Data Center Advisors

BLOGENTRY00013-13APR2010-V1-0002.DAT

There’s been a great internal email discussion here in HDS after someone shared an article about the demise of tape. The article was from “Tape: A Collapsing Star” and the ensuing discussions have carried on for about two weeks now. The initial discussion was about tape not being as bad as portrayed in the article initially, to tape horror stories from the old days and then tape farms, soda machine sized tape drives, vacuum columns and tape stretching. This internal list is primarily mainframe experts and specialists that in some cases have been here at HDS more than twice as long as I have. Then the discussion went on to punch cards and floor sort exercises, paper tape, disk packs and washing machine type disk drives, and now singing CPUs and light shows.

I used to work on mainframe computer equipment back in my early days. While I don’t think I’m as experienced as many of my esteemed colleagues, I did work on many of the same type of processors and peripherals (does the IT industry use the term peripherals anymore?) back in my DEC days. I ended up in the Large Computer Group working on the DEC mainframes. Back then, a mainframe was defined as any computer with a word size greater than 32 bits and the DECSystem-10s and DECSystem-20s were 36 bit computers, as well as, the original VAX and VAX Clusters systems. One of my favorite projects was building one of the largest VAX Clusters ever assembled with about 50 GB of disk storage, approximately 1 GB per half rack (1x RA60 and 2x RA81s with the top available for the top loaded removable disk pack), that ran a specially developed version of System V Unix.

The vast collection of peripherals connected to these systems, were as exotic, as they were large and heavy. Back then “core” meant something very different than it does today. I use to stand inside a single 9 track reel-to-reel tape drive to adjust or fix it. Computer room floors were actually a couple of feet high to accommodate the enormous bus cables linking everything together and stacked on top of each other.

You get to really understand how a disk drive works internally when you have to replace all of the read/write heads in a head tree after a catastrophic “head-surface interference incident”. Then align the servo head to the servo track, align the remaining 19 heads to a single cylinder on an alignment pack, adjust the controller circuits ensure the servo head is optimally reading the servo track and pray that the customer’s data on their other disk packs actually line up. That’s after I would bring the disk drive back into the computer room because the head assemblies and crashed disk packs were metal shavings spread throughout the subsystem and it had to be blown out and vacuumed up before any work could be done to it. All this while wearing a suit and tie.

Today, a disk drive is replaced completely. I’ve watch the technology evolve in this direction directly from removable disk packs, to non-removable Winchester Head/Disk Assemblies, to today’s Hard Disk Drives. There are even discussions and projects of storage bricks where a group of disks are “fused” together and RAIDed as a single component. Once enough disks have failed, the brick is replaced with a fresh brick. Today there are enough layers of abstraction in both hardware and software, that the internal understanding of how a network, disk, tape or CPU works is not a requirement to be productive in the IT industry. In fact, being able to solve today’s business problems and otherwise, at a high level has allowed the consumer population to enjoy this technology as without having to understand exactly what they are using or how it works, so long as it works. This goes for both hardware and software. You can almost state the problem you need to solve or describe an application that you need to have, and in very large strokes of generously applied technology and boxed solutions to actually implement something. A lot of today’s artful skill is in figuring out how to make a solution leaner, meaner and more cost effective, and possibly easier to use.

Sometimes, knowing too much can hinder the creativity of “problem solving” at a big picture perspective. However, deep inside the catacombs of many large and distinguished technology companies, experts have seen the birth of many of these devices, others joined during technology’s infant years, and some joined at the adolescent and teen era. And while many things today seem shiny and new and awe inspiring, mostly these devices are just new approaches to the old ways of doing things. I still say on occasions that many of today’s technologies were thought up decades ago, but were far ahead of the state-of-the-art for its time. So when something actually doesn’t work the way the big picture was envisioned, it usually takes someone who has been around a longtime to understand the root of a problem in a very technical way, even though the names and terms have changed, nothing really has changed all that much.

For some additional strolls down techno-memory lane, check out my old friend Claus’ blog from a couple of weeks ago, Dinosaurs and Comfort Food.