HAstack:
Next-Gen High Availability Hosting Platform by WebMate
A hosting platform specifically designed to provide maximum uptime, maximum performance, lowest latency globally – with built-in failover, resilience and autoscaling for 100% uptime guaranteed.
With over 25 years experience in managing our own network and web hosting for thousands of websites, we’re invested passionately into developing and perfecting our unique platform to deliver the fastest website speeds – especially from power-hungry CMS’s like WordPress – and the highest reliability. Our HAstack platform High Availability web hosting at its very best…
Our single goal was to create the kind of highly redundant hosting platform which large enterprises use for their sites, but without the hefty price tag usually accompanying such a service. But much more than that we wanted it to have multilayer security built-in, and autoscale so every visitor, wherever they are in the world, had the fastest experience every time.
When combined with our EDGE CACHE and Optimizers – which store optimised versions of static elements and images – and at multiple locations on 6 continents – you can achieve at least 400% speed increase. The sub-100ms TTFB (Time to First Byte) that this delivers helps with your page rank on search engines like Google.
To achieve these levels of performance, reliability and scalability, we combined our high performance networks in the UK and US with cloud providers around the globe and specialist CDN services. The result is an optimised website (without the need for plugins) that’s 100% fast and reliable. And the bonus: by utilising on-demand servers to compliment our own, we help the environment a little and keep the costs of this best-in-class hosting service as low as possible.
A self-healing hosting platform combining our own extensive network and know-how with global cloud resources for best-of-class reliability, redundancy and performance guarantees.
Load balancing for HA Reliability
One of the key features of the HAstacks architecture is our bank of load balancers. Three of these work in unison for automatic failover protection with a 1 second switchover. We run ultra-high-performance load balancers which also perform hardware caching of website pages in RAM to make your site the fastest it can be for every device that connects.
These load balancers distribute web requests to the best web server – usually distributing evenly to maintain load but also intelligently monitoring for high traffic to any particular site and isolating or maximising its visitors to maintain speeds for all websites in the cluster. Not only that, these also perform advanced security filtering and DDoS protection. Insane amounts of RAM used on these servers, typically 256GB per server.
Web Servers
Our bank of Web Servers do the heavy lifting work of building pages for visitors from storage and SQL clusters. These are our most powerful servers and feature one “dumb” cluster node which handles primary admin, FTP and runs security sentries – and typically 3 web nodes per cluster to create 4 potentially live servers. Each Web Cluster also features a semi-air-gapped backup node that sits there just in case there’s a major failure. This backup is a bare metal machine with its own storage and SQL which does very little work for the vast majority of time but is on hand if there’s a problem or we need to perform maintenance on all three web servers in the cluster.
Web servers need lots of cores to dedicate to each visitor where possible, typically 48 cores per node. They also need lots of RAM, typically 96GB per node.
Clustered SQLs
Often, much of the “load” in a web server environment is in fetching data from databases – particularly on modern sites which heavily depend on databases like WordPress.
A “family” of SQL servers work together in Multi Master-Master replication to share the load whilst remaining perfectly in sync – so uptime is greatly improved. Often, most downtime with other ISPs during “scheduled maintenance” which needs to happen sometimes daily – with a clustered SQL, that can be performed without any downtime at all – scheduled or not.
These servers are heavy in RAM in order to serve as much data from fast memory as possible. As they are heavy in compute functions too, fast CPUs are essential.
Backups. Then Backups of Backups.
If there’s one thing we’ve learnt over our 25 years of hosting websites is that you can never have enough backups. Server failure is a fact of life – and happens all the time. The key is having the backups and platform that can cope with those outages without breaking a sweat.
But to ensure that we cover all bases, and are prepared for extreme outages or security issues, we keep backups, a lot of them – and at no extra cost to you! Snap-shotting of virtual server images, duplicating those between NAS storage engines and offsite, having scheduled 90 day backups of all sites (and duplicating those to multiple locations) and 4-hourly backups of SQLs are just some of the ways we ensure we’re prepared for any eventuality. It also allows you to revert to any website state in the last 90 days effortlessly – so you won’t need those backup plugins or offsite backup services that clog up your WordPress even more.
The WebMate Hosting Experience: Perfected over 25 years
There are thousands of hosting providers, here’s the four types:
- Resellers: these have one hosting account with a mass-market hosting provider and resell within that. They have very little control of the servers themselves, and they tend to be cheap, often non-UK hosting on a single-server platform like cPanel which cannot provide any form of High Availability. This tends to be the kind of hosting your web designer will provide. Web Designers are talented and creative, but are often not the most capable when it comes to managing a global internet presence, sysops and security!
- Warehouse Hosting Providers: These are the big companies that tend to advertise a lot! Some of these have very capable networks and servers, but as they need to offer low-cost hosting for mass market, their equipment and software is often outdated (and very rarely high-availability). Support levels tend to much lower here too.
- Third Party Cloud Infrastructure: This is hosting on off-the-shelf cloud services like AWS or Google Cloud. The networks are very good (some of our backup services run on these too), but prone to massive outages and attacks. The hosting provider have very little control over performance or security and support can be “less than specialised”.
- Dedicated hosts: Like WebMate, there are other reputable ISPs who manage their own servers and networks – and have very good reputations and skills. Some use cPanel which cannot provide any High Availability but there’s a handful with a network and hosting service almost as good as ours…
And what’s more, we combine our our Dedicated Systems on our own network with Third Party Cloud infrastructure from gloabl leaders like AWS and Google Cloud to deliver site content via global EDGE, load balancing and CDN – with multiple routes of failover built-in by design
Possibly the fastest self-healing, High-Availability hosting platform on the planet
Enjoy a better hosting experience without even having to think about it with our HAstack hosting plans and our uniquely powerful and reliable HAstacks: An award-winning enterprise-grade web hosting solution, meticulously designed and managed in-house by our experienced systems team with three simple goals: Fastest Speeds, ultimate reliability and the best value.
