The changing end-user environment

---

Friday morning I was in a meeting with several of our departmental system-administrators. During that meeting there was much discussion about the increasing prevalence of Apple laptops in Faculty and some staff offices. It was noted that the last year or so has seen a marked uptick in new Apple hardware.

This is definitely worthy of notice. The changing fiscal environment has prompted WWU to homogenize our central IT operations, while at the same time our end-user environment is undergoing the opposite process. Part of the reason for this is a decision from pretty high up to allow users to select their own operating-systems. Area staffers duly advise these people that we're not as good at OS X as we are with Windows (Linux-users tend to be outliers and generally don't consume support resources).

Homogeneity in the datacenter is something that you see quite often. It is an openly admitted to fact that large software companies (and sometimes hardware) structure their licensing costs to be cheaper on a per-unit basis if you go all-in. WWU was able to be heterogeneous for quite some time, using both Novell and Microsoft with a pinch of Sun, for many years. It wasn't until a trio of factors converged that pushed us to ditch Novell.

1. Changes in the IRS rules for the deductability of donated software greatly reduced the amount of Microsoft software we were being donated.
2. Changes on the desktop proved to be intractable to Novell (Vista), which greatly impacted the end-user experience.
3. Watching the housing market catastrophically deflate foretold Bad Things were going to happen to the State budget.
   

The first made using Microsoft software cost much more than it used to, which forced it to compete with Novell for funding. The second made Novell seem a poor overall choice. The third made the need to start cutting costs immediate. Together the decision to go all-in with Microsoft was pretty easy. The pinch of Sun was eliminated in the last year, a move that mostly impacted Professors who had been doing the same thing for 10-15 years and hadn't gotten around to changing.

However... out there in user-land, our Mac percentages have been creeping ever higher. This reintroduced heterogeneity into the environment.

The biggest impact of this is working through the compatibility issues. Take one specific example, Microsoft Failover Clusters with Server 2008. OS X 10.4 and 10.6 connect to them just fine, but 10.5 doesn't. 10.4 (iirc) can get a list of shares if users browse to \\servername, where 10.6 users don't get the list and just have to know what to enter when connecting. We discovered this shortly after the release of 10.6, which ended up forcing a lot of users to pay for the upgrade from 10.5.

It also means that The Latest Shiny in the Microsoft ecosystem may have to be put on hold until we have a way to integrate the non-MS platforms out there, which may take a very long time. VMs and dual-boot may be work-arounds until such time as a native solution can be found. All of this adds up to a perceived degradation in service-levels.

As one of our departmental admins said on Friday (paraphrased):

The worst part are Windows users moving over to Mac. Unlike our long-time Mac users who're used to how that platform works, these ex-Windows users expect things to work a certain way. They want their P: drive. They want their U: drive. They want all of their drives mapped as soon as they log in in the morning. They want their Outlook to work like it does on Windows. They want me to fix it, and I can't. Or if I can, it's kludgy and breaks a lot.

Even though I warned them that this would happen if they went Mac, it's now Western's fault for not working to their expectation. Their perception of our service goes down, and I can't do anything about it. I hate that.

Faculty are our biggest wild-cards here, though there are some staff areas where Macs are starting to show up. Because of the loosely centralized academic environment, the Just Say No policy adopted by so many corporations and governmental agencies is not a luxury we can enjoy. We have to deal with it, not block it. Right now, this means:

• We need to consider OS X compatibility to be critical-path for any user-facing service we provide.
• We need to vet new OS X versions for compatibility with our environment before encouraging users to migrate to it. Or alternately, put in a campaign to discourage users from migrating until the bugs get worked out.
• Somehow find a way to manage the expectations of users so that the more visible costs of the OS X platform don't seem like the University is nickle-and-dimeing them to death.

It's not just Apple Laptops, either. We're seeing a lot more tablets around, and today "Tablet" is synonymous with "iPad". In a year "Tablet" will include a lot of Android, and may see the first glimmerings of Windows. Tablets are popular with staffers, I'm already seeing a lot of them in meetings and in greater numbers than I ever saw netbooks. This entire new computing category will cause its own set of interoperability disruptions as people get data into and out of them and the clouds that serve them.

Because we're Higher Ed we never really got rid of our Apple users. We still had die hards in the dark years of MacOS 7, 8, and 9. Happily, Novell NetWare had very good AFP support so we were able to support these users. When it came time to go all-in with Microsoft, the Apple SMB support had advanced to the point where it was far easier to use that then shim AFP onto Windows. The perceived barriers to using it have dropped, there is now a peer-pressure effect out there in the market, and all those iOS devices running around flagging the Apple brand, have all combined to greatly increase the amount of non-Windows we have to put up with.

We're not alone in this.

The fact that Windows will be entering the Tablet marketplace well after both iOS and Android pretty much guarantees that Windows will never come close to the 90% market-share it enjoyed on the Desktop for so long. Because of our lack of draconian One Platform To Rule Them All policies, we'll be seeing this heterogeneity far sooner than our .com neighbors.

The transition to heterogeneous networking will be a challenge for the entire industry, even those of us who've done it before.