The global pandemic caused what may have been the biggest IT challenge in history. Practically overnight, companies called on IT administrators to find a way for almost every office worker to be able to connect from home, safely and securely.
One fast solution for many companies was to host virtual desktop infrastructure (VDI) on a public cloud. Using the cloud was a good idea for many reasons:
- Immediate availability
- Consistent VDI “broker” software your business had as a standard (Citrix or VMware)
- Device-agnostic access for nearly any device the employee had access to, including laptops, phones, and other mobile devices
- Acceptable data security protocols
- Consistent user interface
- Easily forecast costs
However, what began as a short-term need turned into a long-term or even permanent situation. The global rush to the cloud to support millions of employees has caused a permanent shift in how we work, with many companies offering flexible hybrid office and remote working options and many employees opting for 100% remote work.
According to a recent CNBC article, more than 68% of full-time employees support working from home, at least part-time, in the future, and many companies are supporting this hybrid work scenario. Even on days when they commute to the office, many employees log in beforehand and then come to the office later in the day, compared to before the pandemic.
Post-pandemic, people are also commuting to work later in the day, with a recent article in The Washington Post showing that the percentage of people heading to the office after 9 a.m. has risen steadily since 2019.
Does the return to office lower the VDI need?
Because IT departments are now supporting employees working from home and in the office, the need for VDI remains constant and the monthly cost of hosting all these desktops in the cloud is becoming a permanent cost. Often these costs are significantly higher than was initially projected back in 2020 when many worked remotely. A part of this is general price increases, but more typically it is the ease with which additional cloud resources are effortlessly accessed by the customer without a full appreciation of the additional costs this access incurs.
Options beyond the cloud
Intersect360 Research recently published an analysis comparing a VDI 2,500 desktop deployment on a public cloud with a Cisco compute-based on-premises deployment. The hypothetical customer in the study saved $2 million over three years using the Cisco deployment.
Figure 1. A Cisco on-premises VDI solution saved $2 million over three years compared to a public cloud deployment supporting 2,500 desktops, according to Intersect360 Research.
The savings are significant
If you were to break down the cost on a per-user basis, the advantage is roughly $800 per user, or about the cost of a laptop (see the table below). Additional cost savings can be realized when you consider the on-premises model:
- Adds an additional 10% cost to cover any items not factored into the study
- Includes three-year on-site support cost
- Includes the cost of two head counts to administer the system infrastructure
- The system design can accommodate an additional 1,000 desktops (VMware Horizon licenses are an extra cost)
How the study worked
Although many determinants impact the cost of any VDI infrastructure—on-premises or in a public cloud—the number of virtual desktops and the associated user profiles are the most significant factors. For the Intersect360 Research study, they assumed a fictional customer whose user profile is closely representative of a manufacturing customer. The breakdown included:
- 25% clerical: Call center work, data gathering and correction, data analysis, and reporting.
- 25% administrative: Marketing, accounting, business analysis, HR, and customer service.
- 35% knowledge workers: Typically manage people but could be program managers. Have a lot of contact with customers, suppliers, and partners. Need support for multitasking and high-quality video conferencing.
- 15% specialists: New product development, testing, creating multimedia content, and analyzing huge amounts of data; need significant technical capabilities requiring graphics support and larger amounts of memory. Use of a GPU may be required. Commonly called a power user.
- 2.5% ultra specialists: A subset of specialists who focus on AI, machine learning, CAD/CAM, video creation/rendering, data mining, and hosting scientific applications. Requires a GPU.
These user profiles largely determine the amount of system resources required to meet user expectations, whether their desktop resides on a public cloud (paid for on a monthly basis) or within the confines of a company data center.
VDI doesn’t have to be difficult
Cisco Compute has been a leader in VDI infrastructure dating back to our entry into the server market in 2009. We partner with leading storage vendors such as NetApp to form solutions for specific enterprise applications designed, tested, and documented in Cisco Validated Designs (CVD). These solutions are tested in Cisco labs by Cisco engineers, defining the best practices for all system elements to demonstrate a balanced performance. Your time to implement is shorter, thus saving costs, with the knowledge that these best practices will help provide optimal installation. In fact, Cisco support will provide remote, even on-site assistance to ensure documented performance levels are reached.
Today’s designs incorporate the Cisco UCS X-Series Modular System, which can accommodate up to eight UCS X210c M7 Compute Nodes with the 4th Generation Intel® Xeon® Scalable Processor and the fastest system memory available. The Cisco UCS X-Series Modular System has won a number of awards for its sustainability design and power optimization features, most recently from SEAL and CRN. The NetApp A-Series All Flash array provides the data storage and NetApp ONTAP AI software enables data management such as snapshots and data backups. The complete system is managed by Cisco Intersight, which provides superior lifecycle management. Intersight Infrastructure Service helps prevent problems by delivering proactive support, increasing efficiency, reducing costs, and accelerating resolution times. Telemetry data is automatically synthesized, and recommendations are presented in an easy-to-understand, actionable format. The system can identify and side-step IT issues before they impact your data center.
Cisco CX services can assist your team with initial VDI assessment and cost comparisons. Help architect a system design to meet your current and future requirements, implement the design, and provide full support and lifecycle management.
Intersect360 Research added the cost of VMware Horizon 8 as the VDI “broker,” as well as Nvidia GPUs and related software to meet the power user requirements. The result was a complete design to host the VDI deployment for this manufacturing company.
Consider on-premises VDI solutions today
Is the majority of your VDI still hosted on a public cloud? Read the discussion around, “Is the time right to move VDI deployment from Cloud into the datacenter?” — where you can download the Intersect360 Research report and pass it around your organization. Then, perform your own cost analysis comparing the three-year cloud-hosting costs with an on-premises design. Cisco can assist with cost-effective designs and help ensure that relevant costs are identified. While your numbers will be a bit different from the Intersect360 Research analysis, you might uncover cost savings even if additional head count is needed to manage the system.
Download the Intersect360 Research report
Learn more
- Cisco UCS X-Series for GPU-Accelerated VDI At-a-Glance
- FlexPod Datacenter using Cisco UCS X-Series with VMware Horizon and VMware vSphere 7 for up to 2600 Seats
- FlexPod Datacenter with Citrix VDI and VMware vSphere 7 for up to 2500 Seats
CONNECT WITH US