Our Platform. Part 1: Servers
We’re often asked what servers we use and why we chose them above others. This is the first in a four part series discussing our technology choices; in this part we’ll look at the hardware/compute layer.
We wanted the very best hardware that money can buy and so decided on the Dell 12G range; products from this range have taken top slot in various review websites and won many awards.
Dell has come a long way in recent years in the hosting industry. All due to the creation of a specialist hosting division who asked service providers a whole bunch of questions but actually listened to the answers. We wanted density, we wanted power efficiency, we wanted performance and reliability and we needed all of this at a price point close to what we could build ourselves. Dell delivered in all respects.
Sometimes it’s the little things which make the big differences. Take the Dell Internal SD Module – this is an onboard controller which can take two enterprise grade SD cards and mirror the data between them. It is designed for embedding hypervisor software onto a server without the expense of raid cards, hard drives or SSD, whilst still providing redundancy against a failed storage medium. Absolutely fantastic and adds less than £40 to a server.
We have been seriously impressed by the Dell servers. They are incredibly well designed and engineered and a delight to work with. Performance is excellent, the power draw very respectable and we have had zero issues to date.
Why not use Supermicro?
Those who have followed and hosted with the founders over the years will remember that we were one of the biggest Supermicro based hosting companies in the UK. It might come as a surprise to learn that we’re not using Supermicro again; this is no reflection on the quality of Supermicro hardware but has a lot more to do with the type of business we are.
Supermicro is great kit, but it’s sold as a commodity to companies who want to build and maintain their own servers. If you were a hosting company wanting to provide dedicated servers and provide a guaranteed quick fix on any hardware issues then you’d need to hold your own stock of spare parts/servers since the mainstream providers can do 4 hours at best, which simply isn’t good enough when a customer has a server down. So Supermicro is a good fit for companies with lots of servers where spares and fix times are important.
We don’t have single points of failure in our infrastructure, whether it is servers, network or anything else. We run with so much redundancy that we can easily afford to lose a number of servers from a cluster without affecting quality of service in any way. Our model is also different in that our servers are extremely powerful and expensive and we need less of them. We also need the highest quality and efficiency possible, so for us it makes perfect sense to source hardware from an enterprise server provider like Dell or HP and take onsite warranties from those guys. We chose Dell as we felt they were more aligned to the needs of a service provider and are (in our experience) easier to deal with.
>> Part 1: Servers
>> Part 2: Network
>> Part 3: Software
>> Part 4: Storage