Please note that we will be performing extended maintenance to replace / upgrade a number of ether switches in our London Telehouse facility over the Bank Holiday weekend and install updated firmware on existing switches.
The work will involve firmware upgrades and the implementation of a “Virtual Chassis” between the switches in both Telehouse North and Telehouse West to simplify configuration and improve backbone connectivity between switches (increasing to between 32Gbps and 84Gbps capacity between individual switches)
In total we will be installing 192Gig of new port capacity and a more than 1 Tbit of internal backbone capability (which in real terms means an 80 Gigabit backbone between sites with in-rack capacity of more than 400 Gigabit per rack – which should be enough for inter-VM and iSCSI communication for a while!)
We will also be taking this opportunity to update our internal documentation and replace some old patch cables where necessary (damaged / too long / too short / switch from copper to fibre, etc)
Whilst every effort will be made to keep disruption to a minimum, I do anticipate there will be brief periods over the weekend (most likely between midnight and 4am each night) where parts of your service may be disrupted briefly as connections are either migrated to the new switch fabric or the switch you are currently connected to is rebooted to load up to date firmware.
Customers with backup connections should notice the briefest of blips as traffic fails over to their backup connection. If you are connected via single uplink then you will obviously see a slightly longer (5-10 minute) outage as the work is undertaken to connect you to the new fabric
Should time and resources permit, we will also be upgrading the memory in one of our virtualisation hosts and adding 2 x 10Gig interfaces which we will then work to migrate the existing containers onto these 10Gig interfaces and away from the on board 2 x 1Gig NICs thus providing additional capacity for both iSCSI and realtime traffic.
Should the initial maintenance go well and time permit we are also looking to retire a number of edge routers and migrate their traffic onto the new core fabric which should also increase performance and further reduce latency.
We anticipate that legacy Edge routers in Telehouse North will be retired first – this will affect customers connected to LINX ConneXions through our Telehouse North facility, as well as customers connected via our BT Broadband service as well as leased lines delivered over Virgin, BT and/or Exponential-E into TeleHouse North as we move connections from the old Juniper J6350 and the remaining Mikrotik CCR routers (mainly experimental and most of which were phased out of the core last year) onto the new Juniper MX series hardware now that it has finally arrived, been configured and tested extensively over the last 2-3 months.
Information on the progress of the updates and details of any issues will be posted to https://status.fido.net on a regular basis over the weekend and we will be paying close attention to the support mailboxes in case of any issues.
If you have any questions or require any more specific information then please contact your account manager to discuss in the first instance, and of course I stand ready to answer any questions should you have any.
Jon Morby
Managing Director
FidoNet
5 comments for “AT RISK – Extended Network Maintenance 28th – 31st August 2015”