It’s totally overkill for the article’s use case, but it enables many cool things that need sustained bandwidth, and has been pretty fun.
I went to their home page to try and find more information about colo pricing, but it's all hidden behind a contact form.
Their site design seems very 1998, makes me feel really nostalgic.
> For example, with a 3-year term for a 42U cabinet, it would be $400 a month for 3 years, and that includes 1G already. After that, for another 3 years, would it be $340/mo for 1G, plus $500/mo for the 42U cabinet, which is $840/mo (more than double)? In both cases, is a /27 always included with the 1G port (pending justification)?
> Yes, you are understanding what our standard rates are without the promotion. Yes, we include the /27 with service. Our promotion is currently our most cost effective solution. Let me know if you have any questions.
Do you really want to build a whole 42U rack full of big heavy servers and have your price doubled after your contract expires?
It is certainly possible that the rate increases at some point, but I don't see why either HE or you would want to commit to anything beyond 3 years from now?
Remember to set them crypto mining so you're kept warm at night.
Wonder what if Netflix turns around, and blue oceans [0] its video delivery infrastructure [1][2], like Cloudflare did for text/html. We do know that video has all but eaten up the attention that text/html once had on the Internet [3]
[0] https://en.wikipedia.org/wiki/Blue_Ocean_Strategy#Concept
[1] https://news.ycombinator.com/item?id=28584738
Storing a large number of "cold" or "long tail" bytes is a different physical problem when it comes to storage and IO. And that's also the exact problem you have when hosting lots of third party content (or apps, or lambdas, or anything else). There you're also optimizing for bytes stored:served, but have to place it on cheaper media and with less duplication to make the physical space/power/dollar constraints work.
They certainly could change their business model and infra. But then they end up looking a lot more like a commercial CDN with hierarchal stores, more centralized POPs at IXPs, etc.
Disclaimer: Principal at AWS. Used to work on Amazon CloudFront and a bit of time in AWS Elemental, served a lot of video & streaming bits. Everything above is public domain knowledge.
We used the extra space for spare parts.
Also agreed with sibling comment that 1Gbps is pretty slow nowadays. Do you know if that can be customized?
(And yeah, 42U.)
Is this so different from the cloud? Even an EC2 instance should be assumed to be a single point of failure and you also need multiple (preferably in different AZs or regions) if uptime is important.
Old and boring things, DNS, load balancer, yada-yada.
If you save thousand(s) it is still could be cheaper.
If you want to save some money write the scripts yourself and run them in a cron job. Took me a couple hours and I'm no sysadmin.
To me the best thing to do is move all your services into k8s and automate you DB (use open source). Then you can move to any provider when you realise the cloud pricing is shit.