"How many programmers does it take to change a lightbulb?"
"None. That's a hardware problem" -- old joke
What a lovely thing is work. Putting your brain to somebody else's business and being paid to do so is the desire of every prisoner of the mind. And thus I have been reinvigorated from an unexpected corner. Operations. Hard drives. Routers. /opt. I have just recently discovered the joy of Linux server administration. I have to admit it's a lot more fun when you have three dozen very large servers to administer and everyone around you is quite comfortable that you have root. But this was the joyful gap that I didn't realize existed in all the years I endeavored to pull teeth in corporate IT and get the environment my brilliant developments deserved. I am three degrees of separation from end users, and for the first time in memory, I don't particularly mind it at all. I'm working for the machines.
What's doubly ironic about this is that I made a fairly loud exit from the world of Enterprise six years ago in abject frustration, looking desperately to find my place among open source and the clouds. Now, I want nothing more than to build my own datacenter in my garage. I want a DIY cloud, which is to say a big rack of hardware and some virtualization software. What would be lacking from a real cloud would be a real API, but that will be me. I'm the API of my own cloud and the captain of my soul. I think that making this move towards the machine is the fulfillment of yet another yearning on par with my move to stoicism. I had only thought I was dealing with reality in politics, but what I actually hungered for was the history of political philosophy. Similarly what I want in computing is access to the evolution of hardware and networks. Applications and clicks are as ephemeral as slogans and votes. Everybody does it without thought, what really counts is infrastructure.
Only I see where the whole cloud infrastructure game is going. It's going to oligopoly, with the fourth wildcard. Between the intellectual, legal and industrial capture of Google, Amazon and Microsoft is that nasty fourth thing called the Dark Web, and its next monstrous intervention will be the zero-trust tier of blockchain computing. If you don't know, then consider the history of Bitmain, and know that a man named McAfee is still alive. There is a future out there populated with hackers and systems that are much more immune from the sorts of ordinary disasters and bugs that destroy the lives of Enterprise IT jocks and their sheepy user base. There are two kinds of systems in the world, those who have survived hurricanes of DDOS attacks and those who are still wearing bunny slippers when it sprinkles outside. They don't know what a torrent is. They may not ever learn until it's too late. For those of us who will populate masts we would rather not be lashed to in the coming storms of cyberwar, the hedge is personal hardware and power, private networks and circles of trust. Or so it seems to me right now.
It turns out that grepping log files and shooting pistols are both loads of fun, and skills that require years of practice. But they're also fun. Shooting the zombies of tomorrows apocalypses will not fall merely to the burnt out crusaders longing like for the halcyon fields of home and bipolar yearning for honorable death. It will be fun.
So the first thing that I've done is resurrect a couple old machines from the scrap heap. My old MacBook Pro is now an Ubuntu machine. There's nothing in the Mac world that I need on that machine when I think about it, and bloat has really taken over. It's getting rather obvious and tiresome, Apple. Cut it out. That and an old 386 Dell will host lightweight stuff as yet to be determined. But right now they're running Consul and Zerotier (more on them later). Next I want to get a physical rack and a discarded Dell MD1000. That, I will fill with 7200 RPM terabyte hard drives and lower my cost of S3 storage, which I can and probably will just move all to Glacier. My S3 bill just peaked over $40 a month. No can do. Since I have gotten the new 256GB iPhone, there is essentially no reason for me to have that extra duplicate copy of my 100k pictures on S3. And I'm pretty sure that I have everything Flikr on my own drives too.
The big deal and central object of The Wall is my investment in Vertica and reference data. So I'll have a relatively high powered database server running so that I can practice spinning up entire orchestrated things around referenceable datasets. IE. click on this package and get a full 'enterprise' quality query space, stuff that almost nobody does. I expect to get under the community wire with open source tools, and I am especially looking forward to using Fugue as its applicable to instant-up a data reference stack. The collaboration between AWS and VMWare will help a lot, as will a lot of Hak5 reruns and r/homelab.
Big fun coming.