Mark van Bavel, Dimension Data's CIO, explains how Dimension Data is leveraging consumption economics.

Consumption model enables Dimension Data to streamline and standardise its operations ... and deliver greater business value to its clients.

Global and multinational clients represent a significant percentage of Dimension Data's client base, so it's important that we're able to deliver a consistent client experience across the globe. We've shaped our business to focus on emerging technology and innovation and invested significantly in IT outsourcing, cloud and managed services ... all of which are borderless solutions and services. The foundation for successfully delivering such solutions and services - from anywhere, to anywhere - is a standard set of processes and services, easily implemented in any part of the world, with any partner.

Our Group Information Systems team is currently running a project to standardise our global platforms. Initially we're focusing on SAP and other tier one applications such as IT service management. By aligning our systems globally we're positioning ourselves to augment our ability to deliver client value in new ways: not only will we enhance our clients' experience in interacting with us, we'll also gain a competitive edge ... having the ability to do things like pull together complex global quotes in a couple of days or deploy a new service across the globe in a matter of months will make us more agile and responsive. The project will also have a number of internal benefits: we'll save costs and offer our employees the opportunity to be more productive and efficient, and collaborate more effectively.

Phased approach

We're rolling out the project in a phased approach across the globe. With each additional region comes the need for additional capacity. Initially, we considered purchasing all the hardware we thought we'd need for the project upfront. The problem there was that during the initial stages we'd only be making use of a small percentage of these assets, while the rest lay idle. We'd only achieve full utilisation of our new assets after three years. It became clear that leveraging the public and managed hosting cloud capabilities of our own Cloud Solutions Business Unit (CSBU) rather than hosting the ERP systems within our own data centres made more sense.

In taking this step, we eliminated a steep upfront capex bill and were able to convert much of what would have been ongoing costs into opex expenses. We can adjust our utilisation of resources in line with the fluctuating requirements of different stages of the project. For example, we can spin up multiple development and QA instances for different test purposes ... and then shut them down when they're not in use. This approach also gives us automatic access to 24/7 operations with mature run-books - years of experience with hundreds of clients have enabled our CSBU to optimise its operational infrastructure support. If Group Information Systems had to put in place an operations management capability to support the global SAP platform around the clock we'd have needed to hire several new specialists, as we currently run an 8/5 operation for SAP infrastructure support internally.

Not only has using a consumption-based cloud model made the new platform less costly and time-consuming to build, test and deploy, but the ongoing management of the platform will also be less onerous for our internal team than it would've been if we'd built the platform ourselves. Our CSBU's services are fully managed, so we've effectively outsourced the infrastructure management aspect entirely. Our internal resources can focus on strategic projects and architecture.

Disaster recovery

Disaster recovery is another area where this approach has the potential to add value and save money. Physical data centres are at the mercy of natural disasters such as earthquakes and tsunamis. Dimension Data has experienced four such events in its data centres over the past three years. When the unexpected happens and a data centre goes down, providers are expected to bring the same services up somewhere else, immediately. This means that capacity for disaster recovery must be set aside − and paid for − at a second location, even though it's not being used most of the time ... this represents a significant cost for most businesses. Because we're running our SAP environment in the cloud and can spin capacity up and down as we require, we're able to keep our disaster recovery images on disk and we're not consuming RAM and CPU. We only spin up capacity once a month for a day or two to synchronise the images. This has the potential to reduce our cost of disaster recovery by 80%.

While the benefits of moving from a traditional to a consumption-based model to deliver IT services are compelling, there are potential risks and pitfalls. Developing a strong business case by leveraging the opportunities of cloud and advanced managed services upfront and effective change management are important elements to ensuring success.

IT leaders should be prepared to deal with initial scepticism and resistance from some stakeholders as well as concerns regarding security, connectivity and performance. One often falls into the trap of comparing 'like for like' (i.e. your existing static environment and taking the same over into the cloud environment for cost comparisons); the elasticity of cloud infrastructure means you need to assess your needs quite differently.

That's why it's important that they select their cloud provider with care. There also needs to be a strong focus on timely and effective communication, training, and engagement with affected employees throughout the transition.