A Practical Guide to VDI Change Management
Part 4: IT Change Accelerates in General
In the previous chapters we stated that “the more complex infrastructures become, the bigger the chance for something to fail”. We can easily expand this into a logically related conclusion when we state that “the more complex infrastructures become, the bigger the chance for something to change”.
Updates, They Are A-Changin'
An average VDI environment is subject to many changes, upgrades, updates, and patches, such as:
- New / updated Operating System / Patch releases (accelerated with the introduction of Windows 10, see previous chapter)
- VDI related software infrastructure changes (EU layer and hypervisor)
- Core business application updates and upgrades
- Security patches such as for Meltdown, Spectre, and L1TF
- Hardware changes / Configuration changes / Firmware revisions
- Memory upgrades / Storage upgrades / GPU’s
The flow of changes that can affect a smooth running VDI environment is already significant due to its complexity and many parts. In addition, this flow of changes is growing and accelerating as vendors are more and more moving towards Software as a Service models, where maintenance is a constant fact of life. On average in 2010 a typical core business application would receive 4 updates per year. In the new models it is expected for that number to grow to 120 by 2020: a 30-fold increase.
Looking at a few large corporations we found some impressive numbers in this respect:
- A major US retailer went from 1 update per month to 80 per week
- A major US online retailer pushes 300 changes per day
Like Microsoft, most other major software vendors are switching to a release schedule based on Software as a Service. Sending out many small updates with a short interval e.g. two to four weeks. While this eases troubleshooting as the changes are smaller, it does force IT departments to push a lot more changes to production, to keep up to date and to stay within support.
Microsoft - Changes in Updates
We mentioned the Windows 10 update cycles but there are two other Microsoft products that also must be maintained in nearly every enterprise environment: Microsoft SCCM and Microsoft Office.
Many organizations rely on Microsoft SCCM to orchestrate Windows desktop deployments, application installs and even mobile devices. To be able to deliver two feature updates per year for Windows 10, three SCCM feature updates must be done per year, to make sure the systems are compatible.
Tests executed by the EUC initiative, VDI Like a Pro, show that every Microsoft Office update has a tangible impact on the End User Experience (in some cases decreasing application performance and/or server density with more than 20%!). The Microsoft Office suite is also notorious when it comes to macros and other in-house or third-party built plugins, as updated security features often block applications that contain these additions.
Citrix - Changes in Updates
This infrastructure company now offers three service plans to keep XenApp/XenDesktop environments up to date ranging from their Citrix Cloud offering that’s automatically brought to the latest version, a Current Release (CR) version every 3-9 months and (since January 2018) a Long-Term Service Release (LTSR) version being released every 12-24 months with cumulative updates every half year.
Citrix XenApp/XenDesktop – Virtual Apps and Desktops LTSR update cycles
VMware - Changes in Updates
Current Release (CR) regularly and an Extended Service Branch (ESB) every 3 to 6 months.
VMware Horizon View ESB update cycles
Core Business Applications
Core business application upgrades can have big impacts. Starting early 2018 one of the leading US based EHR-vendors plans to roll-out one major release per year, complemented by one Special Update (SU) per quarter. Rumors are that a performance loss (read extra hardware) should be budgeted for, with additional resource needs of approximately 20% for the major releases and 9% for the Special Updates (results may vary), when a SBC platform is used to deliver the application. This adds up to about 50% (when quarterly compounded) in incremental hardware capacity being required each year.
Whatever the real numbers will be, it is clear that new releases and updates of your core applications need to be taken very seriously. As every environment will react differently, testing with a full user load in your organizations’ own specific production environment is the only way to predict the exact impact, and the only way to be able to react pro-actively in the most appropriate way possible.
An IT manager at a large American healthcare enterprise using Epic as EHR, mentioned that Login VSI will save them millions of dollars each year, and ensures they stay ahead of performance problems.
Meltdown & Spectre
Recent malware releases and security flaws (sometimes even in hardware) force IT departments to be very careful. At the beginning of 2018, when news about Meltdown and Spectre became public, more than 100 affected vendors started pushing mitigations, followed by more permanent workarounds and fixes. The hasty job of putting together these patches resulted in many outages and application failures as well as degraded performance in many enterprises.
L1 Terminal Fault (L1TF)
Security patches for the latest Intel focused danger, L1TF, are predicted to damage scalability and performance in VMware environments with 20 to 30%! Even though the L1TF security problems are not as prominent in the news as Meltdown & Spectre this is not a patch to be taken lightly.
And this is not the end
Following these vulnerabilities emerging in the first months of 2018, we can be quite sure the chances are high that more new vulnerabilities will follow. Making preparing for the impact of security patches, something that needs to be embedded in any well-structured change management program.
Hardware Renewal in Datacenters
Major hardware vendors recommend a hardware refresh every 3 to 5 years depending on warranty and workload. Newer generations of hardware are significantly faster (e.g. Intel promised a 17x better performance in hardware from 2007 to 2013) and more efficient when it comes to power usage and space requirements. Less hardware can also considerably lower licensing costs, but centralizing more services also means more single points of failure that can bring down the whole environment.
In addition to complete servers, smaller portions of the infrastructure may be upgraded due to the availability of new technologies in storage or GPUs.
Conclusion: The more complex environments are, the more (often) changes will occur.
Start using Login VSI today
Our industry-standard software is built to help you avoid problems, lower costs and improve performance. Request a Quote or get your free Trial below, and benefit from our award-winning services.
About the company Login VSI
The company Login VSI provides end-user performance insights for virtualized desktop and server-based computing environments. Enterprise IT departments use flagship product Login VSI (for load testing) and Login PI (for continuity testing) in all phases of their virtual desktop deployment—from planning to deployment to change management—to build and safeguard a good performance, a high availability, and (as a result) a good and consistent end-user experience. For more information about Login VSI or for a free test license contact us.