Over the past couple of years, a battle has been brewing. The battle is not about technology or devices. It is about data and usability. The contenders are Software as a Service (SaaS) and Virtual Desktop Infrastructure (VDI). There are proponents for both camps. And while they are not a direct replacement for each other, they do overlap in many ways. For many, the core value proposition for VDI comes from a specific application or data set. While VDI could be used for other activities, their value simply doesn’t reach the tipping point.
Challenges for VDI
While VDI has been around for many years, it has struggled to make significant inroads into today’s IT environment. In the interim, other solutions have filled the gap. The leading challenges to VDI are the value vs. the cost to implement and operate (as compared to the alternatives). In addition, as we move to a more mobile workforce, the underlying technology that connects clients and servers becomes more challenging for mobile devices.
The Impact of BYOD
Bring Your Own Device (BYOD) presents somewhat of a curveball for VDI. On one hand, BYOD increases the demand for VDI by moving applications and data off personally owned devices. Seems simple enough to implement VDI as a response to address the app/ data challenges with BYOD. However VDI brings a new set of challenges in a BYOD scenario that only adds to the already existing BYOD complexities. In the end, the offset is not as rosy as the surface would indicate. BYOD also creates demand that can be serviced by other means. Meaning, VDI is not the only option here.
Leapfrog to SaaS
Over the past few years, SaaS applications have matured significantly. That is in part due to the marketplace. It is also due to the maturity changes within the IT organization. There are three types of demand that could warrant bypassing VDI and going straight to SaaS. The first is off-the-shelf applications. Movement from an internally hosted software package to a SaaS offered version is one way to make the move. The second is greenfield applications. At this stage, IT organizations building new applications should be evaluating cloud-based architectures. The third are the legacy applications. Some of these will continue to require client-server based architectures. However, at the appropriate time, a change will be needed for legacy apps. It is important to understand where that tipping point is. If SaaS is an option, it presents an opportunity to leapfrog over the incremental improvements and should be considered.
Healthcare: A Case for SaaS vs. VDI
Specialized use cases demand further scrutiny over the requirements and potential solutions. Healthcare is a great example where HIPAA’s requirements around Protected Health Information (PHI) are a key consideration. Many healthcare providers are finding new ways to engage patients while working within compliance requirements. VDI seems like a good fit, right? Wrong. Yes, it could work. But the reality is that usability becomes a strong factor in consideration. Add in the movement to mobile users and SaaS (or mobile apps) start to look more appealing.
Bottom Line: As SaaS becomes commonplace for applications, the existing demand for VDI will further diminish. Unless there is a compelling use-case today, investing in VDI is probably a poor choice.
Discover more from AVOA
Subscribe to get the latest posts to your email.
I don’t think your argument is is VDI vs. SaaS but VDI vs. HTML 5 applications. I don’t see anything inherent in your argument for SaaS that native HTML 5 applications couldn’t fulfill. Organizations have real concerns over not just security issues related to SaaS but also flexibility of SaaS applications. If the SaaS solution doesn’t meet your interface needs what your options?
This is where we look to HTML 5 to deliver rich and consistent experience across light devices such as mobile and browser based endpoints while support BYOD.
Thanks for your comment. I agree with your point that HTML 5 applications are a good alternative to VDI. My point was between VDI and web-based approaches like SaaS and HTML 5. It makes a profound difference in the client experience and connectivity requirements.
Not to mince words, but some might argue the point that HTML 5 can be used for SaaS-based applications.
Makes since. I actually really agree with you. I wrote a similar post on how BYOD is a Failed strategy for the very reasons you mention. Organizations are trying to enable BYOD with VDI which is just a stop gate for Web based apps. If users wanted Windows on their mobile device they would have brought a Windows phone.
http://virtualizedgeek.com/2011/10/30/bring-your-own-device-is-a-fail/
Very true
If the user only needs access to a few APPs then maybe (SaaS vs VDI). But in cases where access to a significant number of APPs is required the upfront costs/time/effort assoc with redesigning and rearchitecting all of them before being able to switch over would be prohibitive. Also this assumes you even have control over the APPs, i.e. for internally developed maybe, but for COTs not at all. In these cases a VDI or simular capability would let you get off the ground very quickly. A VDI solution designed to support mobile use (i.e. designed for the approp form factor) would allow 100% of the existing APPs on the desktop to be available on the mobile device (tablet or netbook form factor anyway, phones would require a complete redesign of the presentation layer, etc…) immediately…. A VDI solution also would provide the ability to reduce the overhead assoc with the desktop device as well, i.e. use thin clients instead of full PCs…. The power savings alone do a long way to pay for the infrastructure needed… then all the other features/capabilities is gravy……
I agree in concept. However, when you delve into the details of the execution, it’s a bit more sticky.
In a large complex org like mine (the VA), every user has dozens of disparate APPs required to be used in the course of doing their jobs, and across the entire org there are several thousand APPs (yep, not a very good track record of ent architecture)…. There really is no practical way to rearchitect enough of them to port even a min number of required APPs to users endpoint devices via SaaS/HTML5, etc.. (many of new ones in development are being better architected, but it will take a decade or more to get completley off legacy APPs). Where as you can give me a check for $100K for a half-rack, 6 servers, 2 iSCSI SANs, 2 switches and a week after delivery I can have have 400 or so virtual desktops in production what would mirror the current desktop expereince and allow it to be accessed via a myrid of devices completely independant as to the end-point device hardware, OS, platform, etc. (only prime dependancy is that the device have a VMware or Citrix viewer APP available).
Interesting approach. My point wasn’t to be black and white about VDI. Just that there are options and alternatives.