As you can imagine, when you have three local governments sharing a technology platform you need a pragmatic and flexible supplier. And because this is government, you also need security and rigour. Code Enigma are fully signed up G-Cloud framework suppliers for government, with services under consultancy, design, development, hosting, support and training, and we are also an ISO 27001 certified service provider at every level (the only Drupal specialist with these credentials in G-Cloud, at time of sale). So we were an obvious invitee when it came to going out to tender for the project.
Because of our relationship with Rackspace, we were able to meet colocation requirements, queries about specialist types of VPN connection, we could order in additional equipment to meet specific needs, nothing in the combined requirements document phased our delivery consultants. We could even wire in a direct connection, site to site fibre, if someone needed it.
Couple this with the naturally flexible nature of our service - our "ask us anything support" and our "no one size fits all" belief when it comes to building out a new customer's service - and we are a perfect fit for an organisation (or organisations!) with demanding requirements and an ever-changing digital presense.
Although the consortium, known as Tri-Borough, decided to purchase one of our top of the line layouts, because the cost is split three ways it still represents a significant cost-saving for each borough individually. So not only did they save money overall, they also gained big with our service levels.
Firstly, there's our VMware private cloud, which provides 100% uptime guarantees on both virtual hardware and network. This is thanks to a number of factors. Our infrastructure partners, Rackspace, have multiple backbone providers and a dedicated 24/7/365 networking team keeping things going. We have doubled up on all cabling and switching, we have a pair of highly available Cisco ASA security appliances in front of everything and we have VMware DRS, which means even if a physical server fails, your virtual servers will keep on working.
At a virtual machine level we have a multiple server layout to ensure there is no single point of failure anywhere in the system. This is important, because while we can effectively guarantee an individual server won't stop performing, we cannot offer the same assurance an application won't crash. For example, if your web server software crashes your server might still be up, but it won't be serving your website. By having multiple web servers, we ensure things keep running if one of them gets into difficulties.
And of course there's our rigorous monitoring. We use a system called Nagios, with multiple endpoints both at AWS and within our own datacenter, to monitor everything imaginable! Server clock, website pages, database response times, Drupal update status, number of connections, DNS capability, disk space, etc. etc. etc. all monitored 24/7/365 with realtime alerts going out to our engineers the moment there's a problem, by chat, SMS, email and smartphone push notifications. We also use a service called Pingdom to keep an eye on things for us, as it is outside of our networks and checks customer websites continuously from different endpoints around the world.
On top of all this, we're supporting three different sets of development teams. RBKC and Westminster both developed their websites internally, and Westminster still partner with another Drupal company, Ixis, for Drupal support. While LBHF are having their site built by another external agency, SwiftCircle, who work very closely with us on a number of delivery projects. So our flexible deployment tools and fully hosted version control, plus the support we give to these teams to help them continuously deliver improvements, is an important part of our work.
The icing on the cake? Westminster told us their site is a good deal quicker with Code Enigma than it ever was before (they've been measuring it ever since it launched). We like that!