As you know we are midst our large infrastructure project and I want to give you an update as to where we are and how progress is going.
In late August, we had new servers, storage and networking installed into our new data centre, and have since commissioned these servers and the new network, and completed formal hand over of the network infrastructure our network operators.
The first priority for servers was getting them ready for our new Parallels hosting system, which will allow us to offer new services like hosted Exchange, SharePoint and Lync AKA “Office 365”. These are now being configured and built by Parallels with a view to launching services in calendar Q2 2013.
The second priority has been preparing a new virtualisation environment, in which we will run our existing VoIP service & existing hosting service. We will complete the hand over into production for servers next week, with a view to starting migration of the virtual machines from 22nd October.
In the New Year, we will then deploy the rebuilt version of our VoIP platform migrate our existing hosting service to the new Parallels hosting service and finally, decommission old equipment.
There are lots of moving parts in the project and as a result, there has and will continue to be a cycle of public activity, followed by apparent periods of in-action – whilst the next phase of behind the scenes work completes. We are very conscious of not making multiple changes at once, so apologies if it seems like it’s taking a long time.
We have been keeping a close eye on quality so as to ensure that customers will continue to receive an excellent level of service. Risk is being proactively managed throughout so that the number of unwelcome surprises is minimised.
We have begun to see a number of benefits in the new platform already and are confident that as we begin moving services, customers will also start to feel the improvements.
Firstly, as befits a new & cleanly installed platform, the quality of design & documentation is much higher; and a much larger, more structured team has been responsible for its implementation. There have been clear & specific hand overs from suppliers, into our engineering teams and now into our 24×7 operations team.
Secondly, all the equipment is proactively managed, with a 4 hour time-to-fix SLA. We have seen some initial promising signs of the improvements to stability that this will bring.
Finally, we are experiencing a significant speed and capacity boost. We are running HP Gen8 blades which have significant amounts of computing power, on a 10 gig network. However, the biggest speed increase is in disk storage, where our 3PAR storage system is supplying the most requested files from flash memory rather than disk. These speed/capacity improvements will, we believe, translate into a reduction in bottlenecks during busy periods.
Virtual Machine Migration:
The next major step, starting 22nd October for a period of up to 8 weeks, is the migration of virtual machines (VMs) from their current hosts to the new environment in. These VMs provide all our services – DNS, billing, sip services, email, hosting, etc.
The services are all clustered with multiple VMs operating a specific service. So, rather than order the migration by service, we have instead risk assessed all the services and their components and will be conducting the migration in order of risk – lowest first!
We do perform quite a number of VM migrations on a routine basis, and are well practiced. Prior to this migration we are ensuring the process is fully documented and are completing dry-runs. Throughout this work, we will ensure that any disruptions are minimal and take place when they will affect the least possible number of customers. Higher impact work will take place at weekends and towards the end of the migration period.
Whenever work is taking place that will affect our service, we will issue Gradwell status updates in a timely fashion to ensure that you are aware of possible interruptions.
VoIP Platform rebuild:
The VoIP team has separately been engaged in an 18 month process of re-engineering our VoIP platform, to improve its robustness and design. This has historically been known as the “Manchester project”.
In Manchester, we have been testing SIP and IAX outbound trunks, which has gone well. We have also now completed the installation of our “Hosted PBX” product, which will shortly be ready for testing and safe launch.
It is our intention to migrate all VoIP customers to operate on the Manchester based servers early next year, before finally bringing our “New London” servers online and moving to a dual-site mode of operation.
Broadband/Connectivity: We are also pending installations for new interconnects with TalkTalk and BT, to re-provision our broadband and EFM/Ethernet interconnects. Once these are in, we will announce plans to migrate broadband customers to new, redundant, higher capacity connections. This is likely to happen in January 2013.
I’d like to add my thanks to all our customers and partners for your continued business, my team and I are very focussed putting these tools into action and we are looking forward to providing you with even higher levels of service – focusing on easier to use systems, greater capacity and a refreshed product offering in all areas.