http://open.citrix.com/blog/328-are-containers-the-right-answer-to-the-wrong-question.html
Russell Pavlicek make some very good points on the containers v. hypervisors race. Of course, he would defend hypervisors; they are his bread and butter after all.
I have to admit, though, that as I read his post, I felt that he was building a straw man as no one is saying that hypervisors will be replaced by containers. However… I went back to my own recent post(http://nexus.zteo.com/posts/the-next-evolution-of-virtualization/), to confirm my thinking, and re-reading my own words I realize that I totally omitted mentioning one of my favorites things about containers: the fact that they can cohabitate with VMs.
When I refer to a future where container-specific applications would exist, I should mention that this is tied not to hypervisors disappearing but instead to the commoditization of VMs. I should have made this explicit.
On the other hand, I still think his post is a straw man: what I hear from containers evangelists is not that hypervisors are bloated but that VMs are: there is a considerable nuance here as, while VMs are bound (IMO) to become a commodity, hypervisors are going to be an increasingly important tool in a network admin’s toolbelt. However, I believe that, if you wish to take the lead in that particular market, you should start looking into hybrid hypervisors that support these different layers of virtualization (For future reference: “Fractal hypervisors.” -- there, I said it)
Comments powered by Talkyard.