What's actually underneath your website
Most hosting pages ask you to take the infrastructure on trust. This one doesn’t.
If you want to know how the platform behind your site is put together before you hand it your business – what happens when a server fails, what happens when traffic triples, and what stops a busy database slowing down your email – this page explains it.
Or ring 01202 237 273 if you’ve got a specific technical question. You’ll get an engineer, not a script.
"It's fine" is not an architecture
Every host says their platform is reliable. Almost none of them tell you what reliable is made of, which is how businesses end up finding out the hard way – on the morning the site is down and nobody can say why.
The difference between a platform that copes and one that falls over is nearly always structural. Is there more than one of everything? Does the platform react to trouble, or does it work to avoid it? Is there a cap on how much of a server you’re allowed to use, and what happens on the day you need more than your share?
Those are answerable questions. Here are our answers.
The short version
For readers who want the summary before the detail.
More than one of everything
Web, database, file and mail servers all run in multiples. If one has a bad day, another picks up the work – no single point of failure anywhere in the chain.
Traffic is steered constantly, not only in a crisis
Global load balancing watches where demand is and spreads it across servers all the time, not just when something breaks.
Separate servers for separate jobs
Your website, your databases and your email each live on their own kit. A problem in one place stays in one place.
Resource arrives when you need it
The platform adds processing power and memory as demand climbs, then hands it back afterwards. No caps to trip over.
All-SSD storage on Dell hardware
Solid-state storage in a mirrored, striped array, replicated across multiple servers. Intel Xeon processors, battery-backed RAID cards, technicians on site round the clock.
Load balancing that doesn't wait for something to break
This is the part worth understanding properly, because it’s where most platforms stop short.
Think of a supermarket. Basic load balancing is a shop that opens a second till only when the first one breaks – the queue eventually moves, but everyone in it has already had a bad few minutes. That’s failover, and it’s what “load balanced” usually means: your site gets moved somewhere else once the server it was on has already died.
Our platform runs a till count that changes with the queue. Global load balancing monitors traffic continuously and distributes it across multiple servers as demand shifts – before anything is struggling, let alone failing. Redundancy is still there underneath, so a hardware failure is handled, but the day-to-day job is keeping your site quick rather than merely alive.
Which means: the improvement shows up on your ordinary Tuesday, not just on the rare morning something goes wrong. Your visitors get a page served from a server that has room for them.
Separate servers for separate jobs
Plenty of hosting puts your website, your database and your email on one machine. It’s cheaper to run, and it’s fine right up until it isn’t.
Think of the electrics in a building. Everything on one circuit means one fault takes out the lot, and one greedy appliance dims every light in the place. Sensible buildings split the load.
Our platform splits it too. Web servers serve pages. Database servers – MySQL and MSSQL – run on their own machines with far more memory than the web tier, so a heavy query never has to fight your homepage for resource. FTP servers handle file access. Mail servers handle email. Each role is its own set of servers, and each set has more than one server in it.
Two things follow from that, and both matter.
A compromise in one place doesn’t hand over the others. If something gets into your website, it hasn’t automatically got your mailboxes as well. That’s containment, and it’s the sort of thing you only value in hindsight.
A busy database doesn’t slow down your email. Or your file transfers, or your checkout. Load in one place stays in that place. Read more about how the platform is protected on our hosting security page.
Autoscaling, and what "no LVE limits" actually means
Shared hosting normally comes with a ceiling. The technical name is an LVE limit – a cap the host puts on how much of a server any one customer is allowed to use. Hit it and your site doesn’t fall over dramatically; it just gets slow, or starts refusing visitors, and you get an email about exceeding your allowance.
It’s the shared office with a fixed electricity supply. Everything is fine until someone plugs in one more kettle, and then the trip switch goes for everyone.
There are no LVE limits on this platform. Instead of a slice of one server, your site draws on the capacity of the whole platform: as demand rises, processing power and memory are added automatically, and released again when the rush is over. Nobody has to notice it, approve it or upgrade you mid-morning.
Which means: the day your product gets a mention, your traffic goes up and your site stays up. Good days stop being risky.
The hardware and the data centres
We should be straight about this: we don’t own the data centres. We host on a platform built and run by a specialist infrastructure partner, and we chose it after looking at what it’s made of. Here’s what that is.
Every hosting server runs dual Intel Xeon Scalable processors, with database servers built to the same class of machine and given considerably more memory, because that’s what databases actually run out of. It’s Dell hardware throughout.
Storage is 100% SSD, arranged in a RAID 10 array. In plain terms, RAID 10 means your data is written to more than one disk at once and split across several, so you get both a spare copy and faster reads. The RAID cards are battery-backed, which protects in-flight writes if power is interrupted. The arrays are then replicated across multiple servers, so a failed disk – or a failed server – doesn’t cost you anything.
The data centres themselves have technicians on site round the clock, every day of the year, sitting on a resilient national network.
The platform we host on runs to a 99.99% uptime guarantee, backed by that redundancy rather than by a promise on a page.
Where this platform shows up
Everything above is underneath all of our hosting. What changes between plans is how much of it you get and who looks after it.
Web Hosting
The standard tier for business websites, with free migration from your current host.
Web HostingHigh-Frequency Hosting
The same platform on faster processors, for sites where half a second is money.
High-Frequency HostingManaged Cloud Hosting
Dedicated compute on your choice of cloud, run for you end to end.
Managed Cloud HostingHosting Security
The firewall, certificates, scanning and certifications that sit alongside the infrastructure.
Hosting SecurityWorking with us
Good infrastructure is only half of it. The other half is who answers when you ring.
We’re an IT company that hosts websites, so your site sits with the same team that looks after your computers, your email and your network. Nobody spends a week working out whose fault it is.
We’ll also tell you when you don’t need the bigger option. If a standard plan does everything your site asks of it, that’s what we’ll recommend.
Getting started
1. Tell us what your site does
What it’s built in, how busy it gets, and what breaks today. A short phone call usually covers it.
2. We match it to a plan
The smallest one that comfortably does the job – and we’ll say so if what you’re already on elsewhere is doing fine.
3. We move it and look after it
Migration, setup, monitoring and updates. You get on with the rest of your week.
See the hosting plansCommon questions
The things technical readers ask us about the platform – failover, limits, storage and what happens when something does go wrong.
If yours isn’t here, ask. You’ll get a proper answer from someone who knows the infrastructure.
Do you own the data centres?
No, and we won’t pretend otherwise. We host on a platform run by a specialist infrastructure partner. We chose it on its architecture, we manage your hosting on it, and we’re the ones who answer the phone.
What actually happens when a server fails?
Traffic is already being distributed across multiple servers, so a failure means the platform stops sending requests to that machine. Your site carries on being served by the others. Storage is replicated too, so a failed disk or array doesn’t take your data with it.
Is there a limit on how much resource my site can use?
No LVE limits, and no fixed cap on processing power or memory. Resource is added as demand rises and released afterwards. If your site’s demand is consistently high, that’s a conversation about the right tier rather than a throttle.
Why does it matter that email is on a different server?
Two reasons – containment and speed. If your website is compromised, your mailboxes aren’t automatically part of it. And a database or website under heavy load has no way of slowing your email down, because they aren’t sharing a machine.
Is this fast enough for an online shop?
For most shops, comfortably. SSD storage, load balancing and automatic scaling all help a checkout stay quick. If yours is large or seasonal, high-frequency hosting or managed cloud hosting is the better fit, and we’ll tell you which.
Can I see this for myself before committing?
Yes. Tell us what you’d want to test and we’ll set you up so you can look at the real thing rather than take our word for it.
Still want to look under the bonnet?
Ring us and ask the awkward question. We’d rather explain how the platform is built than have you find out during an outage.
See the hosting plans → · 01202 237 273 · service@rejuvenate.it