Faster arXiv with Fastly | Hacker News Reader

Faster arXiv with Fastly

blog.arxiv.org

98 pointskickdaddy10 comments

Good promo for Fastly! Fastly is a great company - effectively they're Varnish-as-a-Service, which is really a non-trivial SaaS. I hope this brings in some business for them.

I read an analogy once and its that:

Fastly == Linux (ability to changes lots of things to tailor to your exact needs)

Cloudflare == macOS (prepackaged to work for most general users, limited config)

I've used both and wish Fastly had a mode where a higher level specification could be used for most/simple cases. It's been quite a while since I've used it so may be a capability it has since added.

They have a whole UI that avoids any sort of VCL.

I feel like Cloudflare is becoming more of a cloud computing platform with a focus on global "instant" delivery of services than a CDN company.

Of course, they're also a really competent CDN company, but they're really leveraging all their experience in that domain to bolster up their compute offerings.

Cloudflare wants to be the defacto "Cloud 2.0" company (where AWS is "Cloud 1.0").

Where Cloud 1.0 is laregly just a "lift/shift" of OnPrem to Cloud.

Cloud 2.0 is hosting on the Edge and infrastructure built completely different than anything legacy from an OnPrem (or lift/shift) mindset.

Good job Fastly. I would love to see metrics on bandwidth consumed from Arxiv. I always wonder how it and services like package repositories are able to provide their downloads for free.

That's a cool idea! Arxiv is a member of our Fast Forward program at Fastly, which helps a lot of open source/open web projects, and we've enabled other projects to share that kind of data in the past, so maybe we can figure out how to help them do that too. (One example is Kubernetes; you can see their live request data on our developer site. https://developer.fastly.com/ )

Love how the bandwidth spikes ~10x from HH:59 to HH:00. Looks like people still write crons with manually set trigger times that align on the hour, rather than randomly spreading them out.

This is such great news, I've been annoyed with slow arXiv loads for a while and happy to hopefully see it fixed!

Reply on news.ycombinator.com