Writings on various topics (mostly technical) from Oliver Hookins and Angela Collins. We have lived in Berlin since 2009, have two kids, and have far too little time to really justify having a blog.
I'm sure I'm not alone in my personification of hardware, but in extension to that I like to know that my hardware is doing something. The thought of it sitting there idle just bugs me. So when I install a fresh new Jenkins server and it is sitting there waiting for jobs to be fired off, it saddens me just a little that it isn't utilised more.
The flip-side situation is when the machine is overutilised, or even just adequately utilised. Just as it is frustrating in one way to have a job that completes before you even have time to make a coffee, it is frustrating to have to wait for several jobs to complete that take an hour. In reality, the sweet-spot in the middle where you have the best of both world (latency and throughput) is so hard to achieve you end up having to choose one or the other.
I guess this is what cloud computing is for, which is yet another area I feel like I'm two years behind in. My understanding is that there are no real guarantees of throughput or latency. You have the promise of "infinite" scalability, but no real idea of how things will perform on any given VM. This makes knowing exactly how much performance you will get at any given point in time completely non-deterministic (obviously I am speaking about the public cloud here). What does this matter to an OCD sysadmin who likes his hardware to be well utilised? Probably nothing.
Going to the cloud does make sense, but who in this industry would never miss the endless rows of blinkenlights in the datacenter on the occasional visit? I doubt there would be a single person among us (and if you answered "yes", what is wrong with you? ;))