My network and my network lab at work: Awesome.
I’m lucky to have a very nice lab to test and support my network at work. Just a few things that I’ve got at my (groups) disposal:
This is the latest addition to the network. I have a pair in production with 10Gb uplinks receiving full BPG routes from I1, I2 and native IPV6. I managed to talk our provider into a dedicated 1Gb link for my lab (which is converted to 10Gb for the MX by a Nortel 5530) Thanks to this (and a pile of fiber), I was able to prototype our full redundant config before going live with it. I even managed to secure a /24 that I advertise via BGP for live connectivity. These supply per IP bandwidth policing (across two class Bs), generate default routes into OSPF for everything else and they are fully redundant – both routers mirror every function. I can boot one at anytime with no consequence aside from a BGP flap to my provider.
Tipping Point SMS, Core Controller, 2400E IPS and 1200 IPS:
Before I implimented the MX routers, I had to update our campus IPS. On our edge, we use Tipping Point hardware. The IPSs use FPGAs on several Gb links to do initial, line speed analysis. In depth analysis is done by intel CPUs as needed. (The next generation is even better…) By using a Core Controller, we split the 10Gb link into multiple 1Gb links that are processed by several 2400E IPS units. The signature updates from TP are excellent, and the benefits of using the system on our network are impressive.
Juniper M20 Routers:
Our old border routers. I use these to simulate the multiple router setup in production since it’s hard to justify a pair of MX960s just for lab/hardware spare use.
Nortel ERS8600 Layer 3 switches
Our core aggregation switch. These handle multiple link redundancy, 10Gb core up and cross links and do just about everything you can think of.
Nortel 5500 and 5600 GigE swtiches (The edge switches used across campus)
I have a small pile of these switches for testing code, simulating production
Netscreen 5400 Firewall:
Our core uses four of these in redundant(ish) pairs. There are a few single connected services but they mostly back each other up. These protect our data center and user networks. Again, core connections are 10Gb. 