The beauty of a hybrid can often be in the eye of the beholder. If it’s tempting to try a sushi taco or a labradoodle makes you weak at the knees, it’s not for me to cast judgement. But as weird and wonderful as some hybrids can be, there are some environments where mixing the best of different breeds just works: IT infrastructure can be one of them.
Teaming the most effective elements of physical and virtual data centres with public and private cloud infrastructures can offer flexibility, agility and high performance. And it’s those benefits that are pushing more of us to opt for hybrid approaches.
In fact, Gartner analysts have predicted that by 2020, 90% of organisations will adopt hybrid infrastructure management capabilities: a clear signal of what’s in store for the IT infrastructure.
But along with the upsides promised by hybrid infrastructures, there are risks involved in combining elements that would normally work independently of each other. Slow-downs and outages, which become increasingly hard to prevent, have perhaps the most disruptive effect, ultimately costing the business customers and revenue. This threat is driving a change in the way IT teams need to manage themselves and their infrastructure.
Adapting to change
The interconnected and virtualised nature of the hybrid infrastructure means that applications can affect each other’s performance. A poorly performing storage solution for example could be the result of multiple, completely unrelated applications contending for the same resource (e.g. Array LUN).
And the result? IT teams are already transforming their approach to performance management. Instead of concentrating on the overall uptime and availability, they’re shifting their focus to the performance of the applications themselves.
To do this, they’re identifying the applications that are critical to the organisation, and understanding how each one fits into the IT ecosystem and their relative priority to the business.
>See also: Hybrid cloud: what goes where?
Armed with this knowledge it’s much easier to make informed decisions about where each application should run – whether it would perform better in a public or private cloud, or whether it would be wiser for it to sit in a physical data centre onsite. Increasingly, it’s a combination of all three.
Working more flexibly is a business mantra at the moment, and it’s never been truer for IT teams. They’re recognising the need to adapt easily to changing business priorities and adapt quickly if an application isn’t performing well.
That relies on monitoring both application and infrastructure performance and having the right systems in place to be able to move applications from the cloud to the data centre, or vice versa, if needed.
Availability over performance
Until now IT teams have demanded availability, but implementing a more app-centric infrastructure will bring performance sharply into focus. This is borne out by the increase in the number of performance monitoring tools at work across the IT infrastructure – in fact a recent Gartner report stated that most organisations host five or more infrastructure monitoring tools per data centre.
While these tools provide valuable insight into how the infrastructure is performing, their scope is often limited: they may be able to show a problem with a specific device, but they can’t correlate how that problem is affecting other components or the performance of the application itself. Having a partial view isn’t helpful. How can a problem be fixed if there’s no way of knowing why it’s happening or how it is affecting applications they support?
It’s this frustration, and the push for performance over availability, that will see a consolidation of performance monitoring tools, and a widening of their scope. Monitoring one element of the data centre in isolation does not give the level of insight needed.
As a result, a much more holistic approach is already being adopted: an app-centric emphasis that collects data from and analyses every element of the IT infrastructure, extending beyond the internal data centre and the private cloud to encompass the public cloud.
Pressure on vendors
Vendors too are having to adapt to the changing nature of the data centre. The emphasis on performance means that eventually public cloud providers in particular will be expected to provide performance-related service level agreements (SLAs) for response times.
Until now, they have been hesitant but demand from customers will mean that they have no option – and providers that offer the first SLAs will be in a much better position than those who continue to focus on availability SLAs over performance SLAs.
Performance monitoring vendors are also under pressure to provide a more comprehensive solution. Proprietary vendor-specific tools are not designed for the interconnected, hybrid data centres industries are all heading for.
IT teams need one solution that will monitor across the entire IT infrastructure. And they will vote with their feet. Vendors that are slow to respond to the massive changes we’re already seeing will suffer.
Making the leap
App-centric infrastructure performance monitoring for the hybrid data centre is on the rise, and looks set to be here for some time. The knock-on effects across the IT infrastructure are huge, not least when it comes to performance and performance management.
It may be quite a while before public and enterprise cloud providers will respond to that new focus. And when they do, the impact will be felt by IT teams, and their organisations, across the globe.
Sourced by Len Rosenthal, CMO, Virtual Instruments