Friday, December 2, 2011

The Extension, Part 1


I got to the C-Extension Local Server Junction a little before two PM. They'd been expecting me for a few hours – I wasn't exactly Johnny on the Spot, but I was within my five-hour assignment window, so no problem. Bill the security guy was watching the first access chamber. I nodded to him as I walked by.

“Hey, Bill.”

“Hey, Bannon! Been a while since I've seen you around...”

“Yup. No news is good news, right?”

There was a subliminal flicker as the retinal scanner above the door checked my print. My audite said, Bannon Froyd. Expected on code 3727 maintenance call. Confirm.

“Yes,” I sub-vocalized.

Confirmation accepted. Proceed to access point. I slowed down a little, still talking to Bill.

“Everything good with Janie and Marge?” he asked jovially, apparently excited to have someone to talk to.

“Yeah, great. Gonna get this taken care of. I'll see you on the way out, Bill.”

The door to the router ring opened for me with a swish, and I crossed to the server chamber. Again, there was the flick of a retinal scan... this time, I didn't even slow down. The pressure-locked door popped, opened, and closed as soon as I cleared it. The warmed-over lights inside the server room were already on.

This was going to be a quick job, I figured. The secure gates in my district – doors, locks, exchangers, access points – had started lagging by three or four seconds, and it was pissing people off, and they were complaining to C-Junction KX. They were used to cruising right through doors, letting the scanners run ID's and transactions without even pausing. The three-second delay was totally throwing off their rhythm.

So earlier today I got the debug order... go to the central extension for the district, clear out the cache, make sure all the transistors are locked in right, and double-check the data buffers to see where there's a bottleneck. C-extension does most of that stuff itself – if it can't keep up, it generally means there's a big surge in data, plus some kind of shaky connector. That's the only thing the system can't do very well by itself... make the recursive hardware adjustments to keep up with self-maintenance.

That would probably take about an hour. After that, I would spend another hour filing reports: where was the data surge? Where was there extra noise in the data? Which transistor had come loose? Was everything else up to code? And then it would assign me a general clean-up... another half hour. Then I could go home. Pretty routine. I've had much worse assignments.

This story will probably continue, at least for a while. RSS for this story can be found here.

No comments:

Post a Comment