About the Company
On behalf of 22 of the UK’s largest banks and building societies, Link Interchange Network operates the busiest shared ATM transaction switch in the world, connecting more than 58,000 UK cash machines and 100 million Link-enabled cards worldwide.
With more than 20 years’ experience in real-time payments, the Link Interchange Network has become a fundamental part of the country’s payments infrastructure, supporting a number of payment and settlement processes.
At peak times, the Link switching system has to support up to one million transactions per hour, reconciling two billion ATM transactions annually for UK financial institutions, independent ATM operators, and mobile network operators.
In addition to direct payment transactions, Link also provides a range of outsourced payment processing services via its trading division, Avantra.
Wearing a number of different hats, Alan Gilmour, Link’s head of service delivery, is responsible for assuring the quality of the company’s technology offerings, including the vital service and support capabilities provided to Link’s banking customers.
With a view to creating a more transparent relationship between business services and their supporting IT functions, Gilmour has been working with Hewlett-Packard to implement its OpenView enterprise IT management suite. That forms part of a wider, ongoing project to provide visibility into Link’s IT functions.
Information Age (IA): Link is one of the backbones of the UK banking system. How has it gone about ensuring the reliability and quality of the vast number of payment processing services it operates?
Alan Gilmour (AG): In terms of the core services that Link delivers, there has always been total clarity of understanding as to how each part of the enterprise infrastructure might impact the service provided to our clients. However, there were some peripheral technologies – some of which were less well understood – that on occasion could have an impact [on the business] that wasn’t obvious.
So we recognised we had a disconnect between some of the technology and our visibility of the impact that it might have on the products and services that we deliver.
IA: How have you tried to overcome this disconnect?
AG: Historically, we’ve been a user of Hewlett-Packard’s Integrity Non-Stop [fault-tolerant] systems, with a bit of Wintel. But we foresaw an increase in the different types of technologies coming at us, technologies that from an operations point of view were not our core competency. We realised we’d have to demystify these technologies and to connect the technology to something that [would show] services impact.
So, what we’ve been trying to achieve is [a situation where] we’re able to represent the component parts of our environment – whether hardware, software or comms – and monitor and ‘redline’ exceptions as and when they occur. And to translate that into a business perspective. This would then enable us to understand the severity and business impact [of any event], and then to prioritise the response, to focus our resources.
That was the vision we had two years ago. We set off on an initiative to validate whether that was feasible and to decide how to go about delivering it.
IA: What kind of disciplines and methodologies did you set in place to try to ensure the project was effective?
AG: We were very explicit about what we wanted to achieve. When we went out to reference sites and met users of the main products in this area [of business service management], it became very evident that they had started off in a very similar way to us, that is when something failed they wanted to understand what it meant from a business point of view, to prioritise and put people in the right place to do the right thing at the right time.
Unfortunately, with some of the examples we saw, what was delivered was a number of very niche solutions – solutions that will give you alarms, but no context and no prioritisation. So when the board gave us the authority and the money to bring this in, we were set the challenge: how will we know when this has been a success?
For a software implementation, unless you are undertaking to do a very specific thing, that is quite a challenge. Our approach was to create ‘service trees’.
[At the highest level] these would enable us to view the environment in the context of the business products, then in terms of the services that deliver the products, and then in terms of the component parts that deliver the services. The thinking was that if a component [fails], the tree is destroyed and you get a clear view of the business impact.
By considering everything [in service delivery] we did from that viewpoint, when we considered the use of enterprise IT management products we could say: “Right, now we’ve got awareness of our environment. When certain components [fail] that’s a big red alarm, and [in some cases] we don’t care as much because we have resiliency built in and our response will be appropriate.”
We bought Hewlett-Packard’s OpenView with this in mind, with a clear view of the end game and an understanding that we could get a lot of value out of the product from the beginning of the implementation.
It might seem like obvious stuff, but among the other users of management products we spoke to, there were a lot of examples where that just hadn’t taken place. Most were still happy with the outcome and with the products they’d chosen, but we felt they weren’t generating the full business value. Most were more alerting systems than environments that would help you assess customer impact and prioritise.
IA: So assuming that you are generating real business value, how are you measuring it?
AG: The system is based on key performance indicators and on our service level commitments to our customers. Bearing in mind Link’s systems sit in the middle
of all the banks and building society [networks], we’ve got to make sure – at the core level – that our availability is there at all times. So what we’ve got – with the environment we’re putting in place – is an understanding that if something goes wrong with the infrastructure that sits behind all this, if an exception occurs, it [starts] counting down on availability, so we understand what our availability is to our customers at all times.
A further planned phase involves making that dynamic: so, for example, we have an understanding that the availability requirement of a particular customer is X, relative to what we’ve committed to on the service level agreement. If a problem occurs today that starts to take our availability down further, it starts to count back. When it gets to the point where we’re into the orange (early warning stage), we know this customer is one to home in on.
IA: It’s often hard for the business to understand the benefits of business service projects. How did you persuade the board to fund the project?
AG: That was quite difficult. At a core system level, we are into our 8th year of 100% availability, and we hadn’t had anything go wrong of any major size. So it was a challenge to justify this investment solely in terms of availability.
In the end it was justified on the basis that, as an organisation, we needed to eliminate the barriers to entry of any new technology that we wanted to introduce going forward. We were mainly focused on our core products but some of our peripheral products were becoming more mainstream and so their importance was increasing. So by delivering the ability to look at more [environments] and prioritise better, in theory we’d get some payback in terms of the optimisation of the resources that we’d already got.
Additionally, we couldn’t realistically maintain everybody with their existing, stovepiped skills set. To support a more diverse enterprise, they had to, in a sense, de-skill, so that on the front line they could deal with incidents rather than technologies. Of course, we had to give everyone a baseline knowledge of technology but we couldn’t carry on with the approach of everyone being an expert in every area. At the [senior management] level it was quite well acknowledged that if we wanted to do more and do more with the same kind of cost base, we have got to invest.