Accelerating Virtualisation

One of Data^#3’s Virtualisation experts Victor Medici poses the question…

“Can we get more out of virtualisation sooner?”

If virtualisation makes (much) better use of datacentre infrastructure, and (greatly) improves staff productivity, why are these principles not embedded in the design of virtual infrastructure too?

Think about your experience with virtualisation. Could your business have realised the benefits of virtualisation (a lot) sooner?

If so, the lifetime value of your virtual environment would be even greater.

That was the goal with our 4corners Virtualisation Architecture – how can we increase the value of virtualisation. We set out to create an architecture that
could be used as a base to design – and implement – virtualisation solutions faster, using less infrastructure.

To build 4corners, we took the common functional operational elements used in almost every virtualisation solution we’ve implemented in the last decade – servers, storage, and networks – and investigated how we could accelerate the value that virtualisation can deliver.

In an ideal world, for maximum efficiency (and simplicity), there would be one of each subsystem. In the real world, two issues immediately come to mind: capacity, and more importantly, risk.

Capacity

Capacity is only an issue if the choice of platform is restricted, by application or even simply by preference. Single platforms are well capable of handling massive workloads. For example, IBM’s current UNIX systems perform over 6 million transactions per minute, and its Intel Xeon-based systems over 1.2 million. Interestingly, the more work the platform supports, the more efficient (and therefore cost-effective) it becomes – an excellent example of economies of scale in the real world.

Running all your workloads on one subsystem is possible, and redundancy within the platform further reduces the impact of most unplanned outages. Statistically, fewer platforms and components reduce risk. With the increased reliability of each generation, the case for a
single platform is increasing.

However, not all unplanned outages can be anticipated and designed out. Certain planned maintenance tasks still require a platform outage. Two platforms are therefore factored in to the architecture. Interestingly, this is also the architecture adopted by the majority of
disk array vendors – two controllers.

Unified Network

Unified network switches have been introduced that support both TCP/IP and Fibre Channel protocols. Switch count and adapters can be minimised with the ability to guarantee the delivery of 10Gb Ethernet packets simultaneously. This reduces risk, increases utilisation of the network (efficiency) and simplifies the design and deployment of networks.

Management

Enhanced management is critical to improve productivity. Much like an organisational structure, management is a distinct element within the architecture, and is typically deployed on a separate platform, to reduce dependency risk such as a failed disk array.

The architecture therefore consists of management, hosting servers, shared storage and unified networks to connect the elements together – the essence of 4corners architecture.

The 4corners architecture does not dictate the type of technology used, nor the vendor – that is the function of the design.

Throughout the process, we have distilled our best practices into a repeatable implementation plan, focussed on rapid design deployment. With well over 100 successful designs and associated implementations on the 4corners architecture, our implementations have been delivered in a shorter period of time to full production, using simply less infrastructure, with lower
risk, than designs we have competed with.

Want to learn more about how Data^#3’s 4corners Virtualisation Architecture can deliver faster time to value for organisation, contact your Data^#3 Account Manager, or give us a call on 1300 23 28 23.