Nowadays, when people talks about servers, they would talk about virtual servers as server. I guess this is the age of cloud and I need to evolve my knowledge about such terms.
I had an interesting chat with my colleague once. He was a head of IT in a foreign country university. He was humble and I love how he greeted with warm feelings. We had an interesting chat for a while about each other infrastructures.
He asked me the total of server that we had in my workplace (Universitas Indonesia). Well, I told him that we once had 100+ servers and now we are reducing them to 70+ servers. He said it was too little and his place would have hundreds of server.
I thought, wow, his place must be a big server farm like a hangar. But, to my surprise, he surprised about my explanation and he told me that it was plainly wrong. I should mention the VMs not the physical servers.
Oh, well, what a nice talk and I stopped there because I have no idea how much the 'servers'. I personally haven't memorized all the data about VMs because my workplace had so much VMs. I wasn't prepared for that question. :-)
If it was like that, I think my workplace already had so much 'servers' that we are gradually moving toward 10Gig infrastructure. That's why I was interested in investing my time to learn about bufferbloat. Something that my colleague not knowing.
Yay! Viva Free/Open Source Software! I knew the term and could get to that before even it was published on IEEE. The one who introduced it was FOSS developer and he was on my RSS. *hipster kernel hacker*