Follow

Today we are all learning why centralisation is bad and just how many websites rely on fast.ly for some or all of their static content delivery... :BlobCatCoffee:

· · Web · 2 · 3 · 6

What's the betting that the solution to this somehow creates a new, bigger, even more profitable single point of failure?

@petrichor I'd like to gently push back against this by asking what annual uptime the various effected websites would have had, and at what cost, if they didn't rely on single point-of-failure services like Fastly, Cloudflare, AWS/GCP/Azure, etc.?

I've never been directly involved in the process, but I would guess that if my boss told us we were hand-rolling our own CDN, edge and/or core compute nodes, etc., for a new project, I'd expect an order of magnitude more cost and time.

@22 Thanks, that is an important point. I suppose my response would be that the reduction in cost isn't always from efficiency but sometimes from pushing the cost onto others who are either unaware of it or don't have the power to push back.

In this case the SPOF meant that a large chunk of the internet stopped working *at the same time* so a lot of other people who could just change task if one of those websites were down at a time, basically couldn't do anything at all until it was fixed.

@petrichor hmm you’re right, then it’d be a case of everyone optimizing locally and leading to a global suboptimal 😞 in the face of winner-take-most (edge computing and CDN have so strong rich-get-richer cumulative advantage), who can provide the external coordination needed to encourage sites to diversify, as much as they can?

Sign in to participate in the conversation
Scholar Social

Scholar Social is a microblogging platform for researchers, grad students, librarians, archivists, undergrads, academically inclined high schoolers, educators of all levels, journal editors, research assistants, professors, administrators—anyone involved in academia who is willing to engage with others respectfully.