Microsoft's Windows 10 servicing calendar: a showcase for contradictions

Windows 10 may be half a decade old, but some things never change. Or if they do, that change comes slowly, in fits, starts, stops and bursts.

Witness the recent "Transform Windows feature updates with a servicing calendar" — a piece by James Bell, a senior product marketing manager in the Microsoft 365 deployment group, posted June 18 on the company's Tech Community website.

Bell used the space to introduce what he called a "Windows servicing calendar" designed to, as he put it, "shift your Windows 10 servicing cadence from a project-based effort to a more fluid process that aligns across the release cycles of Windows, Office and endpoint management tools, such as Configuration Manager."

Essentially, the calendar concept is simply a graphical way to illustrate when different parts of the Windows 10 upgrade process take place, how each successive refresh syncs — or doesn't — with those it follows and precedes, and how disparate parts of Microsoft's product line, notably Windows and Office/Microsoft 365, are also scheduled.

Microsoft

The graphical approach is not new. Computerworld has been using it for years to clarify Microsoft's oft-bewildering and ever-changing upgrade scheduling. And Computerworld cribbed it from Gartner Research, whose analysts Stephen Kleynhans and Michael Silver applied it to reports early in Windows 10's history.

It isn't the calendars and their occupying arrows designating support cycles, per se, that drew our eyes. Instead, it's what Bell exposed as he touted the scheme, particularly where he seems to contradict current Microsoft practice, if not strictly policy.

Because Microsoft often obfuscates, reading between lines is a mandatory skill for customers who want to know what's really going on — or at the very least, have a better shot at understanding where the Redmond, Wash. company is headed.

We plucked two elements from Bell's piece that we think are particularly incompatible with Microsoft's public stance to parse.

That horse is long gone

Even after customer resistance to the concept of a greatly accelerated development-and-release tempo, Microsoft still urged users to consider a deploy-every-upgrade strategy. If that was impossible or undesirable for every endpoint in an organization, at minimum, the faster cadence should apply to some of the systems, Microsoft recommended.

"While we encourage organizations to strive towards deploying every release to at least some portion of their estate, we also recognize that organizations with very high device counts, and the need for no/low disruption environments will choose to update less frequently," wrote Bell (emphasis added).

Claiming that a "rapid cadence is within reach for enterprises of any size," Bell also implied that a slower annual rhythm — where a feature upgrade is deployed just once each year — was the starting point for enterprises and thus might even be only temporary. "They are starting their journey with the Windows 10 servicing process," Bell said when listing common characteristics of companies doing annual upgrades. "For those unfamiliar with new processes that support Windows 10 servicing, moving from a once every 3-5 year project to a twice-per-year feature update process can be daunting."

In contrast, Bell touted an every-six-month pace by minimizing the effort involved and portraying it as the better goal for customers. "Once enterprises are familiar with deploying feature updates on an annual cadence, shifting to a rapid cadence is often possible with minor increases in effort, as plan and prepare motions are well established," he said. "Enterprises that benefit from the rapid servicing process ... continuously update supporting infrastructure to unlock new working scenarios."

But this continued encouragement to deploy multiple upgrades each year was irreconcilable with Microsoft's long-running efforts to reduce the number of refreshes. In the five years since Windows 10's launch and the introduction of its radical servicing model, Microsoft has trimmed the number of updates from three to two per year (and last year and this, arguably to just one), and extended support from 12 months to 18 months for all SKUs (stock-keeping units), then later yet from 18 to 30 for Enterprise and Education.

The effect of each of those moves individually and the whole collectively, was to prompt users to reduce their participation in the rapid-release model. Fewer feature upgrades meant fewer of them forced upon Windows 10 Home and unmanaged Windows 10 Pro. (Until a year ago, Microsoft decided when a device running Home or Pro downloaded and installed each feature upgrade.) Extending support lifecycles allowed customers, particularly commercial customers running Enterprise or Education to avoid more of the upgrades without risking running systems lacking released patches.

It's been clear to Computerworld that Microsoft has, as it's claimed, changed its release and support practices because of customer feedback. (What's unclear is how demanding or widespread that feedback has been; it's fair to assume that it took much to move Microsoft from a cornerstone of the Windows 10 philosophy.) What's odd, however, is that it continues to argue for a fast, six-month cadence when its own actions have attested to the benefits of a slower tempo.

Take Microsoft's 2019 announcement that henceforth, all users, including those with Windows Home and unmanaged Windows Pro, control upgrade timing through the Download and install now (DaIN) option. But by reserving for itself the right to forcibly upgrade a device as the current edition neared support retirement — a reasonable demand in light of many users' lethargy in updating — Microsoft also established annual refreshes for those machines (see this Computerworld piece).

Whether Microsoft anticipated that DaIN would result in a majority of users moving from twice-a-year to annual was immaterial: it was the result. That Microsoft continues to tout the faster cadence is the preverbal shutting of the barn door after the horse is out.

From major-major to major-minor

It was also apparent from Bell's explanations that the original intent of Windows 10 servicing — to deliver two more-or-less equally-equipped feature upgrades each year — has not yet died, contrary to the pattern the company set in 2019 and intends to repeat in 2020.

A quick summary is necessary.

During 2017 and 2018, Microsoft delivered two upgrades annually, each fleshed out with new features and functionality. (Microsoft set that cadence as official policy in April 2017, after distributing just one upgrade in each of the two preceding years.) Call that a major-major tempo, with the refreshes roughly equal.

However, Microsoft dispensed with that practice in 2019. The firm released a feature-rich upgrade in the spring, followed by a Service Pack-like update in the fall that was little more than a bugs-now-fixed retread of its immediate predecessor. Last week, Microsoft confirmed what Computerworld had long forecast, that the company would play the same major-minor beat in 2020.

The reasons differed each year, and for current purposes, those hardly matter. What does is that Microsoft, through Bell's advice, continues to assume that major-major is the true Windows 10 practice.

Bell showed that was the case by how he outlined customers' progression through H2, the update released in a year's second half, in the calendar he dubbed "Rapid Cadence." (He put that label on the grab-all-updates option.) There were no differences in the recommended handling of H1 and H2, even though the latter last year and this were but a shadow of their precursors. Under Bell's scheme, each refresh would receive the same four-month process of Plan, Prepare and Deploy.

In a major-minor cadence, customers should not have to go through as rigorous a process for a year's minor update as they did for the preceding major release, simply because so little is changed. Deployment should also be a shorter stage, again because there should be nothing new, or not enough to make IT admins re-run a lengthy roll-out with multiple pools of users.

It's unclear if Bell's calendar was portraying a major-major tempo because that's what Microsoft plans to do in, say, 2021, or because he thought differences in H1 and H2 were unnecessary or potentially confusing to readers. Does Microsoft — after two consecutive years of major-minor — believe it can go back to the more aggressive, less-customer-friendly major-major?

Who knows?

Computerworld has praised the major-minor practice, predicted it would repeat in 2020 and has urged Microsoft to continue that cadence as the next best move to abolishing the year's second update entirely. Those stances haven't changed. Microsoft would be well served to install major-minor as the permanent tempo for Windows 10. It would allow for extended testing (not that that results in a bug-free release), allow enterprises and large organizations to reduce their upgrade frequency to once every two years and put Windows on the same reasonable-for-a-reason pace as every other major OS, from macOS to Android.

Here's hoping that Bell's calendar example was short of detail for some reason other than that it's a prediction of 2021.