Unrivaled Disruption (and Opportunity)
Healthcare IT organizations have never been under more pressure. Changes in the economics of healthcare have sparked disruption that feels like it is unrivaled by any other industry. Mergers here, acquisitions there, expansions so rapid that soon they’ll be everywhere. What are infrastructure teams to do?
I see more large-scale migrations and consolidations in healthcare than any other customer vertical I work with. Stressed out teams are asked to do the proverbial “more with less,” without any idea of what sorts of additional responsibilities they’ll be asked to handle the next day. Beyond the stress associated with these projects, these teams are working in an industry where performance is perhaps more important than any other.
Challenge 1: Rapid Acquisitions (and Workload Growth) Led to EHR Performance Problems
Recently, I’ve been engaged with an infrastructure team at a large healthcare provider in the southeastern United States. This team’s organization has been acquiring smaller hospital systems at an enormous rate, and they have been asked to onboard new workloads at a pace that far exceeds their headcount growth. Beyond that, with the acquisitions, the team has been asked to cut down on their hardware and hypervisor spend.
Although the team had these enormous deliverables in front of them, they had a more specific and immediate issue when they originally engaged with us. In their large EHR environment, they were experiencing tremendous performance pain. Ultimately, this was the challenge that was prohibiting their ability to scale. Their suspicion was that the performance challenges could be attributed to ready queue challenges, but their current alerting, anomaly detection, and problem identification tooling was insufficient to tackle the performance problem. Those tools gave them plenty of data, but no reference point as to what a good state would be in the cluster, only individual recommendations around potential remediation. The problem was only compounded by specific resource requirements from the EHR vendor leading to an inability to understand how VMs they couldn’t resize affected their ability to deliver performance.
Quickly, VMTurbo confirmed the primary challenge was ready queue contention. Through placement, though, and sizing of some non-vendor specified VMs, VMTurbo was able to mitigate the ready queue risk. The team saw a marked decrease in peak wait times throughout the day. More importantly than any resource usage in the environment, the team saw significantly reduced complaints about service degradation from the medical end-users.
(Check out this Principled Tech study, which found that VMTurbo improves performance by 37%.)
Challenge 2: Scaling a Multi-Hypervisor Environment
With their worst performance challenge addressed, and the teams’ customers satisfied, they could finally address the question of scaling. Ultimately, rapidly scaling an infrastructure can never be solved with a silver bullet.
For the last month, we’ve worked with the team on several solutions. First of all, after the team realized the value of application performance control, they were able to logically merge several vSphere clusters that they had segmented for the simplicity of management. With a few clicks in VMTurbo, four 6-node production clusters became one 24-node cluster, which then became an 18-node cluster with additional performance and efficiency gains realized through broader resource sharing.
Beyond that, though, scaling challenges still existed. The thought of a mixed hypervisor environment had been considered for a long time at the team’s organization, but they had never executed due to concerns around the time necessary to manage performance. By replacing their traditional approach of using manual data analysis, with continuous and real time decision automation through VMTurbo, they were able to finally execute on a Hyper-V environment at scale, significantly cutting out software licensing costs.
For Healthcare IT, Doing “More with Less” is the New Reality
Healthcare is a changing landscape. Healthcare IT organizations have to do more with less. They can’t move on anything until they cut out the manual time spent making performance decisions in the data center. Without control and automated performance assurance, planning for migrations is tremendously difficult. With the pace of change, it often seem initially like its more trouble than it’s worth introducing a software component that changes the way you run operations. What our healthcare customers see, though, is value within an hour of downloading the software, allowing them to focus on the higher level business challenges that they face.
To learn more about assuring performance and compliance of EHRs, PACS, VDI, and mission-critical applications, download Maintaining Datacenter Health For Healthcare Providers.