Beyond simple virtualisation

The journey into cloud computing needs a road map to identify the plethora of issues and show the way forward

0ne of the curious things about IT is that the simpler we try to make things, the more we seem to complicate them.

Take for example cloud computing - it's meant to simplify and streamline IT, yet it's near impossible to find two explanations that actually match.

Perhaps we need to quickly review where we are, and what is important on the journey into the cloud, as there is still a way to go.

While many large organisations are already well into virtualisation projects, it's important to note that the real opportunities in virtualisation are only just beginning to be understood.

It is the management opportunities around virtual infrastructure that need exploring, as this is where the richest set of services work lies.

We have moved beyond the notion of fixed function servers with operating systems and applications built for specific tasks.

Storage and network visualisation with VSANs and VLANs reduces cost and improves availability, but it does add to the complexity.

Hypervisors virtualise the operating system and allow us to improve utilisation, dynamically reconfiguring the data centre to better meet the needs of the business.

This explosion of virtual components constitutes one of the biggest management challenges for coming years.

As such, IT managers are looking for tools that will help staff members regain control of their virtual infrastructures.

"Orchestration" is often used when talking about the optimisation of virtual infrastructure management.

An application owner should not care if the service is delivered via a virtual machine, as long as the service level objectives are met. Does the infrastructure manager need to be involved each time a business service is rolled out?

The application owner should be able to select what is required in the data centre, rolling out the related virtual machines from templates, automatically - or manually, by VM operators.

Operations staff should be able to use the same tools to restart a virtual machine, regardless of hypervisor.

So it's safe to say orchestration is about more than implementing a few simple run-book routines: it's about streamlining and speeding up the business-service provisioning processes via self-service and approval-based work flows.

It's about transparently providing heterogeneity across all industry-leading hypervisors, and it's also about automating the processes that determine how, where and to whom resources are allocated.

We rarely wake up in the morning and think about a new operating system - OK, some of us might - but we are usually concerned with finding solutions for our current set of issues.

The application layer is where the real complexities lie and also where most of the solutions can be found.

To make the most of the new dynamic infrastructure, we have applications that can also be virtualised, as demonstrated by Services Oriented Architecture (SOA).

Services can be constructed from existing applications to cope with new business processes, re-purposing existing applications.

With virtual storage, virtual machines and potentially virtual applications, we have a very dynamic environment. This gives the freedom to make the best use of resources, providing we know what is actually happening.

So to make the promise of cloud computing more than simply marketing spin, we need a method of integrating orchestration with the application layer.

Business System Management (BSM) is a discipline focused on meeting the business needs of an organisation by understanding what is happening at an application level.

BSM maps your data centre assets to your business services, typically via a configuration management database (CMDB).

Every component that plays a role in the delivery of the service, from a virtual network switch to a database in a virtual machine, is linked to the service itself.

It is the complexity that will give these new environments the power to align applications to infrastructure in a dynamic fashion, and it will be the automated policies that tame the complexity.

Once orchestration has been applied in a virtual infrastructure, you can increase the resilience of your applications with BSM.

Conclusion

The cloud's real purpose is to provider a wrapper to hide all of this complexity. It should be a super set of technology, encompassing all technology, rather than just a vendor-specific implementation. Think of it as more of a universal ethernet connection rather than a power connector with all those international adapters to make it truly universal.

When you evaluate virtual infrastructure management tools, keep in mind that virtualisation is not the end game, but just one step.

It's important to remember that in the end, everything revolves around business service efficiency. Fast service delivery is one thing, continuous visibility into your running services and fast and automatic remediation of anomalies is another.

All of these elements are essential to get the most of your virtualisation investments. Once you have these elements in place, your Private Cloud can emerge and one day perhaps even link up to those "promised" Public Clouds.