Interview: Simon Hill, IT director, Viridor

Simon Hill is IT director for waste management and recycling company Viridor, which has grown its operations to over 320 facilities across the UK by acquiring many of its smaller rivals.

However, while that expansion may have improved Viridor's market share, incorporating all those additional sites, ICT contracts, applications and other hardware/software resources into some form of cohesive infrastructure presented several problems. One of the biggest challenges Hill faced was how to guarantee application performance at remote sites with very limited bandwidth.

What application performance problems did Viridor face?

Hill: The problem was specifically with our weighbridge application, a third-party app that runs over RDP which we have since migrated to Citrix.

Because of the sites' geographical locations – lots are landfill sites in old quarries where few network carriers reach – they had just 256Kbit/s or even 128Kbit/s of bandwidth. Getting extra capacity out to them can be extremely expensive.

There were some sites with a weighbridge and only four or five people, meaning we couldn't justify the cost of a leased line.

For example, we were quoted a six-figure sum in terms of ancilliary charges for one of our sites in Devon.

The app could cope fine if it was just the staff on site using the bandwidth, but if that site hosted a meeting for several regional managers – where they would all connect their laptops to the network – the weighbridge app would stop working and vehicles could not be processed, meaning staff had to revert to manual methods.

How did the IT department solve those problems?

We needed to be able to protect critical apps when the network was being saturated – make sure HTTP traffic was prioritised lower down the stack – which is where Ipanema's platform came in, delivered as one half of Cable & Wireless's Application Performance Management service with Cisco's Wide Area File Service (WAFS) providing the other half.

We've just renewed our contract with C&W for five years, and Ipanema is rolled into the master service agreement, billed quarterly.

How is Ipanema's platform configured and what does it do?

The 130 sites are set up according to a typical hub and spoke topology with the datacentre in Taunton and all the remote sites connected into it via the Cable & Wireless Cloud. We have a physical resilient pair [of Ipanema devices] in the datacentre, but we are able to deploy remote engines to remote sites – there is no kit [dedicated hardware] there, so there are significant costs savings.

The virtual engine does not provide reporting to the same extent as a physical engine, but it is enough for us.

It means we have no software client working on site, just a central server configured to view all the network sub segments.

The software gives us a more granular view of the traffic on the network whereas before we had bandwidth statistics but we did not really know if there were performance issues on the network.

Now we can log into the software and see IP traffic flows and identify potential issues without requiring third-line support engineers to look at them.

What benefits have you noticed so far?

It is a continual improvement process. The Cisco WAFS piece does file caching, you leave it alone and it does its thing but with Ipanema we can refine it as we go along by altering specific parameters to provide the best end user experience, especially as you add new applications to the stack. At first we set fairly lenient server response times and bandwidth requirements for application sessions, reviewed on a monthly basis.