How we cut user logon times at a major NHS hospital from 3 minutes to just 20 seconds.

0
118

Founded in 2009 and hailed as the first NHS hospital in the country to offer every patient their own room, it opened its doors to great acclaim. In addition to 96 en-suite rooms, it is also home to 11 adult mental health beds, 15 outpatient consulting rooms, a radiology department, and a GP out-of-hours unit. 

Today, the hospital employs more than 13,000 staff; comprising 250 consultants, 1,000 hospital and general practice doctors, and over 6,000 nurses, midwives, allied professionals and community workers. 

The challenge. 

Users were becoming increasingly frustrated by logon times that could keep them waiting for long as three minutes before they could access the system and get to work. Yet a state-of-the-art infrastructure had only just been commissioned and installed, adding to the overall sense of frustration.  

Having deployed an almost identical infrastructure to another client, we hadn’t encountered such slow logon times and as a result felt confident that we could help. 

The assessment and key deliverables. 

In order to determine why such a newly installed system wasn’t performing as promised, our first step was to define the scope of work that would be required; namely to review and assess the underlying infrastructure and the logon process of all the Citrix Windows 10 VDI machines in use. 

This full assessment covered the following items: Nutanix Hypervisor AHV, Storage, Windows environmental setup, Windows group policies, Citrix Infrastructure, Citrix Policies, Citrix Windows Environment Manager (WEM) and Windows 10 VDA. 

As a result of our analysis we were be able to present the following deliverables: a review and report of each technology item in relation to slow logons, highlight best practice recommendations, identify the cause(s), and finally recommend solutions that would rectify slow logons across the environment. 

The test methodology. 

Once we were granted remote access to the relevant systems for assessment, we were able to work collaboratively with the in-house team and commence our evaluation. 

We were pleased to report that overall the AHV Nutanix cluster was found to be in good working order. There was no evidence of any issues with the cluster, its configuration, or the underlying storage, that would result in increased logon times 

After giving the cluster a clean bill of health, we turned our attention to uncovering the cause of the slow logons using iterative testing. This ‘rinse and repeat’ methodology involves making singular changes to the user account, policies, and the virtual machines; while repeatedly testing logons. 

The test logons were performed against the standard HCI desktop and a persistent copy of the same desktop from which changes could be made. They were conducted on both Citrix and Direct Console sessions then re-iterated after each policy, system or service change, with verbose logging configured to capture details for later review. 

This thorough and methodical approach of making slight changes and then identifying any issues took place over a 2-week period; during which we were able to examine the system in fine detail and report back with a comprehensive set of findings and recommendations. 

Summary of the recommended actions.  

As you can imagine the testing procedures and outcomes are extremely detailed, which is why we are presenting a brief summary of the key recommendations (if you would like a more in-depth report, then please don’t hesitate to contact us.) 

  • Group Policy. Reviewed Group Policy Objects thoroughly, as GPO applied to physical machines often aren’t suitable for VDI environments. Better use of WorkSpace Environment Manager (WEM) could be made in order to control user-based Group Policy Objects.  
  • Profile Management. Create a Group Policy Object to control Citrix User Profile Management and remove the configuration from WEM as defined by Citrix best practice. If GPO control is not possible for the desktop team, this policy should be controlled via Citrix native policy.  
  • Anti-Virus. Disable Asynchronous Scanning, review exclusions. 
  • Citrix User Profile Manager. Tune the UPM to set exclusions appropriate to the environment, implement Profile Streaming, or redirect “AppData” to a suitable shared location. Applications such as Chrome should be controlled with separate GPO to prevent profile bloat.  
  • Group Membership. Conduct a full review into standard user account group membership and policies applied because of those groups. 

The outcome. 

By implementing the recommendations highlighted above, logon times have been reduced dramatically; a 90% reduction in time or approximately 6 times faster than before Quadris applied its expertise. 

As a result, whereas previously a user might experience a logon time of 2½ – 3 minutes, thanks to Quadris it has been cut to just 20 seconds. 

Not only does this go a long way to reducing the frustration encountered by individuals, multiply this figure by the many hundreds of users who access the system on a daily basis throughout the year, and it adds up to huge time savings and increased efficiency for one of our already overburdened hospitals.  

To discover for yourself just how Quadris can draw upon its vast expertise in order to solve the most pressing of IT problems, contact Peter Grayson on 0345 163 1080 or email peter.grayson@quadris.co.uk