For years, organizations deploying VDI for the first time have felt the sting of the most common issue that plagues virtual desktops: inadequate performance from shared storage. The storage infrastructure that has worked well for decades when used for servers is found to be dramatically lacking when virtual desktops are deployed to it; for that reason, VDI has received many black eyes (and in some CxO offices, a permanent bad reputation) and that has been very difficult to overcome.
Fortunately, the available options for shared storage are multiplying, with VDI driving a good portion of that change. An organization looking to deploy VDI today has all-flash array options, de-duplication offerings, hyper-converged (scale-out) systems, and even local storage aggregation products available – all of which can not only increase performance of VDI, but lower the cost as well. Although the price point of an individual virtual desktop can be endlessly debated and is largely dependent on the performance requirements of the end user, it is now a safe assumption that a virtual desktop can be made to perform as well as, or better than, a physical machine at a reasonably competitive price.
Obviously, that means VDI performance issues are behind us, right? Unfortunately, the answer is no – at least for organizations looking to run hundreds or thousands of virtual desktops. The hurdle of “slow storage” was so large and so overwhelming for VDI deployments that it was often functionally impossible to see if there were other bottlenecks in the system. The problem these organizations encounter, once storage has been “solved”, is that suddenly mass events, such as anti-virus scans and updates, inventory scans, and software distribution have all the IOPS they can handle, so they begin competing for the other parts of the shared infrastructure: network, CPU, and memory.
What happens when 150 persistent virtual desktops, all hosted on the same physical VDI host, are told to download and install a 70mb patch, and there is no disk I/O contention? One hundred and fifty instances of a software delivery agent spring to life, churning against the vCPUs assigned to the virtual desktops. The patch is retrieved. The patch is expanded. The patch is installed. Post-installation cleanup occurs. And don’t forget the reboot. The performance problem is no longer the disk, but it still very much exists; one bottleneck has been removed only to reveal another.
As you can imagine, the great benefit of virtualization – oversubscription – absolutely cannot deliver even close to adequate performance to these virtual desktops all clamoring for cycles at the same time. Often, the result is an end user experience that ranges from slow to unusable. Desktops that should update in a minute or two are so starved for resources that they take 30 minutes or more, during which time they have nearly zero cycles to dedicate to the work of the user. Ultimately, this turns into hallway chatter: “VDI is awful.” “I want my desktop back.” “This is the worst technology I’ve ever used.”
How do you avoid this? In my latest research document, “Selecting the Right Application Delivery Model for Virtual Desktops” (available to Gartner for Technical Professionals clients), alternative ways to deliver applications (and updates, and patches) to persistent virtual desktops are reviewed in detail. By comparing options such as Server-Based Computing, Application Virtualization, and Application Layering to traditional, physical-focused application management tools, I hope that organizations will be able to see and avoid this issue before it happens, or remediate the problem if it has already been encountered.
Like most technology, none of these solutions are one-size-fits-all. However, among the options discussed, I believe most will find a solution that will reduce or eliminate this bottleneck. And good news – once you have solved storage and solved application delivery, the performance hurdles of VDI finally begin to get smaller. Don’t lose hope!
Which application delivery approach are you currently using for your VDI implementation, and how well is it working for you? Let us know in the comments.
The Gartner Blog Network provides an opportunity for Gartner analysts to test ideas and move research forward. Because the content posted by Gartner analysts on this site does not undergo our standard editorial review, all comments or opinions expressed hereunder are those of the individual contributors and do not represent the views of Gartner, Inc. or its management.
22 Comments
We are in the healthcare industry and currently use VMware View with Unidesk for application layering on top of a Nimble SAN. Application updating is a huge issue as you have mentioned with VPCs, however this has been resolved for us through the use of Unidesk for Application Layering. Using Unidesk for patch management, Windows Updates, application deployment and VPC creation allows us to complete these tasks when we need them to be done, utilizing minimal resources.
The Nimble SAN offers a solid storage solution and works quickly and efficiently with our systems. We have 200 VPCs currently deployed with another 200 on the way, and have had much success with updating applications, operating systems, without breaking the bank, or the hosts available resources.
I would highly recommend this configuration to anyone looking to complete these same tasks.
You bring up lots of valid points and questions that any business must address when considering VDI. We are in higher education and utilize Unidesk and Vmware View for our approach to VDI. We chose Unidesk because of its update once and deploy. We could not afford the time and space required for Vmware’s golden images due to the vast number of unique configurations across our campus. If we were going to do this…we would have kept the physical PC on the desk.
In addition…Unidesk’s support is top notch. They also listen to their customers and find ways to meet our unique needs. For example….we have nonpersistent desktops deployed in labs but occassinally we need to thaw and make changes that can’t be done in an app or OS layer. After opening a case and a few phone calls alter, we had a solution.
We currently have 300 VDI persistent and nonpersistent desktops deployed across campus with another 100+ to be deployed within the next few months.
Our current storage is two Dell EqualLogic SANs on 15k spinning disks. Even though our performance is fine, we have recently purchased Cisco Invicta 24TB all flash storage to run our VDI environment on. The Invicta has IOPS in excess of 200,000 per unit and will be operational by the end of November.
We also utilize Trend DeepSecurity which is a host based antivirus/malware instead of loading software onto each VM. It has some quirks but for the most part….it gets the job done.
I would highly recommend Unidesk and VMware along with the other software that mentioned earlier.
Feel free to contact me if I could be of any assistance to you.
Another healthcare VDI user here. We have implemented XenDesktop 5.6 with vSphere 5.0U2 on EMC VNX5300 arrays utilising flash for FastCache and FastVP alongside traditional spinning disk (FC and SATA). This solution has been in place since mid-2012 serving over 3500 users with a mixture of persistent (1500) and non-persistent desktops.
We wouldn’t have been able to achieve this large scale rollout without the benefits that Unidesk offers; simple image management combined with its layering functionality and personalisation capabilities. Using a single gold base image we present a number of Unidesk layered core applications including Microsoft App-V and a user personalisation layer.
Our delivery model is then built on application requirements/limitations:
Core applications (Microsoft Office / Lync / others) are delivered via Unidesk layers.
Line Of Business applications are delivered via Microsoft App-V or Citrix XenApp.
Users can self-install applications in their own Unidesk personalisation layer.
We also have the ability to use Unidesk layering for those misbehaving applications (crapapps as we call ‘em) that present issues with App-V or XenApp.
The project has been so successful that the next phase of our project is to roll out desktops to a further 3500 users. We have built out a new environment for this development built on XenDesktop 7.x/vSphere 5.5 with EMC XtremIO arrays providing the storage component. We are offloading antivirus activities to at a host level utilising vShield with Sophos appliances. Unidesk is at the heart of this delivery which offers user desktops with sub-millisecond response times.
Migration of our existing user-base is already underway with the aim to be fully loaded by December this year. Virtual desktops are now seen as the first class citizen in our organisations delivery of desktop/application services allowing clinicians true mobility throughout our 100+ facilities.
Whilst I’m interested in what VMware bring to the table with CloudVolumes I’m very excited about the innovation Unidesk have on their roadmap – it’s a company that doesn’t stand still and has been delivering working solutions to us with excellent support for almost 3 years now.
For my organisation 2012 was the year of VDI – 2014 has been the year for better VDI.
(Available for further discussion if required.)
Mark,
I love VDI! Finally, somebody that is focusing on what matters. Traditionally, VDI blogs focus on the wrong arguments. My favorite is non-persistent vs. persistent. Trivial at best; because, we all have an opinion. Considering your blog, here’s my opinion and experience.
Yes, storage was a huge issue for us in our VDI POC, three years ago. We quickly realized that this was the single biggest bottleneck. My “traditional” SAN was killing me, not so much in IOP’s, but latency. I solved that issue, within another VDI POC. For our environment, storage latency is what was key. We don’t even come close to the IOP’s supported by the solution. The storage eats IOP’s for breakfast, before it goes to the gym. Storage latency is what makes the “customer” experience.
VDI will expose every weakness in your infrastructure. Successful VDI deployments are dependent on a solid network infrastructure. This is seldom discussed, but a huge challenge. Bad network deployments at layer 1-3, will result in VDI failure. Fact.
The journey to desktop virtualization is huge. Nobody ever talks about the “big one”. There is a massive investment in making this journey successful. It really is all about the customer experience, do not forget that, or you will fail. Failure to gain acceptance there, will “doom” your VDI deployment. That’s the “bad taste” you speak of.
Gaining acceptance in the adoption of the concept, within our organization, took several years. I propose ideas and concepts as an IT worker. My job is to save time, and time is money. The people that sign the checks, understand that. The “geek” perspective will not sell VDI.
POC’s that generate POC’s? This is “rocky” ground in IT. I typically run, as fast as I can, from this sort of situation. VDI was a bit different. Successful deployment, would result in a tremendous amount of savings in, administrative overhead, time, power, access, BYOD, the list goes on.
I did my POC in a traditional VMWare View environment, 4.x. days. Brutal! Nothing worked with application deployment. Issues and problems, across the board. Managing a gold image circus was not an option, so I was ready to scrap the POC. I will not go into details, but after six months of hard work, I couldn’t let it go.
You guessed it, I’m a Unidesk customer. A project I was not willing to give up on, was resurrected. The project I was getting ready to “scrap”, was reality. Six months of opex/deployment failure and frustration, was resolved in one work day. Less than that actually, about two hours. It really was a miracle for me, as I have limited resources and time.
The goal of VDI, was to significantly reduce the amount of time managing a traditional PC environment, while maintaining the customer experience. Thanks Unidesk.
In July, we finished our VDI project, all faculty, staff, labs, classrooms, etc. We will only grow from here. A previous poster said the solution was not “scalable”? He may need to come visit our deployment, or other Unidesk customer’s, before he makes that bold statement. CloudVolumes seems great, the fact of the matter is, I’m not looking for a replacement solution. I’m happy with Unidesk. That speaks volumes.
I do understand the challenges of a VDI deployment, as I live those on a daily basis, 7/365. There will be “bumps in the road”, as with all new technology, but it has revolutionized our infrastructure. Tasks that normally took many hours, have been reduced to a few minutes.
I have volumes to speak, so if I had any advice, do not forget about your customer. The best VDI deployment will fail, without customer acceptance. You can try to force it on them….good luck.
As someone who built and supports 29,000 1:1 VDI instances I agree with some of your statements but not all of them. While 3 years ago storage was absolutely the bottleneck the reality is that these days I’m more concerned with CPU and RAM consumption with newer application sets. For instance a running instance of Lync 2013 consumes 150mb of RAM as compared to Lync 2010 which consumed 25 mb. Across the board that’s a substantial increase that is difficult to account for. Server-based application delivery solutions alleviate that but with an added cost in both usability and infrastructure.
Where I really see the layering technologies provide a solution is with the struggle between persistent and non-persistent. Any large organization has an issue with persona management and classifying users tightly enough to allow for non-persistent without dozens or hundreds of images. One approach is the RES/FSLogix method of stacking all the apps into the image and simply controlling visibility and usability. Try that with hundreds of apps and it’s a nightmare. The virtualized app route of AppV or Thinapp has its own challenges in just in time delivery to a user. If I send them to a pooled machine discovery of that session and pushing the package down may take too long from a user perspective. They expect their apps to be there and be launchable at logon.
Unidesk, CloudVolumes (now called AppVolumes I hear), and the smaller companies like them provide what I feel is the long term solution with layering but none of them are truly able to play at the scale I need them to… yet. I think they will get there and in the meantime we continue to evaluate them all, but the space isn’t mature enough for large enterprise usage and there are some sizeable drawbacks to all of the current solutions. I have faith it will be resolved 🙂
Another higher ed. Unidesk customer here and while we’re still early on in our deployment have nothing but good things to say about the product and company. As others have mentioned the ability to have a single gold image of the OS, and applications, used by various groups, on both persistent and non-persistent desktops makes patching easy and efficient. As an added bonus it allows for capacity savings on the storage side which drives down the cost per desktop.
The company itself is very open to feedback from customers through a variety of channels. One of the selling points for me was how they make customers the star, and how willing customers were to advocate for them on their own time. From VMWorld 2013 when I first heard of the company and was surprised to see customers working the booth, to numerous webex sessions where customers demoed their production environments live; the common theme has always been about partnering with customers and making them successful.
Which application delivery approach are you currently using for your VDI implementation, and how well is it working for you?
We have here an engineering and architecture VDI user base of about 1k desktops… we have been using VDI for just over 3 years finally feel we are approaching where we need to be with the platform. We tried a number of different application delivery methods, from updating multiple gold images and long weekend recomposes to attempting to use ThinApp. What we found was that even if we could stager the updates to avoid performance issues they were long cumbersome tasks that required our Tier 3 IT folks to maintain and deal with. ThinApp was a much better delivery method but was very difficult and time consuming to package applications, yet our robust engineering applications users needed such as Autodesk Revit and AutoCad.
The product we found that made it all possible for us is Unidesk. It simplified every aspect of application delivery. We went from 6 gold images to only one, it made packaging applications that would have taken days and reduced that to hours or even minutes (even complicated engineering applications). If that wasn’t already a huge advantage we were then able to delegate application delivery to Tier 1 IT staff. This aspect also made a huge impact on our customer’s perception of VDI. They went from requesting an app, then having to wait up to two weeks until a recompose was scheduled before Unidesk, to getting the application pretty much on demand when they call our service desk. Unidesk did a great job of taking something complex, and making it simple. Today we have over 100 applications in Unidesk and close to 1,000 desktops.
We started VDI just over four years ago with high hopes and one year into the project, were very close to calling it a failure. The biggest reason for reaching that point was the inability to handle the delivery of apps. We simply couldn’t not virtualize the unique apps we experienced in our government offices. To overcome that issue, we discovered Unidesk and in essence started over and within three months exceeded what took us over a year to do in terms of total number of desktops.
Our second issue was that VDI was complicated, my most experienced engineers were having to do all the work to support desktops, which obviously is not an ideal situation. Unidesk again address this issue and at this point, our Tier 1 helpdesk staff do everything related to VDI with the exception of managing the underlying VDI infrastructure.
Today, we have no issues with applications and only one gold image to manage. The support time when issues occur are reduced dramatically over that of a physical desktop. At this point, we always try to deploy a virtual desktop when evaluating our desktop needs.
As an early adopter of VDI (back in 2008), I spent my fair share of time fighting the VDI plague. We started by making full clones of a 20 GB Windows XP image for 300 users. This was before linked clones existed. Imagine how much storage we were burning. We were essentially managing virtual desktops just like physical machines as applications, Windows updates and Antivirus definitions were pushed down to all virtual desktops. Yes, our storage couldn’t keep up.
Our users enjoyed the convenience of connecting to a machine from anywhere, especially without the need for VPN connectivity. Though, we still had our complaints about performance. Profile management and application delivery were also issues for us.
As the technology improved, things started to get better. PCoIP also came along and improved the user experience a bit. We began using linked clones and application virtualization with ThinApp. This was the holy grail of VDI at the time, having one image or template to maintain. It was great, but my users hated me for refreshing desktops and nuking their personalization and user-installed apps on a regular basis.
We began using third party profile management tools…. Yikes.
After banging my head against the wall, I realized that what I wanted was desktops that felt like persistent machines, but gave me the benefit of managing them like non-persistent linked clones. I came across Unidesk while searching for a better way to do things.
Unidesk gave me just that, the ability to keep the holy grail of true single image management while providing unique, a la carte desktops for each of my users. We could provision desktops quickly and consistently, with layered applications that were simple to package.
The applications get installed once and the layer gets mounted on the machine that need it. You update it when necessary and mount the new version on a scheduled or individual basis. No more pushing updates, deploying applications or mass anti-virus scanning. We handled that with vShield and Kaspersky. Desktops became easy to manage again.
I have deployed Unidesk as a customer at two different jobs. I now work as a Consulting Architect for a firm that specializes in VDI deployments. Unidesk fits into every VDI opportunity, persistent or not. It has allowed many of our customers to be successful with VDI.
Feel free to contact me for questions or comments.
Another Healthcare/Higher ed. Unidesk customer here. We run Unidesk on VMware Horizon View. Cisco UCS for our compute and Pure Storage for our SAN.
I have to agree with all of the points already made in the blog and the comments here. VDI suffers constantly from two main issues, shared storage performance and managability at scale. These are the two big things that everyone has to look at for any VDI project that’s going to go beyond ~10 VMs.
We’re seeing storage redesigns and new markets come to light from VDI as you said. The next steps will be finding those managability headaches and addressing them in the same fashion. We went with Unidesk and absolutly love it.
We ran a VDI POC on purely View 4.0 and it worked well until we started looking at scaling past POC stages. We quickly had so many base images for just ~100 VMs that scaling past 1,000 was un-imaginable.
We did a POC with Unidesk and it immediately solved both of these issues. Their layering technology is fantastic for getting management of VDI under control at scale. We run our VDI on Pure Storage and the combo is just unbeatable.
Pure Storage is an all flash array that lives by their deduplication and compression. Add Unidesk on top of that to streamline applications into layers that are complete copies of each other and I store only one copy of all of my data on the SAN with all flash performance. Add in Unidesk personalization layer backups being a copy of just another layer and now I have VMs that save me space and I have backups to roll back to when someone breaks an app basically for free.
Now we’re at 1,000 VMs and looking at our end goal of 3,000 with only other project timelines holding us back. There are ways to make it work!
Mark,
I also love VDI and I am confident that I can deploy a fast, reliable, and far more secure VDI system much quicker and cheaper than someone else could deploy traditional hardware with laptops and desktops. Not only that but a VDI system done right will also require a lot less care and feeding from your support folks.
Here at Egan (a commercial construction company located in Minneapolis, MN) we utilize vmware, Unidesk, and Tegile and dummy terminals to deploy VDI to our users. With Unidesk layering it is far faster and easier to deploy application changes to our users. If there is a change to a rolled out app we simply open up that layer, modify it and then choose which desktops to deploy it to. It is so easy to use Unidesk we actually have interns do this much of the time. And we roll the software out during the day so that my team can get home at a decent time. But of course that requires fast storage…
So for storage we use Tegile. Actually our entire VDI deployment lives in SSD only. Our Tegile SAN lets us have 2 separate pools of data. We have a hybrid pool where most of our servers live and an SSD only pool where our VDI and databases live. The reason we can afford to have all of our VDI in SSD is because Tegile supports compression and Dedup! On our old SAN we used about 9TB of data to store our VDI environment. On Tegile that same environment plus some now lives on about 1/2TB (or 500GB). And the great part is we can easily afford to have that small amount of data live in SSD. And of course because of the speed provided by the SSD we can roll out any layer a user may need and we can do it quickly, during the day without bogging down our users.
I guess that in a nutshell is why I so love VDI. But I see your point about it being difficult if you were to choose the wrong partners/vendors to help you.
Healthcare : When first discussing VDI the Sys Admin begins to think they can reinvent the desktop experience with a server based end point. They are quick to adopt, spend spend spend, and implement systems without completing a functional use case.
Understand that bland one application desktop (exam rooms) do not require a lot of integration. VMWare Linked-Clone with a flash based array (I prefer Nimble Storage) can result in an improved experience for that deployment. This also lets you sort out the difficulties with VDI before going full replacement of PCs with VDI.
Once you can accomplish a simple deployment, move to replacing people’s personal desktop. Create a standard image that works and works well, stamp an image to deploy, give it to your helpdesk operators and non-decision makers. As soon as you think you have it almost perfect you will run into the bottleneck above.
You will need a way to update/refine deployments to people based on departments. Accountants need different software than the front desk (Quickbooks vs Webcam). Software Engineers need different products than medical professionals (Visio vs PACS). All those licenses need to be managed accounted for and not wasted.
We use Unidesk application to let divisions, departments, and companies, have exact personalized desktops with the applications they need. Unidesk offers direct level 2 support for help with layering. Recipes for helping activation of software in VDI. Technical aces that are focused on your successful VDI deployment. Once deployed you can mix, stack, rack and deploy hundreds of desktops. The Unidesk Web Interface is simple enough to give VDI deployment to your level 1 or level 2 people. All the features of Unidesk lets you move on to your next IT hurdle and put VDI under the accomplishment column.
First, let me say that the research document is excellent. I would highly recommend it. I agree with many of the comments the others have made. In fact, I thought David Dunaway was a member of my team reading his post. I work for a government organization. We have our fair share of battle scars from storage. I may or may not be typing this from under my desk just thinking about it. Applications are, by far, our biggest challenge. Since the work of my organization is so diverse, the way desktops are used is far from typical. This makes management on virtual AND physical desktops a challenge. I can take all my lessons learned, which I can’t go into, and I only have three words: View, Nutanix, and Unidesk. As to our success, it’s too early to say. Performance at scale will determine the outcome. This VDI thing, and EUC in general, is a journey. Just when we think we know what we’re doing, the tech changes. Truthfully, that’s why I love working in this field. Now there is the introduction of AppVolumes. I think it’s exciting that there are multiple players in this space. It will only continue to drive the rapid rate of innovation, which benefits the industry, it’s customers, and our end users. Thank you, Mark, for doing the hard work and research then giving us the opportunity to weigh in.
Another UniDesk Customer here, the thing that amazes me the most is that Citrix and VMWare both claim to be leaders in virtualization, however they still seem to struggle in providing a true Desktop virtualization solution. To me the result will always be a hybrid solution. In other words it will be a mix of application layering and application virtualization. For products like Office, Java etc application virtualization is a good fit since you can deliver what’s needed and not the entire suite. However especially in Health care such applications are far and few. That is where application layering to me hits the mark. UniDesk understands this problem and solves it in the most efficient and convenient manner. While the early bottlenecks of storage, memory are no longer the hindrance for VDI, Video and Audio still is a challenge.
We are also a big fan of Unidesk. We were looking for a way to lower our storage footprint while improving what we found to be a clunky application delivery system with Thinapp. Assigning just 10 thinly provisioned gigabytes per new desktop (for their personalization layer), we are able to cut out enormous amounts of storage overhead. We are currently running on an Equallogic 15k SAN as it has performed well within our limits. We have Violin flash storage in the wings for once we increase our desktop presence and IOPS become an issue.
We’re a smaller VDI shop with nearly 100 desktops, but Unidesk handles our environment like a champ. We have nearly 40 application layers, some with multiple applications, drivers, and the like. Their template system is intuitive and the layer versioning has saved us many times when we identified a flaw in one of our layer deploys. Their support is excellent, and they tend to QA and release major revs right around the same time VMware patches View. They’re staying with the times and maintaining relevance. Big fan.
Our bottlenecks tend to be with our zero clients. Getting a zero client to perform well with multimedia components (camera, audio) is challenging. Video conferencing, soft phone integration and the like work like a gem when passing through a laptop/desktop’s View client. However, we continue to investigate the best local PCoIP conduit for our user base in house.
Greetings. I am the Unidesk CEO, so confess a non-objective point of view. But I must request to weigh in on two points.
First, Mr. Lockwood, thank you for your thoughtful, practical insight. I hope that organizations who have struggled with VDI in the past will be inspired to try again, and that those, just beginning, will heed your counsel.
Second, on behalf of the entire team at Unidesk, I must thank our extraordinary customers for their comments and advocacy here. Our mission at Unidesk is to be more than a software company. We seek to assist our customers with their whole VDI project and measure ourselves on their overall success. It is rewarding to hear that we are accomplishing that. But not all of our customers are successful and success in IT is not permanent, so we must continue to improve. We will not relent.
Hi Mr lookwood thank you for insight about the VDI bottlenecks.
I am a IT consultant VDI implementation. I have come across the major lamentation projects. I must agree all you have mentioned the bottlenecks of VDI implementation. I have felt there is another bottleneck missing in your blog.
NW latency is the sting of the most common deployments. I could see NW latency also one of the most common bottleneck for the VDI performance. Due to latency issues VDI launching is slow and it hampers the user VDI performance. I think you need address the NW latency bottleneck.
We as well love VDI and cannot imagine it without Unidesk. We are a law firm and are 100% VDI with 80 virtual desktops.
Ease of installing and updating applications, fast break/fix capablilities and storage savings are just a few reasons that we went with Unidesk from the get go and havn’t looked back. Unidesk paired with Nimble storage has made VDI a success for us. I can’t envision ever going back to physical desktops.
We are using Unidesk and VMware View to deliver a huge variety of applications to Provisioning, Operations, Accounting and Development staff at a telecom of roughly 100 employees. Our provisioning people are continually working in a huge and diverse array of Telco provisioning portals which require a variety of java levels to run and Unidesk allows us to manage multiple java installs in order to ensure access to all tools. NOC staff works in VDI to run legacy java based element managers that again require a diverse (and dated) set of java tools. We are delivering a large and sophisticated developer’s desktop environment including large scale dev tools to our in-house development team. We used Unidesk based VDI to onboard 30+ employees across 4 time zones after an acquisition while they were still working in non-secured legacy environments while building the tools they needed to continue to work while integrating into the parent company on the fly. All of this was accomplished with an Engineering/Help Desk staff of 3 people. None of this would be possible at that staffing level if we were continually managing these end user environments on traditional desktops.
Ok… so Unidesk is great. I like all of the feedback. The questions I have are: What has VDI done to your costs? I understand the soft benefits, but what have the various combinations of VDI, Unidesk, new storage arrays, etc., done to the economics? What were your per-user cost for deploying and managing a PC vs your per-user cost for VDI?
I certainly understand that cost is not the only component and that there are other value drivers (like the acquisition in the post above), but there has to be demonstrable economic benefit before execs sign off on a VDI initiative.
very good
I agree that Storage is a bottleneck for VDI. But then again, all is related to the user experience, all parts of the architecture need to be sized for optimal performance: Storage, network, GPU/CPU, etc and especialy the WAN can be a dealbreaker, all can be optimised at a cost.
When storage is the bottleneck of VDI, connectivty, or the lack thereof is it’s achilles’ heel.