped a lot on the job to know just how touchy certain platforms or hardware combinations can be. Nice thing about vmware, the generator image and metrics center is that they are free, the cost was in hardware which doubles as my desktop at home (laptop is a remote) which I needed anyway.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

%d bloggers like this: