Virtualization, a word that started out as one of those industry buzz terms many years ago has managed to find its place in every single organisation I’ve had the pleasure of working for. It’s really quite a nifty technology, being able to run multiple virtual servers on a single physical box which allows you to better utilize your computing resources (and therefore dollars to). I was lucky enough to catch the virtualization bandwagon early on and managed to establish myself as a specialist in the field before most people had a clue how to utilize it properly. Primarily most of my experience was with VMware, since they were the only company doing it right for a very long time. Since then I’ve had a chance to sit down with Microsoft’s answer to ESX in the form of Hyper-V and the differences are quite significant. So in this battle royal-esque show down I’ll pit them against each other and we shall see who comes out the victor.
A great judge of how well a company can do something is how long they’ve been in business. VMware started out in 1998 and delivered their first virtualization product, VMware workstation, a year later in 1999. Microsoft delivered the first version of Hyper-V back in 2008, but we could safely say they were working on it since at least 2007. That gives them 11 and 2 years respectively in the virtualization game, so the advantage of time in the market is going to be handed to VMware.
VMware: 1, Microsoft: 0
One of the biggest considerations that most organisations have when looking to implement virtualization is the cost, as they’re usually trying to save money by implementing it. For the most part they will as long as the design is done properly and they’re not virtualizing for virtualization’s sake. However this is one of the glaring points that plagues VMware, it’s almost ludicrous prices. A single core standard edition license will run you about US$795 and the creme de’la creme costs a whopping US$3,995. That’s not to mention the cost of the central management server either, which will cost you another $4,995 plus the $1,095 required for one year of support. Hyper-V on the other hand is an additional US$28 on top of your existing license fee for Windows Server 2008, which you’re probably going to be buying anyway. The central management server costs about US$1150 which is a darn sight cheaper than VMware. The Hyper-V licensing model also has the advantage of usually being wrapped up with your licensing agreement with Microsoft, flying under the budget radar.
VMware: 1, Microsoft: 1
Now I’m going to get into the nitty gritty of these virtualization solutions: the architecture. Thanks to their marketing departments both of them give you a high level view of their respective virtualization systems.
(Click to enlarge)
Under the hood they’re very similar in the way they’ve designed their solutions and really there’s not much they could have done differently. The biggest difference that’s not clearly highlighted in these marketing speals is that the underlying technology for VMware is basically a modified Red Hat Enterprise Linux kernel that runs directly on the hardware. Hyper-V on the other hand requires that Windows Server 2008 be installed first and then the virtualization run on top of that. While the differences between these two are small, architecturally speaking, the main difference comes from the fact that for Hyper-V you’re running a full blown copy of the operating system in order to get the virtualization benefits. VMware avoids this by using a very slimmed down version of the Linux kernel (the footprint is about 32MB) meaning that more resources are available for the guests. Microsoft has tried to counter this somewhat by introducing a version of Windows Server 2008 called Server Core which removes quite a bit of the cruft, but the large memory footprint and installbase remains.
VMware: 2, Microsoft: 1
The next set of comparisons get a bit fuzzy, since they rely heavily on which version of ESX you end up buying from VMware. For instance if you buy the standard or advanced versions of ESX you’re basically buying Hyper-V at an inflated price, since the feature sets are basically the same. Up until the recent release of R2 for Hyper-V the advanced licensing for ESX would still bring out on top feature wise, still at a premium however. Since VMware has been in the virtualization game for so long they’re no longer focused on what most people would consider core functionality so the more advanced licenses are where they really begin to shine. Things like dynamic resource management and live storage migrations are just simply not available for Hyper-V and have been deal breakers for me in a couple situations. Additionally ESX is very agnostic when it comes to what you install on it, supporting basically every operating system and their respective versions out there. On a pure feature level VMware wins out, but not without incurring some hefty costs.
VMware: 3, Microsoft: 1
One consideration that most people leave out is how easily a virtualization solution will integrate into their environment. VMware is a completely different way of doing things and whilst vSphere is leaps and bounds easier to use than its predecessors it’s still a completely different world to your usual world of stumbling through menu options to fix a problem. Your decent ESX administrator will have to know his way around a Linux command line in order to get some of the more advanced things done (just try to install EMC SAN agents without it, I DARE YOU! :P) and there’s a fair share of problems that will require some obscure commands that can do a heck of a lot of damage if you don’t know what you’re doing. Granted there is quite a good community for all of VMware’s products so you’re not usually left in the dark with no one to hold your hand, but its still no substitute for an ESX administrator who’s dealt with his share of broken ESX environments before. Hyper-V on the other hand has that distinct Microsoft flavour to it and any Windows administrator will be able to administer and troubleshoot an installation without too much hassle. This also has the benefit of not attracting a premium when it comes to hiring in new talent, as us ESX know-it-alls love to charge a bit extra.
VMware: 3, Microsoft: 2
Even though the final score here would lead you to believe that VMware has come out victorious I’m still hesitant to give it the crown of virtualization king (and yes I know I’ve left out Xen but that’s with good reason, I’ve never used the damn thing). The cost of VMware is so massive that I find it hard to convince any shop who’s looking to virtualize to go with it. Sure the name carries a heck of a lot of weight and really if you’re paying the prices on the website you need to talk to a different VMware rep but the point remains that even a single site installation with the basic package will easily run up US$10,000 on top of your current Windows licensing costs. For anyone with a lot of disperate sites that require a virtualization solution you’re much better off with Hyper-V, although ESX really shines in large (100+ servers) data centers. I still recommend VMware to most people since they really are the best, despite their high costs. It would seem as always that you need to carefully analyze your situation and build a solution based on your requirements and not the marketing hyperbole.
Although if you use VMware you’re more likely to get a charming person like myself knocking at your door, and who doesn’t want that? 😉
I see a lot of people these days claim that they’ll never buy a house because it’s a bad idea, usually using the past year of economic turmoil as an excuse for their actions. Similarly with Australia’s housing market in an apparent bubble (that’s been meaning to burst for years if you believe the doom and gloomers) it would seem that it’s impossible for the first home owner to get into the market without ruining themselves financially. As someone who bought his first house 2 years ago when interest rates were rocketing up to 10 year highs and bought the second just this year I know exactly how the market is functioning and what the causes of this so called “affordability” crisis are. Today I want to step through some of the insights I’ve gained into market and why the media reports are, as always, misleading.
Hitting up the ABS for some data on house prices I found that they haven’t updated the price data in almost a year, leaving me with real figures that don’t accurately reflect the current situation. Still I’d hazard a guess that the swing couldn’t be more than 5% either way, otherwise we’d be hearing about it in the news. Taking that all into consideration here’s the median house price for the eight capital cities of Australia (in 1000’s):
You can see there’s a wild amount of swing between the different cities and that’s with good reason. Canberra for instance is filled with public servants who all have steady, above average incomes and this is reflected in the housing prices. Places like Hobart and Adelaide owe their lower prices to the lower population and hence lower average incomes. So what do these figures tell us? Well typically one city or state will be reported on and you can see how that would lead you to believe that all of Australia is unaffordable, especially when you look at say Sydney. Secondly the median figure tells us that 50% of the houses sold in that quarter were below that price, and the other 50% above. So whilst the median price might look rather scary it is in fact far from it and this is why reporting like this which is using the average price can make things look far more scary than they really are. It only takes a few outliers to completely ruin the average price as an indicator of affordability.
I’ve often talked with friends about the house prices in Australia being a crisis of desire and not affordability. Todays generation Y grew up predominately in suburbs with an easy going lifestyle and are seeking the same thing for themselves when they move out. The problem is those closer in suburbs now attract a premium as the urban sprawl has made them far more central than they were when the gen Y’s parents bought into the market. Taking this into consideration the first home buyer should not be looking at median priced homes and should be looking to the mortgage belts if they want to step into the market. It might not be what they want and so they choose to rent seeking to keep their lifestyle. There’s nothing wrong with this, the issue I have is that most of them throw their arms up in the air and call houses unaffordable when really they’re being unreasonable with their demands.
There are also some more financially inclined among us who believe that renting is more financially sound than owning. Now this is a complicated idea and the answer can vary greatly depending on someone’s situation. However there’s a little bit of comparison we can do to get a feel for how right or wrong this idea might be.
Having a quick look through Allhomes I found I could acquire a typical 3 bedroom house for anywhere from $350,000 to $400,000+, but there were quite a few homes available for somewhere in the middle. Checking the rental market for the same area (I’m using Holt as an example) a 3 bedroom place could be expected to rent for $380 per week. Now the one assumption I’m going to use is that property price as a general rule of thumb increases by about 7% annually, or it doubles every 10 years (see here for some data on the matter). So in the buying vs renting argument over 10 years you’d have:
Looking at that it comes out to a difference of around $100,000 over 10 years to own rather than rent which works out at about $200/week. The question is if you are renting, could you use that $100,000 to invest and beat the return gained on the house. If the theory holds true the house you bought is now worth around $750,000 or a capital gain of $375,000. If we knock off the $100,000 of extra money you had to spend since you bought there’s a difference of around $275,000 so the question then becomes, could you do this if you didn’t buy? For most people the answer is no, since Australians aren’t particularly good at saving money. Renting makes fiscal sense if you use the difference in what you spend on a house and what you’d spend on renting to make a return of approximately 7% per annum and don’t spend it. Suddenly the extra $200 per week might not seem so unaffordable.
When you take certain figures being reported by the media it’s always easy to get caught up in the hype. The problem with using a single figure to report on these things is that the real situation is hidden under a layer of equations and interpretations. If you go out there, do your research and curtail your desire you will see that the Australian housing market is no where near as bad as the media makes it out to be. I bought a house at a time of sky rocketing interest rates and everyone screaming that the housing market was about to go bust. Here I am 2 years later and the bubble has yet to burst and they’re all still saying the same thing.
Don’t believe everything you read kids.