When it comes to application and desktop virtualization projects, choosing the right infrastructure for your business is among the most important decisions you’ll need to make. Second only to selecting a strategy and software vendor for delivering End User Computing (EUC) services and applications, infrastructure is the bread and butter from which you’ll build your services, so a highly available and scalable infrastructure is integral to bypass potential challenges that may arise during the deployment and operational phases of your EUC project.
To better evaluate your architectural alternatives and vendor options, let’s break down five architectural principles you should consider during your infrastructure design process:
Entry PointAlso known as the starting point, this principle can often be a make-or-break decision. The entry point determines both how much infrastructure you’ll need and how much it’ll cost to start desktop and application virtualization and delivery deployments.
Let’s take a look at an example. Say your final project is meant to reach 10,000 users, with the initial deployment phase reaching 5,000. You’ll be far less shocked with the initial costs by starting smaller, and depending on your infrastructure, the cost per user will become more clear as you deploy a few thousand users.
On the other side, an initial deployment of 500 users can still allow you to scale up and pay as you grow. Plus, the cost per user will rise steadily at this scale. A lower up-front cost may be easier to sell your project to business leadership, and a higher initial cost will require you to go into detail as to why. Take a look at the visual to illustrate these two examples:
The scalability of your architecture is important both for evaluating your project viability and ensuring that your project resources don’t go to waste. Depending on your architectural alternatives, IT architects will need to understand starting size options and if your design will allow you to start small or require a larger upfront payment.
How large can your architecture alternative scale to? If you’re starting at 500 and aiming to end at 10,000, the architecture may change at both ends of the spectrum, and you’ll need to consider if your organization will be happy with the high points, low points, or both.
However, while the architecture should be able to work at both the initial small scale and the larger final goal, the points in between are equally important. There may be scaling points along the way, and a design that easily scales in buckets is optimal—100-200 users per bucket, ideally.
The end-user experience determines EUC performance, so your architecture must meet user needs. Your architectural alternatives will be put to the test: if one meets minimal user requirements at the starting point, it may not be able to scale linearly and meet later performance challenges. So, choosing the right architecture is important to bypass this and other problems.
Your architecture should be flexible, meeting any and all performance requirements that come about, so the design and evaluation process must thoughtfully consider desktop and app virtualization and performance needs.
There’s no shortage of capacity requirements, so your EUC design will need to meet those head-on, running server VMs, desktop VMs, applications, user profiles, and user data. To complicate things, each layer within the design will likely have different capacity requirements, with some using more data that deduplicates well and others with smaller data that can duplicate to become larger in the end.
To avoid the issue of purchasing too much or not enough capacity, consider architecture alternatives that will provide required capacity while meeting minimum performance requirements. The alternative shouldn’t provide 2-3+ times more capacity than you need; instead, an ideal solution will offer the flexibility to scale performance and capacity at similar rates, allowing both to keep pace.
Too often overlooked, monitoring infrastructure in an EUC environment is a critical consideration. Since administrators tend to focus on performance speed, they need to recognize active issues.
Monitoring should provide detailed information, but it shouldn’t be a round-the-clock job. However, this is not the standard for all vendors, so be sure you consider what it’s like to monitor your infrastructure under each architectural alternative.
Providing performance monitoring at the virtual machine level is another requirement, but a majority of infrastructure vendors do not provide this level of visibility. It’s no longer an option to quickly look at the storage layer and see if the performance issue is global or isolated to a host, group of VMs, or just a single VM.
On the topic of VMs, administrators can manage storage performance at the VM level and see if a VM is temporarily using additional performance or if it’s a consistent consumer, letting you more clearly identify spikes.
Selecting and managing an ideal infrastructure is nothing short of important; be sure you get all the details you need to choose one that meets the needs of your services and applications without sacrificing time and resources. A good starting point to find this information and more is the Definitive Guide to VDI eBook. Once you’ve got the facts on what you need to create successful end user computing environments, see what you can save on VDI with this handy VDI calculator.
© 2018 Nutanix, Inc. All rights reserved. Nutanix, the Nutanix logo and the other Nutanix products and features mentioned herein are registered trademarks or trademarks of Nutanix, Inc. in the United States and other countries. All other brand names mentioned herein are for identification purposes only and may be the trademarks of their respective holder(s).