Virtual desktop infrastructure (VDI) adoption is growing
rapidly and expected to be a $10 billion market by 2023 according to a
report
from Allied Market Research. With increasing dependency for enabling remote
connectivity and anywhere desktop access, the expectation from VDI has also
increased. Users expect the same or better performance as they have from their
physical desktops, so assuring the user experience is a critical success factor
for VDI deployments.
The heterogeneous nature of digital workspaces and VDI
ecosystems create significant challenges for managing end user experience.
Users log on and off their desktops throughout the day, graphic-intensive
applications are gaining popularity and it can be difficult to identify the
source of performance issues. Troubleshooting tends to be reactive, yet the
impact of failure is high-a single server-side or network problem can impact
hundreds of end-users. Resource contention and bottleneck is a killer in VDI
environments.
Based on eG Innovations' extensive experience in monitoring
the performance of VDI environments for close to two decades, here are four best
practices that we prescribe which should be leveraged for assuring end user
experience in VDI environments.
Best Practice #1: Monitoring All Aspects of User
Experience
First, both synthetic and real user experience should be
measured.
-
For example, using a logon simulator can allow
you to proactively identify issues with logon performance before users are
impacted. Simulation lets you test different applications and desktops, from different
locations, even when there's no real user load.
-
In addition, it's critical that you be able to
monitor real users running actual workloads to get a direct measure of what
users are seeing. This includes both real user logon performance as well as
real-time VDI session monitoring of end user performance.
Best Practice #2: Monitoring Every Layer and Every Tier
of the Virtual Desktop Infrastructure
Second, it is essential that you be able to monitor every
layer of every tier of VDI its supporting infrastructure. This includes
user-level, session-level and desktop-level as well as resource (CPU, memory,
disk, GPU) monitoring of all server-side components. Keep in mind that
applications running over a VDI ecosystem may also have dependencies, and users
will not care about what's causing a performance issue-they'll just want good
performance!
Gaps in user experience monitoring are often due to an
exclusive focus on only one technology domain (i.e., VMware Horizon or Citrix only) rather
than every layer of every component that is needed to deliver user experience.
Best Practice #3: Automatic Root Cause Diagnosis for
Faster Troubleshooting
Third, your monitoring must be able to correlate performance
data to isolate performance issues as automatically as possible. The ability to
automatically pinpoint application slowness-
Why is My
Application Slow?
Client side?
User's network? VDI server-side? User's workload? Hypervisor? On-premises
infrastructure (AD, storage, etc.)?
-will be the key driver of the cost of ownership of your
monitoring environment. The correlation should require minimal
configuration/customization, it should be able to automatically triage
performance issues across all tiers, and it should do so with little or no
human intervention. Automation is key.
Best Practice #4: Right-sizing and Optimization for
Maximum Performance
And finally, it's important that you are able to perform
capacity management effectively. This includes right-sizing and performance
optimization of both the VDI and supporting ecosystems. The ability to leverage
machine learning and analytics can help improve proactive reporting and
capacity planning to maintain high levels of end user experience.
Customers should look at all aspects of digital workspace
services, including VDI and its supporting ecosystems when looking to assure
end user experience. Monitoring only the hypervisor platform or ESX hosts and
VMs is not enough. Components such as connection brokers, VDI sessions, network
and others are just as critical.
It's important to be able to map users to specific VMs and
be able to see both external and internal views of VM performance to identify
VM-level issues and/or application-level issues inside a VM.
End user experience has become key to competitive advantage
in the digital world and should be treated as such. Don't let monitoring be an
afterthought. Monitoring
your VDI environment is not a one-time activity; it must happen
continuously and during every phase of VDI implementation.
Having a centralized, correlated view of performance across
the entire VDI delivery infrastructure from a single pane of glass is often the
difference between a successful VDI deployment and an unsuccessful one.
##
About the Author
John Worthington, Director of Customer Success, eG Innovations
For over a decade John's been evangelizing unified
monitoring as an enabler of transformative change for IT organizations. He's
been an ITIL Expert since 2005 and continues to work with clients on
transformational change associated with virtualization technologies, cloud
operating models and DevOps. He has over 30 years of executive experience as a
customer, a consultant and with technology suppliers.