> On Feb 24, 2024, at 20:30, Robert McMahon via Nnagain wrote: >> On Feb 24, 2024, at 4:05 AM, Fearghas Mckay wrote: >>> On 22 Feb 2024, at 18:58, rjmcmahon via Nnagain wrote: >>> Boston University spent $305M on this and it doesn't have an IXP. >> There are two well established distinct IX in the metro already - why would you need a third that is not at an interconnection site ? > For education and modeling. Students used to run on campus radio stations as an example. A modern world requires understanding things like IXPs. That’s a good point. When we do IXP formation workshops, if there are very many ASes which haven’t peered before, we do get the whole thing up and running on a meeting-room table in a classroom sort of situation, before trying to do it in production. And we’ve done joint projects between IXPs and departments of education. So this is a very good reason for a university to host a small-scale or lab IXP. But it’s also a good reason to forge a relationship with nearby production IXPs, such that student projects can get access to no-cost local bandwidth, etc. Students getting access to distance-insensitive bandwidth at universities causes a lot of startups to crash-and-burn when they start to scale up and find out that the Internet isn’t flat, it just looks that way when you’re buying small amounts at retail. -Bill