Wi-Fi that Works.
As hackers started pouring into the EECS atrium of MHacks IV, we were panicking. Because of classes, we had only been in the venue for 4 hours before them, and ethernet had not been deployed. Wireless access points were deployed an hour ago.
Before I go any further, I need to tell you one reason why Wi-Fi was way more important at MHacks IV than ever before.
North Campus really doesn’t have cell signal.
At U-M, people commonly ask each other “Why didn’t you text me back?” and respond with “Sorry I was on North”
Every hacker would be bringing their phone, laptop and tablet. If the network goes down, there’s no Mi-Fi, tethering, P2P internet. If Wi-Fi dies, North Campus goes dark. Might as well cut the power too.
Throughout the venue, we worked with our IT department to double the number of access points in the space, and upgrade them to dual-band enterprise-grade routers. But still, each router was broadcasting 4 networks— MWireless (U-M students), MGuest (guests), eduroam (all other schools), MHacks (us). Tons of networks, tons of channels, and IT told us that they were planning on each hacker bringing two devices.
Hackers don’t bring two devices. They bring their laptop, their desktop, their smartphone, their mom’s iPad and grandmother’s Kindle Fire. All devices would have to connect to Wi-Fi. The probability of something going wrong was huge. We didn’t even need Murphy’s Law.
We planned for Wi-Fi going down within the first two hours of MHacks.
We had several contingency plans, all of which we shared with our volunteers— who serve as our eyes and ears in every corner of the MHacks venue.
If hackers can’t connect to Wi-Fi in one part of our venue:
- Tell all hackers to stay where they are for five minutes. If they moved, they’d bring down Wi-Fi in other areas of our venue like a plague.
- Pass out Wi-Fi 5GHz dongles to get more people on a better spectrum.
- Deploy ethernet infrastructure to 70% of the room (initial plan was to have 30% of hackers in each room connected to ethernet)
- If 1 & 2 don’t work, relocate room to another section of North Campus, and bring them extra food.
If internet goes down throughout the venue (a la MHacks II):
- Tell all hackers to stay in place, or feed them in the atriums, or start laser tag early. The primary objective was to keep hackers away from running all over campus trying to find Internet
- Contact the IT professional we hired to be on site for the first 5 hours of MHacks
- Relocate the entire hackathon to a secondary location we had set up on Central Campus
So why did Wi-Fi work so well at MHacks IV?
We think Wi-Fi succeeded for several reasons.
- We were fortunate. Not only fortunate to have a solid backbone of U-M internet infrastructure, but just luck. We were definitely lucky.
- Because of the lecture halls, discussion classrooms, and computer labs, hackers were pretty well dispersed throughout our venue.
- We worked to place access points where hackers would be, instead of just trying to cover the venue with Wi-Fi twice over. We knew hackers would be in the Quixey lounges, the atriums, and that the lecture halls would be pain points. We focused specifically on this areas.
- Brand-new enterprise-grade dual-band access points. They’re designed to handle this stuff. We also had a huge IP allocation.
- We had our own full-bandwidth network. When we were in the Big House, hackers that weren’t students were only given half the bandwidth that’s available to Michigan students. MHacks IV hackers got it all. Full 100Mbps+ glory.
Huge thanks to our University IT department for supporting MHacks IV and getting 1200+ hackers streaming videos, downloading files and pushing commits to work without problems for 36 straight hours.