Attacking the 35$ (not) computah

they used

reliable.

With a mainframe you are essentially doing the same thing. The data is effectively in the hands of a major vendor with whom you must perpetually renew a big juicy maintenance agreement. Because if you don't ...

--
Regards, Joerg

http://www.analogconsultants.com/

On a sunny day (Tue, 13 Mar 2012 09:17:52 -0700) it happened Joerg wrote in :

I do not know what you guys used, but it was standard practice to back up all data. In an other place I worked in Amsterdam there were also backups from the main frame made on a regular basis.

That is not correct.

Cloud has severe security risks, we all know NSA is also an industrial spy organization, so from that POV companies should keep their data in house.

I have some computahs that do not even have a net connection. They carry the latest weapons of math destruction designs, The real secrets I keep in the little box behind the eyes. Unfortunately they can image thought too these days :-) You will have to stop thinking to be safe with your inventions :-)

Anyways these days it is easy to back up very large amounts of data, That holographic thingy exists that can put a lot on a small disk.

Multiple copies on multiple media in multiple locations is in fact a must for serious work.

There is the transit storage system too, it works like this: You beam your data toward the nearest star. Then when you need it you fly FTL to that star, receive your data, and fly back FTL with it. hehe

they used

reliable.

Reminds me when I was an intern at a company that sold mainframes. They had a mainframe themselves for the administration. At one point they decided to change a the hard drive. The tech responsible asked me to pull the thing out (about 50kg!) and haul it to the workshop where he formatted it immediately. He should not have done that. It appeared the only useable backup was 3 months old...

Nowadays people rely heavily on one server being available. Most engineering stuff can be done locally (I always insisted on working on local copies) but a lot of office documents are on the server.

I can get worse. At one employer we got a day off. The mains breaker for the entire building failed. They had to get a new one from somewhere deep in Germany.

--
Failure does not prove something is impossible, failure simply
indicates you are not using the right tools...
nico@nctdevpuntnl (punt=.)
--------------------------------------------------------------

data.

frame

If the backup strategy is sound. I once lost a weeks work due to bad network maintenance. The make things worse the backup strategy had a hole of one week.

Since nobody seems interested in weapons of total destruction I reveal my evil plan: chili-con-carne with onions.

--
Failure does not prove something is impossible, failure simply
indicates you are not using the right tools...
nico@nctdevpuntnl (punt=.)
--------------------------------------------------------------

data.

frame

Sure, it was done on tape. But that does not help the with dead HD at all. In a mainframe you often had only one because those cost a fortune. They were easily the size of a washing machine. When that died the backup did not help, it was crisis mode.

With a PC system you simply carry another machine into the room, spool the backed up data onto it, leave the old machine with the IT guys, done. No rush, no panic, no major line stop situation.

Running things centralized in data processing for a business that relies on it for operations can be like playing Russian roulette. At our division at Endosonics I would have never tolerated that.

Unless you have an infinite budget it is. I've seen live what happened after a big "platter unit" came down. I don't think I've ever heard that many cuss words in one day.

Ever heard of encryption? :-)

Oh-oh, now the Binnenlandse Veiligheidsdienst will send the goons out :-)

It does not help when your business relies on the cloud for operations. What happens to those companies is this:

[...]

--
Regards, Joerg

http://www.analogconsultants.com/

On a sunny day (Tue, 13 Mar 2012 19:10:47 GMT) it happened snipped-for-privacy@puntnl.niks (Nico Coesel) wrote in :

data.

frame

Garlic, lots of garlic That will do it.

On a sunny day (Tue, 13 Mar 2012 12:21:50 -0700) it happened Joerg wrote in :

data.

frame

I argued at that time with the boss if it made financial sense to have the mainframe. There were also other factors perhaps, our connection to IBM for example. But for the accounting department I think it made sense. I do not remember in those years any major mishap, although IBM techs did sometimes come to play with it. I did see a lot more problems with PCs, but of course there were more of those, for example harddisks that went tick tick tick, sometimes you could fix them by rotating the little shaft that stuck out on the side,,

Not always so simple, and then you lose a days work too (presuming backups were made each day).

I had to repair at one time one of the biggest 'platter' stores you have ever seen. And the result of a failure of that thing did not exactly go un-noticed in the country. Those disks, about 40 cm wide, 2 on a motor, 4 heads on steppers, I have replaced them.. That was 10 years earlier, 1976 or so, Ampex made them (look up HS100).

Were used to store video, and slow motion playback. Those mainframes was just kids stuff :-) Head crashes were quite common, to the point of very often, and to reduce these the thing was finally mounted in an air-tight box with filtered air being blown into it to create an overpressure. I still hold an Ampex license to repair those things,, What it did was record just a few minutes of analog video, and then in a clever way played it back repeatedly to create a slow motion effect. Now that is a real harddisk. It had no processor, it was all TTL logic, and some analog. The disk motor was bigger than your head.

Just today I did read that also skype encryption is routinely hacked, even though it is RSA and Skype could not do it themselves (in German):

I checked out that company, just curious how they did it, Encryption? forget it man.

LOL

all data.

main frame

mainframe.

IBM was always good when it came to service. However, they could not provide a miracle when the big HD unit went kerclunk. They had to truck in a replacement unit and plumb that in, then spool over the data. That could easily take all day.

With a PC you can be back in full operation within the hour. BTDT.

Time to get a new one if that happened.

were made each day).

The one time that it happened to me I lost slightly under 15 minutes. Because that's how often I back up work in progess.

seen.

country.

filtered air

I remember one HD unit from inside. The motor was bigger than in our washer and the whole thing made a jet engine sound while open.

It works. It is not about absolute safety which you never will have. It's about enough safety, just like the additional deadbolt at a front entrance. It'll make a burglar move on to the next house and that's all you need.

[...]

--
Regards, Joerg

http://www.analogconsultants.com/

software they used

THAT reliable.

Oops :-)

BIG mistake. I never do that, ever. Even in this here small consulting office there are two. Plus a code-locked backup I carry with me. Plus an offsite backup.

In some countries people would have walked over to their cars and get the jumper cables out :-)

--
Regards, Joerg

http://www.analogconsultants.com/

data.

frame

Weapon of ass destruction?

Since the '80s most mainframes had redundancy built in.

data.

frame

What sort of "mainframe" has one disk drive? I didn't think you were that old?!

I don't thing a "mainframe" would even run with only one disk drive, in any year there was such thing as a PC.

Don't be ridiculous. It's precisely BECAUSE of their reliability and redundancy that they lasted as long as they did (before changing their name from "mainframe" to "server" ;).

Oh, good grief. Just what do you define as a "mainframe"?

I was thinking the same thing, but encryption has its own problems.

Depends on whether they own their "cloud" or not. ...and of course, how well it's run. I'm quite sure a lot more data has been lost on PCs than it ever has been on mainframes.

software they used

THAT reliable.

Don't be absurd.

all data.

main frame

For example, the one at the Institute for Measurement Technology at Aachen University, Germany. I happened to do my final project right next to the big mainframe room.

Of course it has multiple platters in it but I have seen several that hand only that one box, about the size of a washer.

Like this?

Whatever the guys that bought it and those that installed it defined as a mainframe. Or the guys from the link above.

Yes, it's cumbersome. But I would not entrust my stuff to a cloud anyhow, or to a any centralized computer system for that matter.

I agree, but that was nearly always lost due to sloppiness. In about 30 years I have hardly lost anything with a PC. I had one HD croak on me out of the blue but that had a few rough landings and shaky flights behind it. I lost not even 15mins.

Mainframes with proper maintenance procedures in place don't lose much either. But they can literally shut down a place when they go down. I had one client call me with a request to dig out some really old projects. "Ok, I'll send them over but you've got all that" ... "Yeah, but Bertha took a major dump today and the service folks said it could be a couple of days".

--
Regards, Joerg

http://www.analogconsultants.com/

software they used

THAT reliable.

So what do you think happens when you let a maintenance contract for a mainframe lapse?

--
Regards, Joerg

http://www.analogconsultants.com/

all data.

main frame

You still haven't defined "mainframe". An 1130, maybe, but no one would seriously call it a "mainframe". It was also a bit before '80. I don't think post-'80 models even worked with a single drive.

Model, please. I'm *really* curious what you think constitutes a "mainframe".

Doesn't load. What was the failure? Four to five nines has been pretty standard for a couple of decades.

IOW, you haven't a clue.

As we used to tell people when I was working on the *mainframe* crypto coprocessor, "lose your keys and ICRF becomes a high-speed data shredder" (to which they invariably suggested keeping a copy in the clear ).

The same is true for data lost in mainframes. The difference is that it's someone else's incompetence that did it. ;-) OTOH, if you're smart, you hire competent people to do this junk so you can do what you do best.

Hire competent people. Where I am now, they have similar problems. They didn't appreciate that suggestion either.

software they used

THAT reliable.

They don't come in and auction your data off to the competition. I said, don't be absurd but you're doing an excellent job of it.

No, asshole, it is for the xeon, but it doesn't have to be a server, you stupid f*ck.

I am building a dual xeon here, idiot. Not a server.

We buy several hundred HP DL 360s a year, and they ARE new.

YOU ARE a loser.

all data.

main frame

So you'd call the 1130 a personal computer? :-)

Anything that runs applications for numerous clients in a centralized location. For example the IBM 370. I don't remember which one we had at the institute but I know for absolutely sure that it had a cube-shaped hard drive box, huge. The professor warned us to never ever bump into that because that could bring the whole place to a grinding halt.

Quote "When asked about the nature the mainframe failure, the spokesman was unable to identify the exact problem, although he did confirm the mainframe in question is from IBM". All we know is that it died and left lots of bank customers sans money and without online access. No ATMs, nada.

So you think the installation is not a mainframe while the manufacturer says it is? Interesting.

In those days it was also good policy to keep paper compies. And to some extent still is. Weren't the CAD files to a rocket lost because the tapes became unreadable?

Minor issue: You do not always have an influence over which people are retained to do such tasks. Most certainly not if the mainframe is somewhere else.

Probably because it means a lot of work if they take your suggestions and make them happen :-)

--
Regards, Joerg

http://www.analogconsultants.com/

all data.

main frame

Actually, yes. In fact, it's been called the FIRST personal computer. ;-)

Really? I don't consider a PDP-11 (or 8, even) a "mainframe".

Other than the "PC" sorts of 370s, I sure never saw one with a single spindle. Again, I don't think they'd even work with one. Maybe in DOS, but that was dead long before the 370.

So no information whatsoever. It was likely someone pushing the big red switch. ;-)

The "manufacturer"? You said the user said it was. You don't know what it was but it only had one disk drive. You don't know much about it at all. Hard to argue with jello.

Perhaps. NASA lost a lot of data because they no longer had drives that would read the tapes. No one really expects competency from the government.

BTW, when was the last time you used tape? Disk drives are incredibly cheap.

well

If you're the boss, you do. If you are and don't have that control, YOU should be fired.

All I can say is that it was a good thing someone had a copy of my partners schematic. The network backup was two weeks old. Even the half-bit company I came from had daily rotating backups (not all readable, though). I was amazed that a BIG company was so lax with its engineering data.