Bah #3

I'm cleaning out hardcopy stuff, going over notes and project management nonsense from years ago, tossing it out because I don't need it anymore...the technology is either obsolete or I've memorized what is neccessary...why clutter up good desk space when I can stack trash? It inspired me to dig out this old bit of text and put it back into play, since the Great Blog Deletion took out my archives...this is an older posting, from back when I had a real job. Well, back when the real job was ending and and I lived the slow burning hell that was unemployment in the wake of the popped bubble and subsequent recession.



Get into "work" today. It's not really work. I'm here at ContractorCo central because there's nothing else for meto do but look busy. I've got an office right now, and a bajillion online references and tests I can take through the corporate intranet. Whee.

I've already gone through four of these this morning, just to see if I remember basic things like VLAN architecture, Win2K administration, and the like. I can remember why and how things are done, apparently, but the specific command sequences have been overwritten in my memory by something else. Not that it matters to me all that much, it's more important to me that I know how a thing works rather than how to work it. If I just know commands and menus, I don't really know anything but that...but if I know what is going on under the hood, figuring those things out is a matter of deduction. Makes me platform independant.

The problem with these study guides and tests is that they are so goddamn patronizing, more so than the Microsoft operating system. Every so often a little box will appear while I'm rapidly paging through information that demands my attention. It will say something like "What do YOU think is a good reason to use VLANs in a campus network? A, B, C, or D". Whaddya mean, what do I think? I angrily click at random, and it tells me in a textually cheerful way "That's right!" or admonishes me with "That is not correct. The correct answer is highlighted". The problem is that it is done in some kindergarten way, as if I was five years old printing out a capital A with a crayon inside my Big Chief Tablet. I have a sneaking suspicion that some "wellness initiative" compelled an overpaid contractor (much like myself) to design a question and answer interface that won't damage the fragile computer geek's self esteem. Please! I'm a systems administrator, I don't HAVE any self esteem! I don't need some overgrown Macromedia feebware to hold my hand and tell me it's okay to be wrong, I'd much rather it say "Dumbass, read the specs for 801.10 and come back when you have a fucking clue" and then bluescreen.

Another thing that annoys me even more about these tests and study guides are that they do not accurately reflect the real world. These guides operate in an "idealized" environment without actual chaos and stupidity. It's a fantasy world, where dogs and cats live together, ferrets are actually nice creatures to own and don't have a foul stench, and the Palestinians don't need a homeland because the issue was settled five thousand years ago equitably by all concerned. Such a world DOES NOT EXIST.

So what happens is that these things will pop up another psychologically soft question like "If a campus type flat network has congestion problems, what would you recommend to solve the problem" and gives me some possible answers. Knowing that this is the VLAN study guide, I know the proper answer for this idealized environment is "make VLANs" or whatever the closest thing is. But that isn't my first REAL WORLD answer.

In the real world, knowing nothing else about a network, I would assume that somewhere, something is fucked up. Outdated hardware ranks first, poorly configured and maintained network infrastructure second, and dumb users third. Two of these are rapidly fixable, and the alert reader can quickly tell you the one that needs a stick.

In the real world, baselining a network is vitally imporant. Some contractor types will come in, take a cursory look at the server farms and network closets and announce that it is all totally inadequate, and that a complete rebuild is neccessary. Sometimes, they are right. Most of the time, however, they are right the same reason a stopped clock is right twice a day. You gotta get inside that network, sniff it all out, learn the bugs and reasoning behind the layout and design. If this isn't done, you are wasting your money, even if the client decides that replacing everything is the best way to go. Because if you aren't careful, decisions made in the earlier implementation will carry over into the new one, and the network will still suck...it just sucks on faster hardware.

For instance, in the earlier VLAN scenario. Let's say that I'm contracted out to FeebCo for network nonsense. FeebCo has a huge network in one building, and they have VLANs. Under sane, idealized conditions, this is a reasonable thing to do. But then I ask a question that sounds to all concerned like a dumbass question BUT IT IS NOT: Why were VLANs implemented in the first place?

Sometimes, the correct answer is given by the systems administrator on site: "Because we needed to segment the network, we do layer 3 switching." Sometimes the answer is "I don't know, the guy who designed it was canned." I can live with both answers. But if a manager type, or other clueless person, pipes up with "Because it is an industry standard" or "Nyarlathotep Limited had them when I managed a group there, so I thought we should have them too" or "I read about them in Datamation, and forced the networking group to implement them while being scourged with whips".

Those latter answers are clear, telltale signs that the network may be horribly, deeply fucked. And probably everything else attached to it. Why? Because you are supposed to make a network design based on the expectations of a network, not for political reasons, not because it's cool, and not because it was featured in an article for CIO magazine. (Doing it in a lab for resume purposes is quite acceptable.) And if it's true for something as basic as the overall logical network design, it is almost a certainty that the same quailty decision making was made elsewhere, hiding in the shadows.

I have seen a LOT of that. And not a single online study tool, not a single Cisco certification course, not a single Microsoft handypamphlet, prepares you for that eventuality. Only hard won knowledge and bitter cynicism on the nature of humanity can sustain you though network analysis of this type.

Here's an example. Some time ago, Microsoft thought it would be a great, neat thing to get rid of all their AS/400 (a sort of minicomputer made by IBM. Incomphrensable device. Too many acronyms for non military personnel.) and replace them with NT machines. Sounds reasonable enough, right? The proferred excuse was "Y2K fix", but come on, this is MICROSOFT we're talkin' about here. It drives them up the wall to use non MSoft stuff. So, they dropped in 1500 (yes, that number is correct) NT servers to replace the 40 or so As/400's doing the work.

They couldn't pull it off. They had to go back to the old system, which worked fine.

Now, there are some bright people at Microsoft. I would expect people at Microsoft to perform Microsoft based miracles, and have a Microsoft education far superior to all other types. But they couldn't perform this task...and the task was pointless in the first place. The Y2K excuse doesn't hold water, the AS/400 had been made Y2K compliant already. The real reason was political...Microsoft can't have non-MS operating systems in their networked environment because it looks bad. They want to have NT running everything, despite the limitations of the hardware, the OS, the software, and the laws of physics. The political decision to migrate to NT was a failure.

This is called project management by some. I call it a steady paycheck.

Anywqay, I'm doing these online guides and they are just pissing me off. I tried to raid the technical library here, but it's all stuff like "How to be a manager!" and Microsoft Technet publications. Two books caught my eye, an R/3 SAP book (special edition!) and Oracle, the Complete Reference. The SAP book is totally worthless. It's a special edition all right..short yellow bus special. It's a thick tome, but no REAL information about the SAP system at all...it's a bloated marketing tool. The Oracle book is honest. Right off the bat it tells you that relational databases and fourth generation database tools are dangerous in clueless hands. I can respect that. The SAP book makes you think that pinheads can effectively install and operate the behemoth. Lies, lies, lies.

I pull my head out of the Oracle book around 11:30. I'm hungry, and the day is going by surprisingly fast. I get told that I'm going to be forced to attend product manglement classes for the next two days. Why? Because I am here, not because I need it. I'd rather be doing something else...even the prospect of free food is not enough to compel me to even want to consider attending such a class. Product manglement, like network and systems administration, cannot be properly taught in an idealized environment. I'm sure that after Wednesday, I'll have some sort of problem I'll relate here about the entire affair.

Now, for the funny part.

I zipped across the street to Japon's, the handy japanese resturant. I figured some raw fish would improve my mood. (It did.) I brought with me my book on relativity, and figured I'd have a nice leisurely lunch with no worries. (I did.)

When the waitress arrived and asked for my order, I told her I wanted the number 15 sashimi combination and an iced tea.

"The fifteen?"

"Yes, I'd like that one. And ginger dressing on the salad."

"Uhm, we don't have a fifteen."

Eh? The hell they don't. I opened the menu again, and sure enough, there's a fifteen. In hex. Christ, I'm converting decimal to hex at the sushi bar. "Sorry. I'll have the 'F', please."

Sigh.