RSS - Posts. RSS - Comments. That would be the real comparision. I would not call it so outdated, but yeah it need to be updated. I am sure that update will happen soon, but I can not promise on a date on the ITComparison Team behalf. Latest intelligence, Windows 8 should be on the market in 2 years. One of Microsofts subsidiary companies leaked the awaited release date over the weekend. Lets trust it follows the Windows 7 evolution procedure.
I do not need to have another Vista fiasco. Notify me of new posts by email. Notify me of follow-up comments by email. I have developed strategic designs for few of the most complex environments around where I got to execute on these designs or help the customer execute upon it. Home Management Training Webinars About. Comments Akuma says. June 20, at pm. June 23, at am. Hi Akuma, I would not call it so outdated, but yeah it need to be updated.
The performance of the hypevisor with all else equal is negligable. All hypervisors have optimizations, optimizations designed for specific reasons. You question was about performance, your reply was about features.
I cannot think of a single performance feature that each hypervisor vendor does not have an equivalent feature for. In this small area of performance it is all about the hardware and the configuraiton.
When you get to feature sets, management, and other aspects - that is all different and has nothing to do with performance. Then you need to be very careful that you actually know each platform and how to configure them if you truely want any type of equivalent comparison.
VMware is highly unlikely to release anything that paints them bad, Microsoft the same, Citrix the same, and so on. It is very revealing as to truely underlying performance behavior and scalability, taking out the emotion of the management interface and experience.
The most difficult thing about comparing software from different vendors is removing the bias and being truely objective. It is very hard. You have to install and use them all and work with them and understand them.
You must consider each piece that makes the whole; where it begins and where it ends, what requires an additional license and what does not. What instalaltion options exist and so on. But yes, all vendors want their products to look good.
And denying that fact when reading comparisons is only deluding yourself. This is my point. And again, your link is feature based, not performance based as was your inquiry. With VMware you need vCenter. Features are not apples to apples - you need to look and compare very carefully. I have not seen any comparison that has done this.
And considering that System Center is evolving, and the platform is evolving with Win8 I think you will see a big change. I don't anticipate any new numbers or comparisons until then. You might also want to watch the vendor space to see if any white papers come out. This is pretty constant. But you have to do your own comparison work. Hypervisor has to be very powerful because your resource provision happen on top of that and if have to design a large virtual infrastructure then such supporting certified document help's!
It is not that hypervisors need to be powerful, it is that they need to be efficient and give the most of the physical hardware resources to the virutal machines with little overhead. In the end, the performance of an individual VM comes down the the hardware, and or the design of the virutalization system the architecture of server, storage, network not necessarily the code of the hypervisor.
Features are a totally different type of animal. The features of the hypervisor itself are important in the context of this thread, however the features of the management layer above the hypervisor are not important in the context of this thread.
Hypervisor performance, as you keep stating is about the hypervisor. It is not about the features that vCenter adds, or that System Center adds. They add aspects of the larger system, but they are management, not processing power.
Does the management layer play a vital role in performance? In the perceived performance of certain features, yes. Such as how long it takes an environment to recover from a failure event. If the management layer is necessary for this. But again, this is perception of a feature and generally not considered to be a performance type of item. This is what I am trying to do, separate performance from features. The other issue is falsely comparing features at a performance level when they are not equivalent.
Take Live Migration for example. In the end, the feature is that a VM workload is moved and the VM stays running. What is the usefule way to measure this? VM down time. How long is the VM offline, and can the application you run in the VM tolerate that? Even if it is just a few pings lost, I have seen many applications that cannot tolerate any network disruption, other that take it just fine.
The end result is that your test defines the winner, but not based on performance, rather on implementation of a feature as it relates to your application. This is a better way to look at the comparison that you provided the link to. The very important statement is "with our workload".
0コメント