Deploying Amazon WorkSpaces at Scale

I’ve made no secret of my dissatisfaction with Amazon’s WorkSpaces DaaS platform. While I like the general direction in which the platform is heading, and appreciate the impact that Amazon can have in the DaaS market, WorkSpaces has been slow to implement enterprise-class management features and suffers from too many rough edges to withstand close scrutiny when compared to many alternative solutions.

Nevertheless, it has gained some big-name support; at the recent AWS Summit, Johnson & Johnson’s Director of End User Computing Jeff Mendelsohn took to the stage alongside Nathan Thomas, General Manager Amazon WorkSpaces, to share Johnson & Johnson’s experience implementing Amazon WorkSpaces to support its large contractor community.

The above vendor marketing pitch is confined to the first twenty minutes and is worth review if you haven’t had the opportunity to take a look at Amazon WorkSpaces yet. It’s refreshingly light, if not totally free of hype.

“When we launched Amazon WorkSpaces, the most surprising thing for others was how many of our customers told us the experience of using the Amazon WorkSpaces was much better than the experience of using their on-premises VDI solution, even when they spent a large amount of time trying to optimize those solutions for user experience.”

Seriously, any organization that reports this is seeing the benefit not because DaaS is good, but because its VDI implementation is bad, and no amount of tuning will fix a badly designed VDI environment.

Jeff Mendelsohn’s explanation of Johnson & Johnson’s Amazon WorkSpaces implementation is a textbook example of why an enterprise should consider DaaS. It is offering the service to support BYOD for 16,000 contractors and opt-in employees, and is replacing 8,000 end of life desktops with Zero clients. Johnson & Johnson is also looking to DaaS to cover mergers and acquisitions (M&A) and divestitures. Using DaaS for M&A is an excellent way to get large numbers of new employees on board, running a standard desktop image with no more than a scripted process to build the required number of desktops in the AWS cloud—provided, of course, that your acquisition target has the WAN bandwidth to accommodate the new workload. The benefits of DaaS in a divestiture situation might be as great, except when the target has already been migrated to DaaS. It’s interesting to hear that consideration is being given to the divestiture side of the equation.

Why WorkSpaces? Mendelsohn offers the usual reasons for moving to DaaS: no upfront CapEx, capacity on demand, focus on the service and not the infrastructure, global scalability, and rate of innovation. Plus a few more for the future, including elimination of enterprise WAN connectivity to small branch offices and moving them to a pure Internet connection to the AWS cloud. Mendelsohn goes on to detail all the work his team did to implement WorkSpaces at Johnson & Johnson, building out its own operations console and workflows to provision workspaces which it integrated with its ServiceNow ISTM platform. It also developed a self-service application catalogue and migration toolkit to simplify onboarding.

Not to take anything away from what Johnson and Johnson has done: the level of effort needed to achieve the benefits that it has achieved serves in large part just to confirm my concerns about Amazon WorkSpaces as it stands today. Few organizations have either the skills or the motivation to take on the work needed to take a bare-bones platform like Amazon WorkSpaces and make it a viable desktop service.

Posted in End User Computing, SDDC & Hybrid CloudTagged , , ,