In July, I had the pleasure of speaking at BriForum in Boston. BriForum is a popular vendor-neutral conference focused on virtualization and mobility technologies. My first session was titled Where To Stick It: VDI Storage Considerations; you can read more about that here. My second session was titled DIY DaaS: Creating your own Desktop-as-a-Service.
Unfortunately, the videos from my BriForum sessions are not currently available, but I would like to cover some of the main points of DaaS, based on my experience as a Virtualization Architect and current industry trends. During the session, we focused on these main elements:
• Understanding how DaaS and VDI relate
• Determining what DaaS means
• Creating your own DaaS Model
Understanding how DaaS and VDI relate
You hear a lot of terms when it comes to virtualization: server virtualization, session based computing, desktop virtualization, workload virtualization, workspace, layering, etc. These are ambiguous terms. The same thing applies to VDI and Daas: VDI = Virtual Desktop Infrastructure, DaaS = Desktop as a Service. What’s the difference between the two? NOT MUCH!
Figure 1 – Overlap of VDI and DaaS
One of the key differentiators is perception. VDI tends to be viewed as an internal capitalized expense, which is part of the internal infrastructure. DaaS is viewed as an external operational expense, and viewed as simply a service. Ultimately, though, both are about the DESKTOP.
Another way to think about DaaS vs VDI is Lease vs Buy. Assume your desktop can be provided through DaaS for $70/month, which works out to $840 per year. With a 3 year agreement and 10% contract discount, you are looking at a total cost of around $2300. There are minimal setup costs, you just pay the bill and voila! Of course that is for just one user. What if you have 1,000 users? Do you want to pay close to a million dollars a year in desktop leasing?
Well, maybe you do… it’s an operational expense not a capital one. Or maybe you look at that 2-3 million you would be spending in a 3 year period and think, “If I build it myself, I can capitalize that expense, spread it out over time, and have exactly what I want, how I want… and I can keep it longer, thus reducing my TCO…” Either way is a viable option.
What does DaaS mean to you?
As previously mentioned, DaaS is an ambiguous term, so you need to decide what it means to you.
DaaS can be a mix of any of these ingredients:
- Hosted Desktops with a Service Provider
- Applications “in the cloud”
- Remote Windows Desktops
- Citrix XenDesktop VDI
- Desktops “in the cloud”
- Hosted Apps in the datacenter
- Remote connectivity to Desktops
- VMware Horizon View
Let’s compare DaaS and VDI a little more closely, to highlight the differences. These definitions can further refine what you need DaaS to mean to your organization:
But what if you could have the best of both worlds?
- Merge the best of both concepts
- Create and manage your internal infrastructure
- Use the service principles of DaaS
Creating your own DaaS model
The first step, when creating a DaaS model, is to identify your business drivers.
- WHY are you investing in VDI or DaaS?
- WHAT is the benefit you are seeking?
- WHO is sponsoring the imitative?
- WHAT is your end-state goal?
- WHAT constraints are there?
Then you need to understand your user, software, data, security, and service requirements. This includes items such as:
- Who are our target users?
- What applications and business processes do they need?
- Are there any special requirements you must support?
- Do we need dedicated or shared resources?
- What about network isolation?
- What operating systems must you support?
- What level of availability are you planning?
- What level of fault tolerance is required?
- What level of performance is expected?
The focus of any DaaS model should be SERVICE. Whether you are doing this internally or externally, using cloud platforms or traditional virtualization, the focus on service is key. The first step is to define your service level agreements to determine what your needs and requirements are.
Figure 2 – Service Level Cost vs. Complexity
Once you determine WHAT you need, you can then focus on HOW to deliver it. Insourcing or outsourcing does no good if the requirements are not clearly defined. If you want to deliver your DaaS using basic VDI architecture, you can leverage traditional VMware View, Citrix XenDesktop, or Citrix XenApp architectures.
For internal clouds, you can build your platform on Microsoft Hyper-V with Azure Pack, VMware vCloud Suite, OpenStack, or Citrix CloudPlatform. For external providers, you can leverage Amazon Workspaces, Microsoft RemoteApp, or even the Google Cloud Platform, among others. Of course, if you want to outsource your entire desktop strategy, you can do that to. Companies like o4it with CloudNow, NaviSite, nGenx, and Apps4Rent offer complete ready-to-consume DaaS solutions.
Granted, this is all just the tip of the iceberg; there are still considerations around financial models, licensing, maintenance, security, and more. To learn more about DaaS models, check out Brian Madden’s Desktop as a Service book. Also, to see more detail on the planning phases as well as building your own solutions, check out my Citrix XenApp 7.5 Desktop Virtualization Solutions book.
Ultimately, there is no magic bullet, there is no one-size fits all solution. However, at OpenSky, we have the tools, experience, and expertise to help you plan your desktop strategies, including assessment and design around your VDI or DaaS objectives.