Amazon tries again with AppStream because customers didn't like it
Tired of desktop apps that respond instantly? Banish them to the browser
Acknowledging that the initial version of application streaming service AppStream failed to appeal to customers, Amazon Web Services is ready to try again.
Introduced in 2013, AppStream offered companies a way to stream Windows desktop apps to users via browser across a variety of devices. It was intended as an alternative to the potentially costly processes of managing apps on premises.
"We thought game developers and graphics ISVs would embrace this development model, but it turns out it was more work than we anticipated, and required significant engineering investment to get started," said Gene Farrell, VP of AWS enterprise applications and EC2 Windows, in a blog post. "Those who did try it, found that the feature set did not meet their needs."
Exhibiting a flair for self-flagellation that's unusual in an industry convinced of its infallibility, Farrell drags out the mea culpa. "With AppStream, we set out to solve a significant customer problem, but failed to get the solution right," he said.
This time it will be different, or so Farrell would have us believe. Version 1.0 required customers to set up their own streaming service using an SDK. Version 2.0 is fully managed and doesn't require server engineering. It provides more instance types, streams applications without plugins, and works with Amazon Virtual Private Cloud (VPC). It also supports identity federation.
Amazon claims that AppStream 2.0 is beneficial because it provides rapid app access, cross-device compatibility, freedom from device hardware limitations, centralized updating, and the security of running apps in Amazon's cloud rather than on a local device.
App Stream 2.0 supports any application that can run on Microsoft Windows Server 2012 R2 OS. Apps are accessed through an HTML5-capable browser, for better or worse.
Amazon suggests AppStream 2.0 can be used for CAD, 3D graphics, simulation, gaming, media editing, medical imaging, and life sciences applications.
Yet in its 1.0 FAQs, the company warns that AppStream may not be suitable for some online games. It recommends against running first person shooters and other games that require low latency.
Amazon also notes that AppStream requires a network connection, making its use on mobile devices – often used in situations with slow, intermittent, or absent network signals – questionable. Google took a similar stance when it first launched Chromebooks, and only later acknowledged that offline functionality is worth having.
Though AppStream may find fans among companies that want to offer legacy desktop applications without the chore of managing them on local IT infrastructure, the service comes at a cost. Starting at a minimum of $0.10/hour for an instance with two CPUs and 4GiB of memory for each user, provisioning resources for eight hours of use per day over 216 or so business days in a year comes to $172.80 per user, plus a $4.19/month fee for each to cover the Microsoft RDS software license (for organizations that don't already have an RDS license). That's more than four times the annual per user cost of Google's G Suite.
For a company with 100 people, the price would be $22,308 annually. And some companies will want 24 hours of app availability every day, so workers aren't limited in the hours when they can access the software. That would cost $92,678 annually.
But if AppStream allows organizations to fire a few IT people, it could pay for itself. ®