You can get a decent generic barebone from Tigerdirect for less than $300 (have to watch for a deal) with a quad-core processor, 8GB of RAM and a TB hard drive. I have one with Xenserver free version because I like the tools and driver support. I have used VMWare 2 GSX and ESX, then ESX3, VMWare Server free version and ESXi, but have been using Xenserver free version in both test and production for the last three years, though I understand that VMWare's solutions are also very workable. A UPS is helpful as well.
My current test Xenserver has at time of this writing 4 VMs on it - two Linux boxes, a Windows server 2003 and an XP instance, all used for testing and development. I have a Windows 7 instance as well, but it happens to be turned off at the moment. I use an external USB storing snapshots of test VMs - get a clean config, store a snapshot of it, then you can test, muck it up, blow it away then start up a copy of the snapshot without having to re-install. Mine has been running continuously since early summer.
This setup can get you started with minimum cost and effort if you are doing development and functional testing that does not include anything too exotic like clustering or a database with a large transaction volume. You're not going to break any speed records but you can build VMs in all the OS types you want to test and limit the number currently running to 4 or 5, and you'll do just fine.
I'm not a big fan of Oracle's 20th century business model, which like a lot of other big name proprietary software companies and other types of companies as well is predicated on doing everything possible to obtain vendor lock-in, then charge through the nose for licensing and support, forcing upgrades, and basically squeezing customers at every opportunity. That's the downside of the model - in one way it sees customers as prey to be devoured.
The flipside of this is that proprietary companies like Oracle do make considerable investment to create solid, reliable product offerings, and they try to provide high quality support.
There are other proprietary companies out there who have Procrustean approaches; they don't spend time developing or innovating but rather continue to ride the gravy train of code that was written years and years ago. Customers have to alter their problems to fit the proprietary solutions. This is true in part of some of the niche applications aimed at specific vertical markets Oracle has acquired, but Oracle's acquisition has actually brought new life to languishing applications and brought Oracle's support processes to those same small app vendors.
Oracle targets customers who are willing to pay high prices for high quality software and willing to pay high prices for support. Is the cost justifiable? It depends - for some companies the risk exposure of getting 90% of the functionality of Oracle-type products for free or for very low cost is worthwhile, and the risk exposure of being without an enterprise-class support organization (or paying for support on a per-instance basis, sometimes through a consultant if no such support plan is offered for a given application) is justifiable. It's a decision each technology using company has to make for themselves.
Oracle's acquisition of the K Splice project is consistent with their business model.
Their business model is not amenable to me personally, but in some cases it might be a good fit for some of my customers. In those cases I can recommend Oracle's solutions, even though I am not fond of Oracle's business practices, which to some may seem avaricious, but to others may simply be a sign of an aggressively run profitable company that offers high end products and services and demands concomitant prices.
As to whether Oracle will contribute to the K Splice community or hold its own code contributions proprietary is their call. Past history indicates that they may not be enthusiastic contributors to the community but any prediction of how they will act in this case is pure conjecture. We'll have to wait and see.
I tend to be favorably inclined toward Open Source but I also grant that people whose business model is built around proprietary software licenses have every right to market their software under whichever license they choose. Open Source licensing preserves some very specific freedoms - if you are going to use or distribute Open Source software, you need to comply with its license and keep it open.
If you're going to use proprietary software you should comply with its licenses too, or you should find or code an alternative that has license provisions more to your liking. When I am searching for a software solution I look for Open Source solutions first, followed by proprietary ones if I can't find anything that is free and works. For some functions - and the narrower the vertical market niche that's being serviced is, the more likely this is true - there simply aren't Open Source solutions yet, and you have to go proprietary.
A proprietary niche vendor who sells crap (and there are more than a few) just needs some competition to make him learn how to code. If that competition is Open as well as superior, he may have to not only learn how to code better, but may have to re-examine his business model.
This is what distinguishes cyber-libertarianism from cyber-anarchism, the view that all licenses are bad and all rules are oppressive, which view I don't think is warranted or even possible, but that is for another discussion.
There's another trait that's missing here: Professionalism.
Real Veteran Unix Admins are true professionals who know how to function in a business environment without being arrogant or prickly. They are comfortable enough in their own skins and emotionally secure enough that they never need to engage in put-downs of others who don't rise to their level of technical acumen. They never play at being the high priests of the sanctum sanctorum of Unix administration which should not be desecrated by mere mortals. They are capable of doing their jobs without needing constant stroking by management or self-stroking by engaging in endless pissing contests and self-aggrandizement.
Unfortunately there is a small subset of Unix admins who are technically brilliant but emotionally insecure. They can do amazing things with systems but they are often difficult to integrate into a team or an organizational culture. They are high-maintenance employees, both blessing and curse to the companies they work for. With proper structure and coaching (and sometimes therapy) some of these can be developed to become true professionals. Some may be impervious to all efforts to help them mature and become more stable - they will have to be compartmentalized for the good of the organization or let go.
I will leave it to my gentle readers to decide which group the author of the article belongs in.
BASIC is the Computer Science equivalent of `Scientific Creationism'.