As a company we are growing fast and that means we’re bringing lots of new members into the team. We’re determined to ensure that our new starters feel part of our culture straight away and have the passion for the services we provide – whatever department they work in.
To enable this we run a two day induction process for all our new starters, and welcome in presenters from across the business to explain what they do. The one rule? No slides!
One of the most popular sessions is run by our head of operations – Nigel Wareing. Nigel is responsible for keeping our platform performing and stable and we think its essential all our new starters understand how the ‘service’ in ‘software-as-a-service’ is maintained.
As most of the attendees are non-technical Nigel uses cardboard boxes to represent various aspects of our datacentres starting off with a pair of boxes representing our redundant telephony servers. These boxes handle the telephone calls on behalf of our customers and interact with our core call centre application. Nigel attaches a coloured tape to each box to represent multiple communications providers coming into each datacentre.
In between the two telephony cardboard boxes Nigel then places another ‘storage’ box to represent the call recording archive we provide both in our datacentre and the replication to a third party for disaster recovery.
Now Nigel adds a wooden plank across the top of the three boxes and explains that this represents a VMWare layer that allows us toseparatethe running of the applications from the physical hardware. For the non-technical members of the team this was a really useful introduction into virtualisation and inspired a number of questions!
On top of the wooden plank Nigel now adds a pair of redundant database boxes explaining that this is where all the information about our customers, their agents, their call plans, and the current status of each agent is held. In between these boxes Nigel adds a monitoring box to represent all the automated tests we run on each datacentre to alert us on any dip in performance before customers experience any issues. These tests are used internally, but additionally the results are also made publically available on the Trust page on our corporate web site.
Another wooden plank is added on to these three, on which Nigel places two web farm cardboard boxes. Each of these web farms contains multiple servers and runs our core call centre software. Each datacentre will have a production web farm and a secondary web farm. This enables us to load up the next version of code and seamlessly switch customers across without them seeing any break in service.
The final box Nigel adds is a firewall box to represent the security that sits in front of our web farms. Nigel then adds one of our customers to the top of the entire stack, and attaches two coloured tapes from the customer into the firewall. These represent the multiple data connections into our data centre.
The finale is to demonstrate why we have this level of redundancy, and that is by asking team members to remove any of the boxes – a web farm, or a telephony server. Because of the wooden planks the entire structure stays solid and in real life would give our operations team time to replace kit without the data centre being compromised.
In addition to this, because we run multiple datacentres even if an entire datacentre was affected customers would still be protected and receiving calls. As Nigel says “we make sure there are no single points of failure – something that would be very costly for a customer to achieve on their own.”
Our new starters completed what could have been a very technical session with a fantastic understanding of how we build a datacentre, and the cardboard boxes have meant they are able to relay this story to other colleagues, customers and partners.
We hope this has been of use and helped you to understand what goes into a NewVoiceMedia datacentre. If you have any questions for Nigel just let us know in the comments section below.