Three major run-time performance hurdles to avoid

Network Function VirtualisationEvery investment in major IT transformation arrives with its own set of structural and institutional challenges, not least migrating to a virtualized infrastructure. Get it wrong from the outset and you will be faced with the mother of all headaches. If the issue of performance is not addressed at inception, your data centre and its services – although being able to run – will be increasingly plagued by service problems, poor user satisfaction and inadequate Return on Investment. So, companies looking to transform their virtualised infrastructure should look before they leap and consider three major hurdles.

#1: The heavy lifting of remedial actions

Shift happens. Seasonal traffic spikes, buggy drivers, fluctuating head counts, updated applications, worn out drives and a thousand other factors all contribute to the constantly shifting foundation upon which your infrastructure must stand and thrive. But in any dynamic environment, the unforeseen throttle will inevitably occur and issues will need fixing.

End-user issue remediation actions are exceedingly important, so when a VDI (Virtual Desktop Interface) user calls or files a trouble ticket, IT must have the tools for translating complaints into troubleshooting. Pinpointing problem sources manually can take hours; sometimes even days or weeks, because virtual infrastructure tools tend to be designed for high-level observation, not granular examination of what exactly is, or isn’t, going on. If multiple problems crop up from different sources, the burden of remediation magnifies, potentially risking an organisation’s business continuity.

#2: Isolated end-user experience metrics

Amazon famously publicised how each 1/10th of a second of added site latency sacrificed 1% in sales. Nothing will persuade a worker not to use a cloud-based software tool, particularly on a new rollout, quite like staring through a login hourglass. Of course, management feels the same pain. Aberdeen Research found that poor application performance can slice up to 9% from corporate revenue.

Citrix proposes that there are five key, measurable metrics that constitute the lion’s share of poor user experiences within virtualised environments:

  • Launch time
  • Logon time
  • Load time
  • Latency
  • Operations such as printing, screen updates, etc.

Clearly, most issues will stem from obstructions in resource flow, but determining the root cause or causes of any decline in these metrics can be arduous.

#3: Inefficient allocation of IT resources

Virtual infrastructure resource utilisation wastes mountains of money when it runs too cold and throws up bottlenecks and failures when it runs too hot. Efficiency is the relationship between performance and resource utilisation and maximum efficiency does not mean maximum utilisation. The physical underpinnings of your virtual infrastructure are a mesh of CPU, memory, storage, network and – particularly in VDI and HPC deployments – GPU (Graphics Processing Unit) assets. The key is to deduce not only utilisation levels but overall efficiency within each resource type, network region and entire organisation.

The solution for companies looking to jump these hurdles is to invest in better predictive analytics-driven strategic levers. They enable the reduction or even elimination of persistent IT problems. To prevent the heavy lifting of remedial actions the tool needs to increase real-time prevention of problems and in cases when prevention is not enough, step in with automated remediation. Either way, the end result is massive savings in resolution time as well as a much improved chance for IT, and thus the whole organisation, to meet its SLAs (Service-Level Agreements). The ability to help pinpoint the problem’s source and whenever possible, cross-silo insights for balancing load optimisation to alleviate bottlenecks is crucial to prevent isolated end-user experience metrics.

In order to combat inefficient allocation of IT resources you must be able to monitor the real-time characteristics of these physical resources and how end-points and applications are using them across the network. Clarity of vision is a strategic lever not only to utilisation levels but overall efficiency within each resource type, network region and entire organisation.

Written by Atchison Frazer, CMO at Xangati