VDI Options - Modernization
-
@dafyre said in VDI Options - Modernization:
@travisdh1 said in VDI Options - Modernization:
@jimmy9008 said in VDI Options - Modernization:
@travisdh1 said in VDI Options - Modernization:
@jimmy9008 said in VDI Options - Modernization:
@travisdh1 said in VDI Options - Modernization:
@jimmy9008 said in VDI Options - Modernization:
I get what ya'll are saying but thats just not how it is here. My options are replace what is there with new, or keep what is there and let it grow older.
I'll keep looking at options on my own, but thanks folks.
If you just want to buy a solution without doing your homework to figure out what's right for the business, just get new servers and keep paying the crazy license fees for VMWare/Citrix (I'm assuming you've got the HA VMWare license.)
Without knowing what apps are running in the VDI, all we can do is generalize.
Are you stuck with VMWare and/or Citrix because of management? Big cost savings in moving away from those, even if you keep paying for support IE: Scale or Starwind
More details would be needed to make any solid recommendations.
I am more than capable of being able to appraise solutions to meet our business needs. My question was asking for a list of solutions "What would you suggest we look at?", not to be told to not look at VDI as its wrong. I'll decide that. I was hoping the community could point me to solutions, vendors, resources which you have used and had experience of. I see the people on here as experienced so wanted to ask here, I should have just looked at g2.
Well, I think @scottalanmiller already explained much better than I ever could that VDI Modernization is a contradiction in terms. If you're stuck using VDI, then you by definition are not modernizing.
As to different platforms to run it on, that's why I suggested Scale or Starwind to run the Citrix solution.
Oh come on, seriously. How on earth is that a contradiction in terms. I like this forum but some time people on it can be ridiculous with rubbish like that. You can modernize many things in a wide range of ways and saying that a 'VDI cannot be modernized' as that is not how you think something should be done is just pure rubbish.
You are running Windows Server 2008r2, and are considering migrating to Windows Server 2022!... that is not modernizing... your workload should be SaaS/Cloud! Yeah, BS. You can modernize without being SaaS/Online services.
Oh! You want to modernize and move from HDD/Spinners to NVMe... well tough luck, you cant modernize like that dumbass... your storage should be a blob in Azure.. Local storage, pfft. No way is that 'modern' anymore!
You want tomodernize your compute and use PMEM. Oh shoot! That cant be modernized as you should be using a VM in AWS. BS!
You can take outdated infrastructure and modernize it in many ways - just because ya'll believe in narrow minded dogmatic BS like 'my way is the right way' you think this is a contradiction in terms. LOL. WOW.
You can take old VDI infrastructure and modernize it. Contradiction my ass! If this is what I can expect from this forum I may as well post on Spiceworks. Gosh.
So, are you even considering basing things on Scale or Starwind as your underlying solution, or are we just ignoring half of what I say?
If you're serious about checking out other platforms, I can vouch for Scale. I had them at my last job, and they now have two x 3 node clusters and the guys there are extremely happy with it.
Do they support their stack using ESXi? We have to stay standardised to that.
Be it a ton of servers backed by a physical SAN, a ton of servers backed by a vSAN, be it a ton of servers with local storage, or whatever - we have to run ESXi.
-
@pete-s said in VDI Options - Modernization:
@jimmy9008 said in VDI Options - Modernization:
@jt1001001 said in VDI Options - Modernization:
@jimmy9008 We have a use case involving a legacy client/server app that we've determined we're going to have to go VDI for in order to secure it. One lousy app for approx 5 users that I hope we eventually move away from. We are currently reviewing Azure VDI for this and it so far will fit the bill though we had to go throught a lot of "hoops" to configure networking, VPN back into our infrastructure, etc. We have not yet presented budget numbers to the bean counters but Im hoping when we do they will see the $$$$$ wasted for 5 users and will force them to a new product.
What other products do you plan to look at? Still VDI or something else? Any experience of VMWare Horizon?
We have around 600 - 1000 users globally (mostly developers) on the VDI I need to replace. The company dictates that the VDI must be in the same datacenter as the rest of the developers environments, so I don't think Azure VDI would work for us because of that mandate.
If you have a solution that works, and at the moment VDI is a must, then it makes no sense to change the fundamentals of what you already have. That's just an unwarranted risk.
So keep Citrix and VMware as is. Just replace the hardware and consolidate it. You are only averaging 16 cores per physical server and 370GB RAM per server if my math is correct. You could easily cram 3 to 8 times as much into each server. 128 cores per server is nothing special today as well as several TBs of RAM. AMD is the leader and the way to go.
You could replace your 20 servers and have 384 cores and up to 12TB of RAM with only three Dell R6525 or R7525 dual CPU servers. You might want 4 or more though. But no need to go to blades when you only need a couple of servers. No need for complex hypervisor management solutions either when you only have a couple of servers.
Use vSAN instead of SAN for the VDI. With the proper drives these servers are certified for ESXi and vSAN. You should use U2 NVMe drives and avoid SAS. It will outperform your old SAN - by a lot.
Since you have 1 PB of data, storage for non-VDI workloads needs to be researched. I think I would want to separate VDI from the rest. Gut feeling would be to have completely separate physical environments for everything VDI related and the rest. Consolidation is good but overconsolidation can be too risky.
This could be a great option. New servers with more horse power, running the existing software stack. One problem is that due to another departments projects the underline storage is going from this solution. The storage is being ripped out leaving 20 servers, which need to be replaced, without any of the 1PB storage.
I was considering Dell PowerMax for this storage presented over iSCSI to a server on the Dell VXRail, also running the VDI.
-
@jimmy9008 said in VDI Options - Modernization:
@pete-s said in VDI Options - Modernization:
@jimmy9008 said in VDI Options - Modernization:
@jt1001001 said in VDI Options - Modernization:
@jimmy9008 We have a use case involving a legacy client/server app that we've determined we're going to have to go VDI for in order to secure it. One lousy app for approx 5 users that I hope we eventually move away from. We are currently reviewing Azure VDI for this and it so far will fit the bill though we had to go throught a lot of "hoops" to configure networking, VPN back into our infrastructure, etc. We have not yet presented budget numbers to the bean counters but Im hoping when we do they will see the $$$$$ wasted for 5 users and will force them to a new product.
What other products do you plan to look at? Still VDI or something else? Any experience of VMWare Horizon?
We have around 600 - 1000 users globally (mostly developers) on the VDI I need to replace. The company dictates that the VDI must be in the same datacenter as the rest of the developers environments, so I don't think Azure VDI would work for us because of that mandate.
If you have a solution that works, and at the moment VDI is a must, then it makes no sense to change the fundamentals of what you already have. That's just an unwarranted risk.
So keep Citrix and VMware as is. Just replace the hardware and consolidate it. You are only averaging 16 cores per physical server and 370GB RAM per server if my math is correct. You could easily cram 3 to 8 times as much into each server. 128 cores per server is nothing special today as well as several TBs of RAM. AMD is the leader and the way to go.
You could replace your 20 servers and have 384 cores and up to 12TB of RAM with only three Dell R6525 or R7525 dual CPU servers. You might want 4 or more though. But no need to go to blades when you only need a couple of servers. No need for complex hypervisor management solutions either when you only have a couple of servers.
Use vSAN instead of SAN for the VDI. With the proper drives these servers are certified for ESXi and vSAN. You should use U2 NVMe drives and avoid SAS. It will outperform your old SAN - by a lot.
Since you have 1 PB of data, storage for non-VDI workloads needs to be researched. I think I would want to separate VDI from the rest. Gut feeling would be to have completely separate physical environments for everything VDI related and the rest. Consolidation is good but overconsolidation can be too risky.
This could be a great option. New servers with more horse power, running the existing software stack. One problem is that due to another departments projects the underline storage is going from this solution. The storage is being ripped out leaving 20 servers, which need to be replaced, without any of the 1PB storage.
I was considering Dell PowerMax for this storage presented over iSCSI to a server on the Dell VXRail, also running the VDI.
For compliance and security our development environment is totally separate from the production environment. So the cluster of development servers, network hardware etc lives in their own ecosystem.
We have large enterprise customers and they are splitting up their workloads in the similar ways. They're running on EXSi and vSAN but everything lives in different pools and on different networks.
I would try to keep the VDI solution separate including it's storage. Besides technical reasons like noisy neighbor and security there is also management reasons - for example the other department ripping out the storage is one such reason to avoid having all your eggs in the same basket. With software and OS you have dependencies and you have the same type of dependencies in the organization when it comes to who manages, who pays, who decides when to upgrade etc.
That's what I mean about overconsolidation. Just because it's technically possible to put everything into one box, doesn't mean you always should.
-
@jimmy9008 M365 is what covers the licencing, so Compute is your biggest cost, and it's just like compute for any Azure VM. Use Reserved instanced if you will have some VMs on throughout a 24 hour period. Use burstable VMs if a VM will be kept on continuously and doesn't need a ton of power. Or just turn the VMs off (deallocate) when not in use. At my last company, we found turning VMs off to be the most effective with the personal VM route. If using host pool, you would only keep enough VMs online to meet demand.
-
@jimmy9008 said in VDI Options - Modernization:
Oh come on, seriously. How on earth is that a contradiction in terms.
No, he's totally right. VDI is a legacy holdover technology specifically a high cost to allow those that have not modernized to get some advantages or access that their legacy infrastructure doesn't otherwise allow. If you have modernized, VDI has no real value.
VDI is a stop gap until you modernize so that you don't give up all of the flexibility that more modern network designs allow. And it's a Windows-only technology. It's not just legacy from a design standpoint, but effectively a "unique to Windows" problem. That's not good or bad, just that modern OS agnostic designs naturally don't need it, even when overall not modernized, because it's two different aspects of legacy to need VDI.
-
@jimmy9008 said in VDI Options - Modernization:
You are running Windows Server 2008r2, and are considering migrating to Windows Server 2022!... that is not modernizing... your workload should be SaaS/Cloud! Yeah, BS. You can modernize without being SaaS/Online services.
Sure. But that's we weird reaction. Absolutely no one has even hinted that you should be looking at SaaS or Cloud. Why would you jump to that? Why mention stuff no one is talking about.
Just because you update your operating system, but not your infrastructure or design, doesn't make you modern.
-
@jimmy9008 said in VDI Options - Modernization:
I get what ya'll are saying but thats just not how it is here. My options are replace what is there with new, or keep what is there and let it grow older.
I'll keep looking at options on my own, but thanks folks.
I'm confused. You are asking for advice, and there's no legitimate way to give that advice without asking these questions - if we didn't ask these questions, nothing we told you would have any value. You work in IT, you know we'd be negligent if we didn't do that. If we just said "here's a way to do this task without considering your needs" we aren't IT, we're just sales people.
-
@dashrender said in VDI Options - Modernization:
@jimmy9008 said in VDI Options - Modernization:
I get what ya'll are saying but thats just not how it is here. My options are replace what is there with new, or keep what is there and let it grow older.
I'll keep looking at options on my own, but thanks folks.
I'm lost - why are you bailing on this thread because one person said VDI is not how you should be moving forward? Other options were presented.
No one said it wasn't how they should move forward. We pointed out the fact that it's a legacy holdover and that by modernizing you don't need it (and haven't ever, VDI didn't come about until it was already a legacy need). That it's legacy doesn't mean it isn't the right way forward. People need to stop thinking that everything legacy is always bad, that's an inappropriate reaction to IT basics.
Old isn't always bad. It's well worth evaluating if there is a reason not to modernize, sure. But just because something is legacy does not automatically make it bad.
-
@jimmy9008 said in VDI Options - Modernization:
@dashrender said in VDI Options - Modernization:
@jimmy9008 said in VDI Options - Modernization:
I get what ya'll are saying but thats just not how it is here. My options are replace what is there with new, or keep what is there and let it grow older.
I'll keep looking at options on my own, but thanks folks.
I'm lost - why are you bailing on this thread because one person said VDI is not how you should be moving forward? Other options were presented.
I don't see value in discussing why we have a VDI. The fact is we do and that will not be changing. Being told 'grumble grumble' that is not how to do it 'grumble grumble' is of no help to me. Regardless of what it does, VDI is staying. My options are keep the old stuff and hope it works for another 5 years until the next cycle, or use the budget I have to replace it for a new VDI stack. The project is not to asses the needs of requiring VDI, but to replace the VDI with a new VDI.
Most of the comments on the thread do not help with that so I gave up with it. Sure, if the project was 'get rid of VD' - but its not.
Somebody suggested Azure VDI, will take a look at that and keep looking at other options.
The value is that if we don't know WHY you have something, then no one can tell you which solution is even an option for you, let alone which one makes sense. This is IT, you can't do any meaningful aspect of your job in any other way.
No one, absolutely no one "needs" VDI, that's not a possibility. It's possible that VDI is a good option for you, but it's impossible that it's a necessity. No one is grumbling. Do not call "doing ethical IT diligence" grumbling.
If you have no control over the environment and have strict constraints, that's fine. But this is an IT question, so we need to know those constraints. No one is judging or saying you are doing something wrong. But we have no idea of the requirements so the first step of any investigation is "what is needed", then "how best to do what is needed." THat's where we are. We don't know why it is needed or when modernization will take place. We have to know if the company is on the path to modernizing and so VDI might be a very short term stop gap, or the plan is that modernization isn't even on the radar and VDI might be needed for decades to come (a common thing.) And we need to know all those needs and more to know what types of VDI, products, approaches, etc. will best apply.
-
@jimmy9008 said in VDI Options - Modernization:
Sure, if the project was 'get rid of VD' - but its not.
Why not change the project to "do whatever is best" instead of "deploy what we say"? Maybe you can, maybe you can't. We don't know. It sounds like you are being told to blindly deploy something without evaluating any needs. Do they realize that they are forcing you into that? Maybe they do, then maybe you shouldn't push back. Or maybe they have no idea that they are suggesting something that is considered legacy and a stop gap for companies trying to get to modernization and would really appreciate your insight on how to modernize and your evaluation of it that is or isn't valuable in this case.
If the company is a real business, why not go to the powers that be and ask "Have we evaluated the value in modernizing rather than throwing money at bandaiding legacy solutions? Maybe there's a better way, we should run the numbers and see." It's a chance to shine, a chance for IT to prove it can provide value, that it is doing its job. If they shoot you down, that's fine, you did your due diligence. That someone dictated a VDI project, but then assigned you to find VDI options, guarantees that they didn't do their IT diligence at all... because in order to make that VDI decision, they would have to have known the options in order to have evaluated it in the first place. So we can guarantee that something is amiss. We totally understand if you feel politically that your organization is vindictive and would punish you for attempting to do a good job and expose upper managers as just spending money to try to not have to do proper evaluations or just looking the other way as money is spend because it's the easy thing to do, but if so, just say so, there's no reason to feel like it is something personal.
Don't go after it assuming that legacy is bad and modern is good, that's the wrong attitude. Lead white paint still looks the best. Fountain pens are still a joy to write with. Reading paper books is still easy for your brain to retain information (and easy to read under a tree in the park.) Legacy is not wrong by definition, modern is not right by definition. But it's highly useful to know where you sit when evaluating, but to always evaluate.
-
@jimmy9008 said in VDI Options - Modernization:
Oh! You want to modernize and move from HDD/Spinners to NVMe... well tough luck, you cant modernize like that dumbass... your storage should be a blob in Azure.. Local storage, pfft. No way is that 'modern' anymore!
You want tomodernize your compute and use PMEM. Oh shoot! That cant be modernized as you should be using a VM in AWS. BS!All of these statements make zero sense in the context of this thread. None of that is modernization.
I think you are using "blinding do the popular thing" with "moderizing". That's never the meaning. Similarly, people often use the term "old school" to actually mean "just doing what they were told without evaluating." All terms used negatively kind of mean the same thing, lol.
-
@jimmy9008 said in VDI Options - Modernization:
You can take outdated infrastructure and modernize it in many ways - just because ya'll believe in narrow minded dogmatic BS like 'my way is the right way' you think this is a contradiction in terms. LOL. WOW.
You can take old VDI infrastructure and modernize it. Contradiction my ass! If this is what I can expect from this forum I may as well post on Spiceworks. Gosh.Of course you can. THat's OUR point. We were telling you that there are other options and you need to evaluate. But VDI IS legacy, it's entire purpose is legacy. It is conceptually legacy. You can't modernize legacy, by definition. You can deploy legacy systems fresh, but that's not modernizing. You can manufacture a brand new Model T, but it's still a legacy car, even if you build it from scratch today. It's new, but not modern.
You can make new VDI, but you can't make modern VDI.
-
@jimmy9008 said in VDI Options - Modernization:
Just had a very quick look at Azure Virtual Desktop and the calculator for 3 years up-front shows around 1.5m usd. Were looing upper limit of 1m usd, which would cover us for the next 5-7 years, making Azure look expensive.
Azure is generally the most expensive solution you can imagine for anything. Azure has a brand name that is worth a fortune regardless of what they do under the hood, so MS will always capitalize on that. If you want to consider price (as you should), you can generally just rule Azure out. Once in awhile they will come in with similar prices to other solutions, but never cheaper (that I've seen) and often like this, wildly more expensive.
Cloud, in general, is a price premium solution. If you haven't already totally modernized your network, cloud should mostly be off of the table. Moving legacy apps and designs to cloud is the worst of all worlds. If you need like a single VDI instance and have no office, sure, Azure will kick butt. But at any scale, building your own and even putting it in colocation will be far, far better.
-
@travisdh1 said in VDI Options - Modernization:
@jimmy9008 said in VDI Options - Modernization:
@travisdh1 said in VDI Options - Modernization:
@jimmy9008 said in VDI Options - Modernization:
I get what ya'll are saying but thats just not how it is here. My options are replace what is there with new, or keep what is there and let it grow older.
I'll keep looking at options on my own, but thanks folks.
If you just want to buy a solution without doing your homework to figure out what's right for the business, just get new servers and keep paying the crazy license fees for VMWare/Citrix (I'm assuming you've got the HA VMWare license.)
Without knowing what apps are running in the VDI, all we can do is generalize.
Are you stuck with VMWare and/or Citrix because of management? Big cost savings in moving away from those, even if you keep paying for support IE: Scale or Starwind
More details would be needed to make any solid recommendations.
I am more than capable of being able to appraise solutions to meet our business needs. My question was asking for a list of solutions "What would you suggest we look at?", not to be told to not look at VDI as its wrong. I'll decide that. I was hoping the community could point me to solutions, vendors, resources which you have used and had experience of. I see the people on here as experienced so wanted to ask here, I should have just looked at g2.
Well, I think @scottalanmiller already explained much better than I ever could that VDI Modernization is a contradiction in terms. If you're stuck using VDI, then you by definition are not modernizing.
As to different platforms to run it on, that's why I suggested Scale or Starwind to run the Citrix solution.
We have a small amount of legacy VDI and we do it on Scale. It's not at scale, just on Scale. For our small scale this works really well.
One MAJOR question is reliability. Scale you pay extra for high reliability. VDI often doesn't need that level of reliability. So that can be a huge decision factor (And one of the reasons we HAVE to know what the use of the VDI is to understand what approaches even make sense.) The majority of our VDI is application testing, so a huge mix of operating systems (hence why VDI instead of terminal services.)
So a highly available VDI approach vs. a normally available VDI approach will make a huge difference in recommendations.
-
@jimmy9008 said in VDI Options - Modernization:
I’ll take a look at Starwind, but not Scale. Unless I am mistaken Scale do not use ESXi as the hypervisor layer. Don’t they use KVM? I didn’t write previously but we have to stay standardized to VMWare. That would remove Scale as an option.
Scale is KVM. That you need to be on VMware changes everything. It's not really VDI (so much) that you are evaluating but ways to deploy VMware for use as VDI. That's a wholly different picture as that rules out basically everything, and certainly all of the interesting stuff.
Starwind could be a great PART of the solution, but they are not the VDI portion. So while I love Starwind, it's not necessarily applicable, at least not entirely, to the discussion as it's the storage component of the hypervisor that you will use for VDI.... so two steps away from the conversation.
If you are stuck on VMware, I'd just install VMware on your own servers, use Starwind IF you need HA, and stick to the VMware stack. Not a lot of other pieces to consider.
Since VMware was a requirement for one piece, are there other requirements also like VMware vSAN instead of Starwind?
-
@jimmy9008 said in VDI Options - Modernization:
For Starwind, I do not see a specific VDI product. The site seems to say they can build for VDI, but it feels a bit ‘Mum and Dad’ shop. The highest model stack listed, well the one on their site, seems a lot smaller than we would need.
Starwind is the top end of the industry. It doesn't get better and they are part of Veeam, so about as far from "Mum and Dad" as you can get. They are who you turn to when VMware's own solutions can't cut it.
However, they are not a VDI vendor. Their storage improves the underlying performance of VMware deployments that are HA to allow for better VDI, that's all. So while they are built "for VDI", it's in the same way that Intel and AMD CPUs are "built for VDI". Any VDI solution might use AMD and Starwind, but they could use Intel and VMware vSAN, too. It's just pieces of the underlying platform on top of which your VDI will be built.
VDI doesn't require any toolset at all, but obviously benefits from them in almost all cases. VDI is a way of licensing and deploying end user systems that's all. So from the context of this conversation, we'd be looking at designs or products existing above the hypervisor (VMware) while the server hardware, storage, hypervisor itself, etc. are components below that level.
-
@pete-s said in VDI Options - Modernization:
I would try to keep the VDI solution separate including it's storage. Besides technical reasons like noisy neighbor and security there is also management reasons - for example the other department ripping out the storage is one such reason to avoid having all your eggs in the same basket. With software and OS you have dependencies and you have the same type of dependencies in the organization when it comes to who manages, who pays, who decides when to upgrade etc.
Starwind actually addresses both of those concerns in their design. Not saying that their design is always best, just saying that noisy neighbour and "other department ripping stuff out" are both able to be protected against with Starwind even in a single pool environment.
-
A key example of where NTG is stuck on something legacy is our financial systems. Due to reasons I don't totally agree with, we are currently on QuickBooks (and I could write a lot about what I think about that and how my suppositions played out, but that's for another time). For now, it is a legacy app, deployed on a legacy platform (by the vendor), hosted in a legacy way. It's SaaS, but legacy (SaaS has existed for a REALLY long time and doesn't have any modernization suggested in its use.) We are stuck with it. But we know it is legacy and causes network design complications. It is isolated and we can work around it mostly. But it is annoying and a red flag, we know it is a place that needs to be modernized and we regularly discuss plans around it.
QuickBooks Online is a modern replacement to QuickBooks, but more limited. We can't use it as it is less capable. In this case, legacy is better. We hate it, but it is better.
-
I know much of what Scott posted most recently in this thread is things he's said before - but why not start out this way - instead of the - that's legacy way... after reading the broken wall of text it definitely comes off much less - you way is dumb, and is more inviting.
-
Now with that said - I'm amazed management is willing to spend over a million dollars updating that PITA of a VDI solution instead of paying developers to make a new system that wouldn't require that VDI knife to do to the job - with the expectation that the long term costs would be much lower.