IT evolution seem to be a never-ending sequence of extrapolating upwards, enclosing complexity into simplicity and increasing managed complexity through encapsulation, focusing on the service. Really, we don't care about speeds and feeds regardless of what the salesdroid says, we want the service working as requested. Getting dialtone is interesting only when I can call someone.
Mark Burgess: In the 90’s again, it was "the host is the thing". The host was your best friend but today it’s services – which it’s been that for a while. I mean, ITIL has been pushing the idea of services for some time from a business point of view but IT is finally catching on, I believe.
And of course, with just the scale required to deliver these services, we now find that we need things like load-balancing and in order to scale-up means it’s no longer a host or a process running on our host but rather a little cluster of machines, that is representing a service.
So we have maybe the idea of infrastructure blocks or -- what would be the word, organs -- let’s call them organs. And, if the host is a cell and we’ll think of a special organ to deliver web or database or applet, or whatever, your favorite services and we want to -- we want to think in terms of those resources rather than with the low-level hosts, themselves.
At the same time we have virtualization has created a lot of possibilities in this space and made it very easy, much easier to do resource management than most. I believe a lot of companies today would typically, even if they are doing bare metal themselves, they would typically just run a hypervisor on the bare metal and then put a virtual machine on that, even if it’s only one. Just in order to be able to move that machine maybe, somewhere else with a minimum of disruption, in order to upgrade the hardware or something like that.
And I think, what the Cloud does for the host, and by cloud I mean, this approach to managing virtualization essentially. It’s what packets did for the network and what paging did for the disk. It allows you to swap and change resources in -- on the fly, in a very dynamic way and manage those things in a simple accountable unit which gives us a lot of flexibility to deliver things like services in mission-critical applications and focus on the important stuff.