Login VSI is today announcing the availability of its newest product, Login AT for Application Compatibility Testing.
Back at Synergy, this product was the finalist for the New Technology category in the Best of Citrix Synergy 2018 Awards, and after getting a demo from Mark Plettenberg, I can see why—it’s useful for practical reasons, and it takes an innovative approach.
Login AT overview
The more frequent updates that are part of Windows 10 mean that IT has to do a lot more app compatibility testing than in the past.
Login VSI’s Login AT addresses this by using a script to launch apps and take a screenshot. Compatibility issues are likely to show up visually, in the form of error messages, apps not loading properly, or otherwise appearing different.
The idea is that you can compare screenshots from before and after OS updates (or any other changes—Mark suggested that customers could run tests daily). If the screenshot looks the same, then it gets a pass, and if not, it fails.
Administrators have to approve initial baseline screenshots, and then the software uses an image manipulation library to detect changes in subsequent screenshots. Administrators can also define a threshold for passing (say, if an app screenshot is 98% similar), or approve a new app screenshot as an additional passing baseline.
Of course, this is only an approximation for app compatibility, as some issues might not show up visually, or might only show up later on, but this is meant as a broad, mostly automated tool to help with the first pass.
Login AT is delivered as a self-contained virtual appliance (so there’s no need for a separate database or anything like that) and the screenshot comparison is done locally. It will be available standalone and through Login VSI Enterprise Edition XL. You can find admin console screenshots on the Login AT product page.
Mark said that the next release will focus on alerts, but for now, you can always integrate against APIs in Login AT.
The computer vision angle
Mark noted that Login AT is not using any artificial intelligence or machine learning (the library they’re using today is ImageMagick), but the concept behind this product fits right into a related trend: computer vision.
Computer vision is not a new concept, but it has been getting a lot of attention recently—the idea that software can pull just as much usable data out of a bunch of images as it pulls out of text and numbers is quite exciting. All the big public cloud providers are working on computer vision services, and many of these indeed use ML and AI.
We’ve written about how machine learning trends are affecting enterprise IT, and this is a prime example for those of us in the desktop virt and VDI space.
There are a lot of directions Login VSI could take Login AT. For example, one feature Mark mentioned would be to train it to read error messages.
The door is open for Login VSI to use more advanced computer vision services, and I’m looking forward to see where they go next.