I need to evaluate two options for hosting email and web server needs.
First option, I can pay “full price”: the going rate of top-end hardware and service at Linode, Digital Ocean, Vultr, etc. (My favorite company at on this tier is Linode.)
Second option, I can pay a reduced price for a package advertised at Low End Box: older hardware, fewer features, but a great value judged purely by number or CPUs, RAM, and disk storage. (My favorite of the LEB purveyors is Wishosting.)
We use email constantly. Gmail is the heavy hitter. But I like having email at a vanity domain. The address firstname.lastname@example.org tells people something about me that I want them to know.
But if email is important then it has to be done right. Your email server should not go offline. Your email server should follow best practices for reducing spam and forged addresses. You should have some plan to get your email back online in a hurry if something goes wrong.
Two types of value from web sites:
I also have a lot of online projects that stem from the lack of impulse control. Domain names I've registered because I thought they were funny or clever. Blogs or wikis I've started because I thought I would keep at them that have fallen by the wayside over time. These aren't inherently valuable, they just exist alongside the valuable stuff.
Services like Linode come with more than just modern hardware. They have fully automated control panels that allow customers to make changes without waiting for staff to get involved. Provision a new server, clone an existing one, change RDNS, resize your system resources, schedule backups, and get billed by the hour. This is an environment that ensures stability and encourages experimentation.
If I want to test the impact to the web server of installing some additional software, just clone the existing web server and test the changes on the clone before committing to them on the production server.
If my server becomes flaky or misconfigured, spin up a clone from a backup image taken before the problem appeared.
So, in fine, you're paying more for the reliability of the service and the flexibility to experiment or get failed services back online quickly.
It costs less. At the scale of human perception the response time between a Xeon E3 vs. E5 isn't going to be noticeable. Not for serving web pages anyway. But a web server running 2 E3 CPUs is going to outperform a server with a single E5 any day. That's where a LEB system acquires its competitiveness: you can get more older hardware for the same price.
The same is true for the speed benefits of new RAM memory vs. years-old RAM. New RAM may run faster but if you don't have enough of it your system will slow down. Having more than enough RAM on a LEB server will prevent your system from thrashing.
So there's the value proposition: LEB sells you hardware that is older but you get more of it for the same price.
The down side is that LEB is sold on month- or year-long contracts rather than hourly billing. If you want to spin up a clone server you'll have to buy it for a whole month. Also, you'll have to do it by hand because their control panels don't allow you to spin up servers based on cloned or backed-up images.
I've tried hosting all my services at Linode. I've tried hosting all my services on a LEB. I've decided that given the small amounts of money involved, and how devastated I'd be if I lost my web content, I'll host my services at Linode.