Managed Server Hosting

Case Study

Background

Lynx Internet Solutions is a leading supplier of e-commerce solutions with over 1600 clients. Their flagship product, Evolve, is targeted at companies that need high performance software and exceptional support to keep their sites trading 24 x 7. In their tenth year of trading, Lynx really understand the requirements of their customers, which include some of the busiest e-commerce sites in the world. To keep control of the entire solution their policy had always been to host and manage their own servers in London where they are based. By 2006 they had established a sizable server farm in Redbus and found that a considerable amount of their time was spent maintaining, replacing and upgrading servers. Their customers were growing and ordering new developments to their sites. However, whilst this generated more revenue it also brought with it new challenges.

"We manage some of the busiest ecommerce sites in the world and the global nature of our clients means that the concept of a planned maintenance window is rapidly disappearing. Griffin's Managed Hosting solution solved all of our problems and whilst it was initially a big decision to outsource a vital part of the business, our hosting costs have reduced and we actually feel we have more control now than when we ran our own server farm".

Paul Meyers - Managing Director, Lynx Internet Solutions

The Challenge

Each time a customer needed more disk space they had to carefully plan the project in advance. The customer's site had to be taken down at the quietest time and a new hard disk built, tested and installed. As the world became more used to trading online, e-commerce volumes on some customer sites grew unpredictably, making it hard to plan upgrades. Plus the global nature of Lynx's customers meant that quiet periods were increasingly hard to find. The alternative was to over-provide for disk space but this was just not economically viable.

The more changes customers wanted to their sites, the higher the chance of bugs and corrupted data. Lynx have a comprehensive development and test environment but innovation in software brings challenges. They had on-line and tape back-up solutions, but it could still take several hours to roll-back to uncorrupted versions of data and software when problems occurred. Often when they did restore a customers system it meant overwriting corrupted data and inevitably losing transactions taken on the site after the corruption occurred.

Finally as servers age, they start to fail - the hard drives are often the first to go and it seems, all at the same time. Naturally Lynx had back-up servers, but failures meant staff needed to attend site at any time of the day or night and inevitable downtime for the customer. Furthermore, recent terrorist attacks had made them nervous about hosting all of their sites in Central London and to cap it all prices for hosting in London were going up dramatically.

Lynx realized that hosting was not their core business and was taking management focus away from their key differentiator - their lead in the development and implementation of e-commerce solutions. Something had to change.

The Solution

Lynx decided to host their customer sites on Griffin's IPSAN network choosing our Data Centre in Derby as the primary location. Built around EqualLogic's ISCSI IPSAN technology, Griffin's Managed Hosting service offers 100% uptime through a RAID 50 disk array. Upgrades to storage for individual customers can be done through a web portal in minutes, with no downtime and the storage capacity of each IPSAN is practically limitless. Upgrades to storage are on a pay-as-you-go basis, so you only need to buy as much as you are going to use.

The IPSAN boots the server. Everything is stored on the SAN and so servers are disk-less meaning there is less to go wrong. In the unlikely event that a customer's server does 'blue screen' they are all fitted with remote access cards as standard, which can be accessed even if the server is dead. If the server cannot be revived then all customers have access to a shared hot standby server as standard. For the really paranoid they can even opt for a dedicated standby server - attached to a SAN in a different Griffin Data Centre.

The Benefits

Probably the biggest attraction of the Griffin solution to Lynx was the ability to recover from failure in minutes rather than hours or days. The unique 'Snapshot' ability of the EqualLogic SAN means that Lynx can now restore a customer's data and programs in minutes from any one of up to 255 snapshots taken hourly, daily, weekly or monthly. Furthermore restoration to another live volume leaves the corrupted volume intact, meaning that Lynx can still recover transactions in the corrupted volume for their customers.

Lynx felt more comfortable with hosting their business in Derby. The Data Centre sits on a self healing MPLS ring linking other Data Centres in Leicester and London. The network is fully redundant such that there is no single point of failure and like the Managed Hosting solution the network carries an SLA of 100%.

"Lynx are an ideal customer for Griffin and we have enjoyed working with them. For customers that cannot be offline for more than a few minutes our Managed Hosting product is really the only affordable alternative to traditional dedicated hosting and back-up solutions that are increasingly being shown to be inadequate for high-end customers".

Andrew Dickinson - Sales and Marketing Director, Griffin Internet