Of course, everyone says that they deliver good quality, but why is it so much different in performance and reliability? Below we present some details that enable us to deliver high quality at very reasonable prices.
Shared servers
We start with the very core – the hosting server, the actual service a shared hosting solution is using, to then work our way back in the parts that mainly affect quality and performance.
Rackfish web hosting is careful with the performance for sites to respond quickly, which affects our customers’ results – whether it’s pure information pages whose readers are staying or if it’s a store that sells more.
One way we’re handling this is by using a CloudLinux as an operating system, which is adapted specifically for servers with many customers on shared servers. CloudLinux makes sure that every customer is better isolated and not being so affected by their neighbors.
We are also using Kernelcare, which is a software from Cloudlinux that gives us opportunities to make updates without rebooting (so customers don’t have to get downtime at the wrong time).
In addition, we use CloudLinux platform Immunify 360, which is very competent protection against various kinds of attacks on a web server or specific services such as WordPress. This is a so-called WAF – Web Application Firewall that finds, for example, intrusion attempts.
Our shared hosting servers also have “High Availability”. This means that they automatically start up on new hardware (a new server), in case something should go wrong – here you don’t need to wait for a new hardware and restore backup – it only takes a few minutes and then we’re up and running again, even if the hardware itself broke down.
Of course, we still have backups, in form of snapshots (taken several times a day on the server and stored for a few days) and server backups sent to our other data center for longer storage. Some packages include extra backup, such as our Managed WordPress where you as a customer can back up and restore, or get automatic backups before each update.
Servers
On the next level, we have physical servers. These are all provided with dual connections to our backbone network. If something happens with the infrastructure, there are always backup routes and equipment in reserve.
Our servers are virtualized with VMware. This gives us great advantages, such as automatically balancing the load so we always have the capacity to deliver what we promised. The possibility of servicing hardware without the customer getting any downtime at all (not even due to reboot) and have the possibility to automatically restart servers on new hardware in case a server breaks down.
Our storage is in different storage network solutions. Because we have separate storage from the servers, we can easily move operation between different servers, even when one server is down.
Internet
No matter how good all operation works, of course, internet connections are super-prioritized to actually reach out.
We at Rackfish have built failsafe double networks in our centers for our internal communication. Our network infrastructure is from Arista and Cisco and we use 10 GBE to not get bottlenecks between servers.
To get online, we have agreements with 3 different suppliers, each of whom can handle all our traffic.If one supplier gets any problem, we shut it down and wait for better days – it’s certainly expensive with so much overcapacity, but oh so safe!
In addition to this, we have traffic exchange (“peering”) directly with some 20 other Swedish and foreign operators which will be another way we can reach your visitors.
We have protection against DDoS, through market-leading Arbor Networks that stops attacks outside our network (before it reaches us, that is!), and even the protection on the servers can handle minor attacks.
For visitors to always get the best performance, we also have a system that optimizes the roads traffic takes – this means that 30% of the traffic gets a better performance (and fewer time-outs) compared to a traditional solution!
Data center
We have our operations divided between two data centers in Uppsala and Stockholm. The sharp operation of servers takes place in Stockholm, while for example backups are stored in Uppsala at a safe distance.
We have two power supplies for our equipment. Of course, we have battery backup and diesel generator, in case the power fails – even parts like that is at least doubled. So we have two power cords for each server that comes from different battery backups, with separate diesel generators (which in turn is doubled) which are connected to each switchgear in the electricity grid. Expensive and very, very good! Not even a wrecked diesel backup stops us.
In addition to the power supply, both centers have security guards 24/7. There is everything from climate systems with reserve and fire extinguishing to physical obstacles to sabotage. High quality in every step!
