DaaS Incrementalism

So you've decided that you're going to use VDI for some of your desktops. Great!

So you've decided that you're going to use VDI for some of your desktops. Great!

Then you decide that you're going to pay for someone else to build and operate that VDI in their cloud environment. So now your paying for cloud desktops. (DaaS) Great!

Thanks to today's storage technologies, you can get fully persistent DaaS desktops for a decent price. This means you can easily replicate what your current desktop environment looks like in the cloud. Great!

Your current desktops have drives mapped to do storage. When they were on-premises, this was just regular network drives. But now that your desktops are going to be in your cloud, where will your file shares live? Will you map drives from your cloud desktops across the internet to your on-premises file servers? Hmm.. that doesn't make much sense. It's probably better to buy some file server services from your cloud provider. So now you're paying for cloud desktops and cloud file servers.

Of course, you're not moving all your desktops to the cloud, which means you need to keep operating your on-premises file servers and you need to figure out how to do some kind of file replication between your on premises servers and the file servers at your cloud provider. So now you're paying for cloud desktops and cloud file servers and replication services.

Now you think about your applications and the databases that run them. If you install those desktop applications on your cloud desktops, how will they perform when they connect to their databases running at your on premises location? Hmm, you either need to use XenApp to host the apps or buy database services from your cloud provider (along with replication software). So now you're paying for cloud desktops and cloud file servers and replication services and database software.

Okay, so how are your users going to login to their desktops? You already have Active Directory and all the users and groups and policies set up, so let's just extend that to the DaaS provider. (After all, you don't want your users to have to use a different username and password to authenticate to their cloud-based desktops, and you want to be able to manage those desktops with your existing management tools and methodologies.) So now you're paying for cloud desktops and cloud file servers and replication services and database software and Active Directory sync tools.

Hmm. There's a lot of IP in your cloud environment right now? What happens if your provider goes under? You better make sure you're getting regular backups of your environment. So now you're paying for cloud desktops and cloud file servers and replication services and database software and Active Directory sync tools and backup services.

But now you're done, right?


Join the conversation


Send me notifications when other members comment.

Please create a username to comment.

:)) entertaining read. That's why I've maintained that if the desktops go to the cloud everything goes to the cloud. The question now becomes can you move your datacenter to the cloud? And is it financially viable? And is it operationally more efficient? In many cases I've had customers give up on premises datacenter and are 100% IaaS. But there are plenty of AS400 and other non virtualizable systems to put a wrench in my suggestion.... Unless the DaaS workload is self sufficient you are absolutely correct, it's not a good fit.




does the IaaS definition exclude LPAR? LPAR virtualization existed long before x86 was 'virtualized'. AS400 can easily be virtualized these days as well.

Even OpenVMS (R.I.P.) eventually introduced virtualization support in 2001.

Saying that AS/400 and others are non-virtualizable is simply not true.

Many customers and consultants might just think that IaaS is limited to x86 compute.




yes IaaS excludes LPARs, and yes I am referring to x86 virtualization.



IaaS is crazy expensive.  You still can't fire your network admins, they still have the servers to manage, users to deal with, upgrades to perform, etc, etc.  Sure maybe you'll save a couple of bucks (I doubt it though) being that they are not administering storage or x86 hardware to run your hypervisor, but I've looked at going that route and for what 3 providers wanted, i could refresh all of my hardware once per year practically.  And for bandwidth???  Who knows how much I would need to cough up to ATT to get a large pipe to corporate for my vpn tunnel back to the data center.  You just can't run 100% DaaS, dedicated workstations and laptops will always be around in some form or fashion.  

PS - And every time I added storage, memory or CPU, my monthly price goes up, that's not true with on-prem, your operating costs remain fixed for x number of resources till you need more.  


Many VDI environments are in co-lo data centers today as exisiting datacenter capacity is a premium for many.

Also as part of a broader infrastructure strategy, many simply refuse to build out new data centers and depreciating them over time. This effect is magnified when you consider the desire to refresh hardware faster due to industry innovation. How many people are investing to refresh their existing data centers vs. just squeezing the lemon harder? If anything, I've seen people refactor for new must keep on-site workloads and looking actively to move non-critical workloads elsewhere, whether public cloud or private co-lo data centers.


So, AS/400 cloud providers shouldn't call themselves IaaS providers then?

Alluding that AS/400 is "non-virtualizable" is just factually incorrect.

And AS/400 cloud providers exist and I'm sure they're making good money with their offerings.


Think you've basically summed up why VDI has/will continue to be a niche technology - even with DaaS.

Users want to use notebooks or shiny silver Macs.  They'll increasingly access native apps and web applications.  If they HAVE to, they'll access some legacy application via VDI/XenApp.

That's not to say there isn't a nice little ecosystem around VDI/published apps that has to be fed  :-)


@christoph, take a deep breath my friend.... Take a deep breath I was simply referring to x86 IaaS and when I say non virtualizable I mean it on the x86 platform. Sure CSP for AS400 exist but that doesn't solve the problem if my AS400 is with one CSP and my x86 is on Amazon then my problem persists, it's as if I didn't virtualized it. Unless they are all on the same platform or in the same datacenter or latency is super low it would not matter.

I don't understand what you are arguing, I'm not disagreeing with you



Great points Brian!  I have been preaching this for quite some time now as well.  This is why DaaS is way more hype than reality and a waste of time and money for most enterprise organizations.  Legacy desktops and applications are messy; drive mappings, SQL connections, printer connections, etc…  What we have referred to as the "golden rule" for Citrix implementations for well over a decade is that your desktop/application needs to be in the same data center as your data. That is why WinFrame/MetaFrame/XenApp became so successful. I cannot for the life of me understand why anyone would even consider moving to a DaaS solution without simultaneously migrating their servers as well.

The greatest thing about VDI is that it is an enabler of data center consolidation.  That involves customers collapsing smaller data centers into larger ones, or simply migrating all their servers to a public cloud.  When customers started collapsing their smaller branch data centers years ago, they realized that user experience sucked because the desktop and application executables were now separated from the data servers.  XenApp has been a great remedy for some of these issues, but not everything runs on XenApp.  With VDI we now have the ability to move the desktops AND the servers to a consolidated data center at the same time and still give users great performance and compatibility.  DaaS offers absolutely zero benefit for most organizations unless the servers and data migrate to the same cloud at the same time as well.


I am not sure why DAAS looks complicated in this conversation and un benefitial when the world is moving towards it. We ourselves have been using DAAS in house for a long time and just outsourced hosting of it to desktopasservice.com for enhanced reliability and reduced our cost of managing and letting the experts do it. However, our critical servers are still with us and only users personal / shared data has been moved to the cloud. The benefit of DAAS is work from anywhere and BYOD. In the horrible climates we have been experiencing in our region our users can work from home, without needing to feel the pain of difference of an office PC and home pc  as DAAS is accessed from all locations and it actually has been working cheaper and more productive. Don't you agree that DAAS requirements could vary based on each user? Isnt it easier for a user to use the same desktop from home, office, cell phone / laptop than getting comfortable with changing desktops? DAAS is too complex to design and host yourself but easy to just subscribe. Isnt it?



I tried DestopasService.com and found it very good!.

Anybody knows is there is a human behind it you can contact?

My clients and partners don't feel at ease not knowing who is behind this excellent DAAS Company.