Let's Ginger-Fi it!

aka the one I didn’t submit 😉

We currently have over 21,000 students, 6,000 faculty and staff, and 130 buildings across our different campuses. These buildings range from 2 story houses to 15 story towers and everything in between.

Just before covid lockdown, a new network plan outlining a switch from Aruba wireless & Cisco switches to a full Cisco SDN (Software Defined Network) was announced. This new network is being implemented by outside contractors and since many companies cancelled their Cisco orders during the covid lockdowns, we were able to get our switches and access points by June of 2020. While the campus was empty, we completed the switch rollout, but until the new Cisco SDN infrastructure was standing, our 2000 Cisco 9130 APs would wait on their pallets. 

The Issues:

As the beginning of the new network rollout approaches, one of the big parts of my job has been planning the “seamless” change from our existing Aruba wireless network to the new Cisco network.

We exist in a mix of high-density areas and offices, hospital and clinic spaces, a morgue, large sporting areas, animal barns, and every type of research area you could think of including a few radioactive spaces. We even had our own nuclear reactor until it was removed in 2011. This space was then turned into offices. We are also a BYOD environment resulting in approximately 17% 2.4GHz only, capable clients on our network. 0.04% of clients are 802.11g.

Since we were in the process of upgrading our indoor Aruba APs from a mix of AP105s, AP115s, AP125s, AP135s, AP205s & AP225s to AP315s and AP345s when the decision was made to move to a completely Cisco network, we have a mix of all the above (minus the 125s due to being EOL). This causes issues in spaces and buildings that have a mix of old and modern technology with the older APs only being capable of up to 802.11n technology.

Another large issue is that all the 2.4GHz radios across campus are enabled leading to very high interference rates in the 2.4GHz airtime utilization. In spaces I have monitored, the noise floor in the 2.4GHz spectrum has an average of -73dB with an interference rate of 74%-81% while >1% utilization is coming from the clients.

The Solution:

I created a detailed plan for rollout including printouts of existing floor plans with AP locations, step by step instructions for AP installs and how to provision the new Cisco APs. This plan also includes a document we will use to note any anomalies during installs (non-drywall walls, ceiling height changes, mechanical equipment in area, etc.). With the included maps, we will also be using this as a time to verify AP locations. Our current floor plans are quite out of date for renovations completed in the last 10-15 years and often have human error in AP locations added to Airwave after installs. All of this will be used to improve our future troubleshooting and monitoring capabilities.

I came forward with recommendations for the new network based on how we had been doing things, knowing that an evaluation had not been done in over 15 years. Because of the large spaces we cover and the high-density of our environments, I suggested that we change from our current setup for our radio profiles. Everything is currently set to 20/40MHz channels. Having a mix of 20 and 40MHz channels in the same space causes issues where the 20MHz primary channel could be the same as the primary or the secondary of the 40MHz. It also limits the number of channels we have available in each space and causes co-channel interference more than it helps with coverage.

Also, since every time we widen the channel, it adds and extra 3dB of noise. This will affect the SNR in the space and affects the MCS rates the client device can choose from and therefore lowers the data rate and throughput of the devices, removing the benefits of bonding the channels.

There was a comparable situation with the macro/micro cells offered with the 9130 APs. Management thought they sounded like a great idea but then I pointed out the issues with channel reuse and the fact that, a lot of the time, client devices do not choose to connect to the micro cell. It is better to use the single 8×8 radio than two 4x4s with one being underutilized and potentially causing co-channel interference in high-density areas.

I also added the requirement of an RF profile that would not include the 2.4GHz radio on an AP, finally giving us the ability to disable unnecessary radios is spaces and reduce our co-channel interference in as many places as possible.

In Conclusion:

I am grateful to have the wireless community members to reach out to and bounce my thoughts/ideas off and all the knowledge obtained through my CWNP training. Being the first wireless person in a large institution, I sometimes worry I will not come across as knowledgeable as I have become. The ability to reach out and gain confidence by speaking with an existing wireless expert really helps boost my confidence in what I am saying or suggesting is correct. So far, I am happy to say I am being heard and we are creating a more modern wireless network for or students and staff.