VDI - A Solution in the Real World -Pt2

From my previous post on VDI, I detailed what VDI is and how it is deployed. Now we move on to how do I get my users on this type of system and what are the benefits of this type of system within my organisation. 

When I introduced my previous post I detailed how I’ve seen VDI come on over the years. With VDI performance being improved extremely by Vendors new technology leaps and bounds. The Type of Workloads the end users place on the VDI systems was something that had to be looked at carefully when deciding the type of infrastructure. When determining your user requirements most organisations thought that they would replicate what was under the desk and put that in the data centre. This added expense and complexity to the solution as this then made VDI too costly to implement. You only had very few users who were granted the privilege to use VDI as the business required.. We now have system networks that are more than Capable. We have Infrastructure that is more performant than the previous years and now we are more than ever more able to allow flexible working and more inclusive working environments.

The below are some of the determining factors:

The Network:

Has always been a challenge and with the introduction of 10Gb/25GB/40GB/100GB etc, these speeds have given benefits to all organisations that have looked at VDI as a solution.

The Storage:

Virtualized desktops are hosted on a centralized infrastructure. Users can access the same desktop with a variety of end-point devices using a remote display protocol over a network. Enterprise grade storage is critical for VDI deployment success:

The CPU:

A virtual desktop can only run as fast as the core on which it is running, so this always seemed to create a problem, as server platforms never seemed to be running at a high clock speed in comparison with Workstation/Desktop CPU’s

The GPU:

GPU used in VDI deployments is gaining traction thanks to a number of technological advancements and emerging use cases, such as the ability to share GPUs and free up CPU resources. In the early days of VDI, IT administrators used GPUs as a way to appease power users who would otherwise reject VDI.

Application Layer

Cost of these Apps being put into the datacenter rather than on a workstation under the desk

Application Portability and Compatibility:

Can your Application be in a Virtualised state?

This has also led to the introduction to a form of build or buy VDI systems which has fuelled the introduction by Vendors of their own Hyper-converged systems specially formulated to bring Core speed, storage, and network performance in a system, namely #DellEMC, #ciscoucs, #hpe #hpeconverged , #hpesimplivity #nutanix But just because hyper-converged computing and VDI seem to go hand in hand doesn't mean that you should think of hyper-converged infrastructure as plug and play for VDI use.

What are the components of hyper-converged infrastructure (HCI)?

HCI consists of two main components, the distributed plane and the management plane. The distributed plane runs across a cluster of nodes delivering storage, virtualisation, and networking services for guest applications—whether they're VMs or container-based apps.

The type of workloads the end users place on your VDI will primarily shape your options for an HCI system. The work patterns of those VDI users are also a key factor, as is the software they will run on the virtual desktops.

How to Determine user requirements:

How many virtual desktops do you need? Most hyper-converged computing systems have a minimum number of nodes in a cluster and can handle hundreds of virtual desktops. If your IT needs aren't at that level yet, implementing HCI would be over- provisioning and quite expensive and maybe a simple Build it on your own would be best suited.

Once you know how many machines will be virtually deployed in your VDI, you will need to understand the work profiles of the users of those machines. Typically, VDI users can be put into three groups: task users, knowledge users and power users. Each has different levels of resources they need from the shared pool of an HCI system.

Analyse the user group as a whole to determine the concurrency rate -- how many users are on and working at the same time. Your hyper-converged computing system will need to be able to operate efficiently at the peak demand time. This is usually when end users all log on if they are in the same time zone, an exercise commonly known as a boot storm. Having a geographically dispersed workforce eases resource demand peaks, but you will still need to find out when those peaks occur and provision resources to meet those peaks.

Benefits of HCI for VDI

The primary benefit of using a hyper-converged computing system for your VDI needs is easily scalability. As your company grows, you simply add HCI nodes to add new system VDI Users

• Cost savings are also one of the draws for HCI Buyers. The main cost savings come in the form of reduced management costs, because the HCI has most of the management functions built in and at least partially automated. That means fewer specialised types of IT administrators, such as storage admins, for most operations that use HCI.
• An additional benefit of hyper-converged computing is that it can ease or even eliminate one of the primary problems VDI faces in traditional IT structures, the storage issue. Storage capacity demands grow faster than any other aspect of a VDI system. This can lead to storage becoming the resource that slows the performance of VDI for your end users.

The ease of expansion of resources in an HCI system makes the storage bottleneck all but disappear. Simply adding more nodes to a cluster will add more storage capacity to the virtualised logic pool to be shared among your VDI users.

Drawbacks of HCI in VDI

The ease of expansion can also lead to one of the drawbacks of an HCI implementation. 

1. You might over-provision one resource to meet the demands of another. Simply put, since each node contains storage and compute capacity, you may end up with more compute resources than your VDI system needs in order to meet the storage demands.
2. Not all Workloads are addressed by HCI and thus one size doesn’t fit all requirements.
3. Workloads which are single threaded suffer as a result of HCI performance offerings

#vdi #virtualdesktops #virtualisation #pandemic #dellemc #citrix #isilon #datacentre #vmware #xendesktop #nvidiagpu #hyperconverged #ciscoucs #hpesimplivity