TechNet Subscriptions, Part 2: I'm Not Crazy, and Neither Are You
Back in the day, I subscribed to the very first Microsoft TechNet offering, which consisted primarily of a CD-based offline Microsoft Knowledge Base. It was tremendously helpful, and only cost a couple hundred bucks. Until recently, you could pay not much more to have access to non-expiring evaluation software of Microsoft's major business products, with the Knowledge Base being better served from the Internet. Sadly, TechNet Subscription's days have come to an end, a move many see as evidence of Microsoft's increasing detachment from the "IT Pro" side of their audience. So what replaces TechNet?
First up is Microsoft's official answer: free-to-download product evaluations and "test drive" VHDs. Unfortunately, these expire, so while they could certainly be useful for evaluating software that you might purchase, they're not much good beyond that. Even as evaluation software, many have commented that more complex trials can simply take longer than the allotted time. Trying to do a comprehensive eval of the major System Center products can take months just to set up properly, before you even get into the testing and actual evaluation stage.
Beyond those free eval downloads, your options depend a bit on what TechNet was doing for you. Reading the comments in a recent article I wrote about TechNet , it's clear that my perspective comes from using TechNet in large enterprise-scale environments, but that a myriad of other uses exist.
Let's start with that perspective of mine. Unfortunately, I definitely worked with customers whose companies purchased a TechNet subscription, and then used it to set up permanent lab environments that were used by multiple IT team members. These environments weren't often used to evaluate Microsoft software, but instead were used to stand up a small-scale mirror of the production environment so the organization could test in-house software, deployment scenarios and so on. That use seems to be a pretty clear violation of at least the letter of the TechNet license, if not its spirit, and those organizations will have to pony up more money going forward to have that same lab environment.
MSDN is one option, with the cheapest "MSDN Operating Systems" package at $700 for the first year and $500/year thereafter. You don't get server products with that, though, only Windows itself; to get all the server products like Exchange and SharePoint, they'll be spending at least $6,200 for the first year and $2,600 thereafter. But the MSDN license is still for a single user -- it really isn't intended to be used in a shared lab environment. These companies could, I suppose, buy an MSDN subscription for each team member, but that's going to add up. Still, it won't add up as fast as paying for production licenses for all of those products, which is the other option. Some of the organizations I work with are looking at the idea of negotiating "lab licenses" as part of their overall licensing agreements with Microsoft, to get those server products at a lower, non-production cost.
Another perspective is that of the very small -- even one-person -- consultancy that needs to build out simulated customer environments to diagnose issues, test new techniques, and so on. That's a use explicitly allowed by MSDN, but at close to 10 times the price of TechNet, it's been a bitter pill for these small businesses to swallow.
CloudShare.com offers a possible alternative, although the financials aren't necessarily more compelling. For about $500, you can get an online "environment" that contains up to 8 GB of RAM and 10 vCPUs to be shared between a set of cloud-hosted virtual machines. Those VMs can be built from templates that include the OS and software licensing, and products like SharePoint are amongst the templates offered. VMs run for only an hour or two unattended, then they suspend automatically -- making them useful for a true lab, but not for production costs. Upgrade fees get you more RAM, more CPU, "always on" capability, and more -- and those add-ons add-up quickly. For a small consultancy looking to build one environment per customer, it might be possible to justify the price... but it's certainly a compromise situation. Larger environments can get incredibly expensive (think $6k/year and up), though -- and it certainly isn't a stretch to think that larger environments would be needed.
What about Azure? Unfortunately, Azure's pricing model is intended for production uses, and it's an expensive lab option. Its pricing also doesn't include server software aside from the OS itself, so you're really just shifting the hypervisor to the cloud, not creating an alternative to the software that was licensed as part of TechNet. I could certainly see Microsoft positioning some kind of "Azure Lab" product, providing VM templates that include software and perhaps doing something to restrict run-time or incoming connections to tie the lab to non-production uses, but there's currently no such offering. And, as many have pointed out, cloud-based labs aren't going to meet the need in every situation.
A third perspective is that of the hobbyist, student, and self-learner. The folks who build a "basement lab" in their home, and who were in most cases paying for TechNet out of their own pocket. These folks might be able to get by with an option like CloudShare; it depends a bit on what-all they were testing. Certainly, they can't build out complex networks, which in many cases was part of the point. CloudShare and options like it don't offer the full range of MS products, either -- you won't be doing System Center experiments, for example. Like the smaller-sized consultancy, these folks are sadly screwed, and it's a shame, because they're also some of Microsoft's biggest fans. It'd be nice to see some sort of "spark" program from Microsoft that provided software licenses to these people, at a low cost, for their own self-learning purposes... but that's pretty much what TechNet was, and the company's made its interest in that audience pretty clear.
A fourth use-case is the small business that maintains a small lab environment, but that isn't engaged with Microsoft at the Enterprise Agreement (EA) level. In other words, they don't have the negotiation option where they can try to wrangle some lab-use licenses from the company as part of a larger licensing scheme. MSDN pricing for even a small team of IT folks still seems cost-prohibitive; you could end up spending as much on MSDN licensing as you do on healthcare for those same employees. Now, I do believe that labs provide a business benefit in terms of increased stability, consistency, and reduced headache, and I do believe businesses should expect to pay money for those benefits. It's called Return on Investment, right? A 10-person team could have used TechNet for about $3,500 a year; MSDN would be more than seven times as much money. It does seem as if some middle ground should exist. Again, these folks seem a bit left out in the cold, or at least left in some kind of "gap" in Microsoft's strategy.
So what could the company do?
We've actually got some glimmerings of precedent, here. SQL Server, for example, allows you to run a free "hot spare" instance of the product in scenarios like log shipping. Provided the hot spare server doesn't do any production work, and is only used for failover, you don't have to buy it a separate SQL Server license. There's no enforcement of that licensing; it's something Microsoft would catch in an audit, and it, by and large, trust its customers to comply with licensing terms. So let's pull that thread a bit. What if, for example, your Software Assurance (SA) agreement allowed you to mirror the covered software in a non-production "mirror" environment? In other words, what if every SA-covered piece of software included the right to use that software once in production, and once in a production-mirror lab environment? For many organizations of varying sizes, that'd solve the problem instantly. It'd add real value to the sometimes-sketchy SA proposition, too, which would be good for Microsoft. It wouldn't require any extra infrastructure on Microsoft's part. It isn't a complete solution; evaluating new Microsoft software would still be dependent upon 180-day trials, but it would solve a big chunk of the problem for organizations using SA.
Of course, that isn't a one-size-fits-all solution. Your basement hobbyist, your smaller consultancies, and the significant number of businesses who don't use SA would still be left out. Microsoft might manage to get more businesses onto SA if lab licensing was an included benefit, but it still wouldn't be everyone -- not everyone can afford the SA surcharge of 25 percent of the product price per year, and smaller businesses would be impacted the most by that surcharge.
Frankly, Microsoft's not offering a lot of answers. Most folks fully agree that maintaining a lab environment of any size should cost something; nobody's asking for free software. Microsoft's official answers so far -- free evals -- don't support a permanent lab, which organizations clearly need. Microsoft seems to be saying they either have to pony up for MSDN on a per-IT-user basis ($$$), or outright buy full software licenses ($$$$), neither of which is exciting their customers or their community. Cloud-based options only meet a very small subset of lab needs, and it seems as if Microsoft is either unaware of, or uncaring of, the wide array of things TechNet Subscription was being used for.
Honestly, I could actually buy the idea that the company just didn't know how customers were using TechNet. As I wrote in the beginning of this article, my perspective is coming from large enterprise, and those organizations can better afford to build out labs and negotiating licensing to some degree. It's certainly possible that Microsoft didn't consider the small consultancy, small businesses, or even hobbyists and students. Of course, by this point, the hue and cry should have made it apparent that those scenarios do indeed exist, and Microsoft hasn't given any indication that it cares.
It's a frankly confounding situation. There's a lot of anger from Microsoft's community over the entire issue, in part because it just makes so little sense. The uses to which most people put their TechNet Subscription -- entirely legit uses -- benefit Microsoft in a number of ways. The company's answer -- time-bombed trials -- don't satisfy many of those uses. The company could have met those uses by providing things like connection-limited trials, as opposed to time-bombed trials; both would prevent production use, but connection-limited software will still support a permanent lab or learning environment. But... Microsoft has chosen not to go that route, at least so far.
Much of the anger around the TechNet Subscription is not, I feel, about the discontinuation. It's the fact that Microsoft seems to have turned a blind eye to such a large and diverse portion of its customer base, taking something away and offering nearly nothing in return. For folks who rely on TechNet, it's entirely analogous to Microsoft simply discontinuing, say, Exchange Server, and suggesting that everyone hop on Outlook.com instead. The anger comes from bafflement that the company would so completely disregard its own customers.
I'd like to say that it'll be interesting to see how this plays out -- but I suspect it won't have a satisfying conclusion. While I'll entirely admit that my own feelings about TechNet Subscription weren't strong, mainly due to the nature of the environments in which I work, I'll also admit that I'd not considered a huge number of use cases where TechNet made all kinds of sense. Microsoft's sin here is, it seems, in not making that same admission about itself. The first step in solving a problem is admitting you have one, and Microsoft doesn't, to this point, seem interested in taking that step.
Posted by Don Jones on 09/03/2013 at 11:42 AM