Amazon announced this week that they are offering their own Desktop as a Service (DaaS) offering. Before I dig into this let’s set the record straight: technically speaking there was absolutely nothing new in this announcement.Many DaaS service providers have been offering near identical services for years. Since I cover this space I’ve had to suffer through the tweets of @guiseblue for the past few years (CEO of a DaaS start-up TUcloud), so I can tell you very confidently that for the most part this has all been done.
- Citrix has many partnerships with service providers that are doing this
- VMware has vCHS and also with the acquisition of Desktone they picked up other service providers doing this
- There have been a ton of leaks about Microsoft planning on doing this
- There are plenty of small guys doing this
So there is nothing drastically new here BUT there are some pretty cool things about this announcement. So lets discuss those:
For those that don’t know Teradici, Teradici is the maker of the PCoIP protocol and they make zero clients which are a thin client alternative ASIC solution that are more secure, easier to manage, etc, etc. This announcement is huge win for Teradici and in my opinion its on the same scale of magnitude when they partnered with VMware to put PCoIP in VMware’s Horizon View product.
Why would Amazon partner with Teradici? Well I could go into the business reasons (I can think of many), but I’ll focus on the technical reasons which boils down to this: PCoIP is a powerful protocol. If you are going to offer a DaaS offering protocol is going to be an important aspect of the user experience (UX). Protocol is the foundation of the UX so you need a strong foundation to start from. VMware learned this the hard way when they first put out their VDI product. Originally they did it with RDP which didn’t deliver the UX their clients needed so VMware eventually teamed up with Teradici. Someone at Amazon made a great call by teaming up with a company that has already been through this before and has made their protocol one of the defacto protocols for a remote Windows experience.
I have a very strong belief that DaaS is the future of the VDI/RDS market. I’ve devoted a good portion of my 2014 research schedule on DaaS based on that belief. Anyone in the VDI space knows that VDI cost to much, the licensing is broken, and its too complex, DaaS fixes everything but the license issues. This is why almost all DaaS models get around the license issue by not using a Windows Client OS (which requires VDA) and instead use a Windows Server instance (which requires RDS and has a SPLA). I don’t want to get into another #FixVDA debate but Amazon doing this at such a large scale should be a wake up call to get this license fixed. I don’t think Microsoft ever intended for Windows Server OS to be used in this way but hey it works and it saves money. Licensing aside DaaS solves the complexity (it has none because the DaaS provider is in charge of that) and it solves cost concerns because you know exactly what it cost to own it.
I am happy with this announcement from Amazon because it validates a lot of the points I’ve been making over the years including:
- Servers used on a one to one basis avoids the VDA cost
- DaaS removes complexity
- DaaS removes cost
- DaaS is the future
- I’m the coolest guy you’ll ever know
Okay maybe not that last one but the others are definitely getting validation when a huge cloud service provider like Amazon decides to jump into this market.
However there are some other points I’ve made about DaaS that Amazon will soon be running into:
- Desktops follow servers: If your servers (that host your mission critical applications) are running on-prem but your desktop is off-prem I’d argue that there is negative value of moving to off-prem DaaS. This adds four hops to the UX. User to remote desktop, remote desktop to application, application back to remote desktop, remote desktop back to user. All of that is probably traversing the Internet which has no sense of Quality of Service (QoS). This is a problem, but one that is fixed with the next bullet.
- DaaS will be a hybrid model: I picture the future of DaaS as a hybrid model whereby orgs manage the solution via a SaaS based management platform but the desktop you use will actually change its location based on the user context. (This is where Workspace Aggregators come in). I need to know who you are, where you are, and what device you are on, so I can provide you with the best UX based on that context. To support this hybrid model on-prem DaaS appliances will start showing up. I’m talking about a piece of hardware where all the org has to do is give it an IP address. The org would then go to their SaaS based management console and plug in the IP of that appliance. The DaaS provider (and DaaS software) will then replicate the desktop/data/etc to this appliance. In doing this users will benefit from a LAN UX when they are in a location with an appliance like this.
My theory on the hybrid solution solves many of the UX challenges that a full off-prem DaaS solution will face. The thing is even my theory is nothing new, its just reapplying the branch office problem that no VDI vendor has really solved yet. I’ve been looking at DaaS to solve that challenge, I think it will, and I think VDI complexity and cost may soon be a thing of the past.
I applaud Amazon’s entry to this market, and think they’ve already made some great decisions (Teradici) upon entering the market. However, I don’t expect it’ll take too much time before we start hearing about the need for better UX, be it on-prem DaaS or how just doing a good protocol like PCoIP isn’t enough, UX features that extend the protocol are just as important. Oh yes, Amazon has opened a can, and it should be a fun market to watch because of it.
Final Note/Question: How would you define the Amazon WorkSpaces service? I put out a one-question survey to see what you think.