Harvard student thrown off 14,000-core super ... for mining Dogecoin A Harvard University student is in hot water for using the Ivy League school's 14,000-core supercomputer to mine Dogecoins. According to The Harvard Crimson, an unnamed student has been stripped of access to all of the prestigious uni's research-computing labs and systems after it was discovered earlier this month that s/he …
A somewhat poor quote, as other non-TV but still official novels establish that others had done so too. Scotty in particular, by realising that the computer used a simplified almost-realistic model for disrupter/shield interaction and calculating an attack that would cause this model to glitch - though he admitted it couldn't have worked outside of a simulation. After that stunt, he was strongly urged to switch career track from command to engineering. Other tricks included invoking an obscure Romulan custom to challenge their captain to a 1-on-1 duel, thus buying enough time to achieve the primary rescue and escape objectives, even if at the likely cost of the commanding officer's life.
"A somewhat poor quote"
It's a quote from arguably the best of the Star Trek movies with the original series crew and you call it "a somewhat poor quote"? Books may be popular but I guarantee you (internet guarantee!) that more people have seen the movie referenced than have read the book you're talking about.
I'm not kidding about this, but if you do something stunningly audacious, and pull it off successfully (and don't hurt anyone) any punishment will be a formality. People love stuff like that. They also don't like broadcasting their oversight failures, so it just goes away. But, that word 'successfully', among other things, means you don't get caught until after you've done it and walked away. People do not love walking in on you when you've got one hand on their (thing) and one leg already out the window.
Do as you will, don't get caught.
Besides, if this kid was paying regular tuition and not a gifted hardship case or employees child they'll get a single semester of academic probation and pay a fee to the school. Most universities here have 'forgiveness' programs and all the high end universities do.
Do you know what happens when you get funds for non-profit, or 'not for profit' research grants, than then get diverted to 'for profit' activities?
Depending on where the money came from... it goes away.
Then there's the damage to your reputation.
Then there's the issue of theft.
The student was guilty of theft of services.
At 'Hah vahrd' there is still a concept of an honor code. This act would have been a violation of said honor code.
The point is that the student either didn't give a rat's ass about the consequences or didn't stop to think about the consequences.
I think he should have read more William Gibson.
I hear what you're saying, very well (there's another comment I made below about gifts/donations) but you're not seeing the larger picture.
Students, and young people in general, who are otherwise well behaved, are never, and have never been, subjected to the stiffer sides of punishment/behavior modification unless grievous injury/death occurred or just colossal damages occurred. And that's as it should be.
It has been known for millennia that over prosecution of young people for fundamentally harmless breaches of the rules is the #1 way to create an angry person with fury escalating as the repercussions of a overly stiff punishment reduces their opportunities for the rest of their lives.
So you smack their wrists, just so they don't forget they aren't as clever as they think, and send them back a little embarrassed and foolish, but also with a better idea of what is acceptable and what isn't. A punishment that isn't a crippling burden gives them a bit to think about.
Once they leave school and enter the professional/career phase of their lives they aren't going to get a second chance. For the 'good' kids that's more than enough to at least think about future actions. For the shitty little Jr. Despot probably nothing will work, but we've got a system for them.
There is less than nothing to be gained by derailing an otherwise 'good' student. Being corrected by your school is part of the 'growing up' phase. You live, learn and move on.
What you're talking about is organizational intent to misuse funds/resources in aims beyond their purview. Yes. That is bad and could have huge implications for a lot of people. That's just premeditated conspiracy to engage in criminal behavior. Fuck them, off to jail you go. A stupid student doing what stupid students do is part of being a student though.
If you never got in trouble at school you really wasted that opportunity, that's a shame. I was in my second year of university when I found out racketeering wasn't getting a tennis racket restrung. I could have gone to prison all for what I just thought was a clever way to fund the alcohol and marijuana so crucial to the success of any good academic program. If that had occurred it's entirely possible that some of the IT hardware in your office would look different and be a lot more expensive.
All the things I learned about business and manufacturing optimization in prison really wouldn't have translated well to the commercial world. After all, how many shanks can you sell to a person. A lot of my success was in enterprise oriented products, maybe I could have tried something similar after getting out of prison. I'm not sure what enterprise or institution would want millions of dollars worth of shanks made from toothbrush handles, but who knows!
A hundred years ago I got in a little bit of bother at the Human Genome Mapping Project where I legitimately had an account and (stupidly) ran a setiathome client on one of their Sun bioinformatics servers. I had fantastic through put but it slowed down legit work for the other users. They found it, deleted it, and I received a polite email telling me they had found it, deleted it, and please don't do it again....
Hey, if your bit of the work had included a genuine alien signal, you would have been hailed for creative use of resources!
Me, I'm just hoping the bodies frozen at Area 51 are *not* in a frost-free freezer. Bad stuff can happen when it periodically warms up to melt the frost...
As a student in the late 70s I got a copy of the old "Colossal Cave" adventure, and after some work managed to get it running on one of the university mainframes. Along with a few (I thought) friends I had a great time playing this, until a few weeks later I got a summons from the head of the computer centre. He dropped a large pile of fanfold paper in front of me, which I quickly identified as ADVENT.FOR. He then asked me to explain why 80% of billable university computer time for the last month had been down to that single program.
oops.
I was pleasantly surprised when he only told me to take steps to restrict it to authorized users (undefined) and off-peak hours, or else it would be deleted. I authorized anyone who asked, but never heard any more about it.
Reminds me of when me and many of my friends were using the university mainframe to play MUDs. Most of the people involved were using accounts they'd acquired through the physics department, suposedly for accademic use. When the department found out they really wern't happy and placed a ban using those accounts for MUDs.
Some of us however had acquired our accounts through a student society, so there was absolutely nothing the department could do (the mainframe admin didn't mind, just asked us to keep to the less popular terminals and surrender them if anyone else wanted them).
A.C. since my username here is my account name here is what used to be my username on that mainframe.
hehehe. My High School was the HS for a college here, which is pretty common over here (most High Schools are either linked to a university's "system" or are wholly part/owned by one) but in my HS case, we actually were in the university campus… so we shared everything with the actual college dudes. So one of the students had done some work for some company which had a Red Hat Linux 5 server somewhere. Naturally, he compiled a CircleMUD variant and put it up for everyone to use. So we used the university workstations to play on a MUD secretly running at a server. hehe. The computer lab admin for the lab where all the RS/6000s were (the ones we used for MUDding) actually allowed us to do this as long as we didn't do it when real work was needed to be done.
After a year and a half of this, someone got wise on the company and killed the MUD. A sad day that was...
His mistake was to be sneaky about it. If he'd just applied for runtime for his "experimental novel internet connected computational fluid dynamics program" he might have got a whole week of runtime without anyone batting an eyelid. They might even have commiserated with him when his results turned out to be "garbage" (but in fact comprised a tidy stash of virtual coinage). How often do sysadmins actually look at a student's results?
Presumably these supers are connected to the internet, even if hidden behind some firewalls, etc. I bet there's plenty of hackers out there who must be thinking about trying to steal some runtime off them for bitcoin mining.
And I'm guessing that a supercomputer, in the interests of achieving best performance, isn't running an AV package or many of the normal defences. So once inside the owning institution's outer protection it might be quite easy to take control and mine away, at least for a short while. I'm hoping that their sysadmins are wise to that sort of problem.
It's not so much the risks of using the computer, it's a matter of the rules attached to it. Kind of like 'student edition' software. When universities, especially high end universities, want to buy something like a supercomputer they go to their alumni for the funds (why spend your own money right?).
The school will have a 'general' set of policies and rules about how alumni funds and gifts are used and for really gigantic, specific purchases the donors add caveats of their own. I funded a materials research lab at the school where I did my graduate work and non-academic, commercial work is strictly prohibited for students and the school itself. Some schools have a nasty habit of edging students doing pure research out and sticking faculty and industry funded students in to do work that the university will protect (patent) and license out for big money.
I'm sure the computer in this story has similar rules attached to it. Paying for the equipment and its maintenance to further education is one thing, but I'll be damned if I'm going to do that, plus hire graduates from there, if they are competing with me in my own space and trying to license discoveries to me that I paid for discovering. They can do that to the government, but not to people who throw money at them. In my case, the maintenance and operating costs of the lab are tied to compliance with my, simple, rules. I'm sure this computer is the same. So if they catch someone breaking the rules it is very important to broadcast the issue very loudly.
Even though some universities are very wealthy, they can't just move money around internally like a regular business. If the funds to operate the computer are cut off they simply can't replace them in a timely fashion.
I mentioned it in an earlier post, but it won't be too bad on the student unless he's just a Class-A Fuckup who causes trouble all the time. Nobody got hurt, it'll all be OK and honestly, it might serve him well when he graduates and finds out there are 300 other Graduates from his university in line in front of him for every job. Anything to get you noticed in the sea of expensive blue suits covering shoulders still too thin to fill them out.
Besides, even though it isn't required, if it's a silly, fairly funny and audacious thing the school will ring the people who fund the asset for guidance on the punishment. Like I said, they really, really want you to know they're complying with your wishes.
I responded to your earlier comment.
Its not alumni money, but grant money.
And that's the thing.
You lose grant money sources, there goes your R&D and then your researchers and then the school slowly goes down the drain. However this is 'Harvard' we're talking about where the students have more money than common sense.
Grants are 'gifts', as are trusts, alumni donations and plain old donations from 'outsiders'. It really doesn't matter where the funds come from if the funds came with commercial restrictions. You're correct that breaking those restrictions can result in pretty serious penalties.
But a student being dumb is not remotely comparable to an organization engaging in actual commercial activities. Very, very few donors are going to screw an organization to the wall due to the actions of a student. Not only are most donors not that big of a cock, but it would look really, really bad if they advocated serious punishment.
Groups who issue grants and those who make large consolidated donations are 100% dependent on their reputations to keep donations coming. There's usually a pretty good incentive, as most philanthropic entities get 100% of their operating funds through those donations.
"His mistake was to be sneaky about it."
Well, yes, he could commit fraud.
The problem with that approach is that if you construct an elaborate smoke screen and get caught, then you can't pull the "oops, sorry" argument. It is most definitely a well sonstructed frad and opens you up to far more serious charges.
"Sounds more like his mistake was to overdo it."
Nah, he under did it. If he'd made a fortune out of it he could have split the proceeds with first the BOFH or, if the BOFH was unaccountably honest, the University. "He's a very Naughty Boy, but he has given us a nice fat cheque so we'll let him off with a First"...
At any rate, he may be sitting on a stash of crypto currency that in future years becomes worth a lot of money. I wonder whether as a notable Alumnus he would be prepared to me a donation?
This post has been deleted by its author
The CPU 'black holes', as you so crudely describe them, are very well defined processes crucial to modern IT.
How can you not know these things? Are you actually in IT? If so, how long? What color was the enclosure on your first computer? Jesus man. Do you know how stunningly ignorant you look right now? Since you obviously skipped chapters 1 - 4,351 I'm going to, this one time, explain this for you. Do not ask again.
About 50% of a CPU's 'invisible' capacity is devoted to the NSA compliment 'Freedom Safe' and 'Dogmatronix' data collection and analysis tools. Some variance in overall resource use is expected and is down to whether or not the processor manufacturer has a agreement with your country. All CPU's made anywhere in the world within the last 12 years use a USA Theme. The technology is quite impressive.
The CPU is slipping 'freedom expressions' into everything displayed on your PC. Thousands of images (freedom expressions) are seen (by your subconscious anyway each hour you are at the computer. Bald eagles, George Washington, 3,800 calorie cheeseburgers, guns and images of tax collectors being tarred and feathered along with pictures of Jesus and apple pie.
The impressive part is that the CPU is context and content aware so if it detects something 'unfavorable' it can increase the speed and duration of the images it shows you. For example, if it detects anything to do with evolution, safe sex, charity, women being allowed to wear pants, birth control, budget balancing, old people, poor children, gun control, subjugation of a populace through commercial entities, black people, brown people, voting fraud, education, social responsibility or the inherently flawed 'trickle down' economic policies the images attributes are adjusted accordingly.
But, some countries don't want the USA Theme so themes are created for them and may impact performance depending on how many things they hate.
The remaining 'invisible processes' are related to analysts of your financial situation to determine what sorts of things you like to buy throughout the range of your financial latitude. That data is sold to marketing companies.
In the future please try to have your basics covered before you start making stupid statements.
Years ago I was in a senior project role at one of our National Laboratories and one day myself and the other team leads were disappeared by facility security. I mean disappeared like in the movies. I came through the lobby oft building and as soon as I walked through the metal detector four guards materialized (I honestly still don't know how they did it. I went back many times trying to figure it out) two put hands on my shoulders and we went right out the back entrance, into a van, and ended up being interrogated in another building for hours by myself, then the other team leads were brought in.
The issue was that a fairly serious accident had occurred at a small, private, chemical company a few weeks prior. A bunch of people had been hurt (nobody died or was disfigured) and millions in damages done. The problem was that the company was suing a small engineering firm that was owned, through an alias, by the guy who managed resource use for one of the (at the time) stupendously powerful supercomputers there.
He not only managed resource use, he led the QA and efficiency team that polished the code others had written for the computer. The people who do those jobs are fairly rare even today, back then more people had been on the moon than people in his field, so oversight was kind of lax by default. This guy had opened an engineering consultancy and was using the labs supercomputer to run simulations for just about everything it turned out and he was charging big bucks for it as well.
My project was currently the biggest user of the computer and they wanted to know how we couldn't know how this was going on. It wasn't very hard to explain. We were dealing with stabilizing state changes in a number of exotic unstable alloys and the only thing we knew was that a couple of times a month we would feed data into the computer and a few days later the computer would tell us we just wasted several weeks of work (obviously, not really wasted, but not what we wanted either) and if they would care to look, there are five tons of tractor fed printer pages in queue for the document disintegrator (furnace) and they were free to dig through all that shit if they wanted. Didn't even need clearance either, it was thousands of pages of weird numbers with a reference code for that part of the project (actual codes are classified material) and if you looked hard enough, without blinking, you could actually see where FAIL and NO CAREER PATH were buried, but that was it.
Anyway, this guy had been writing the simulation code (with outside help) and forging work orders for his team to test and run the code and nobody knew because those guys never know what they're working on, they just do as ordered. Since the work orders were forged the resource and timeline stats never got input into the system and he altered the authorized code the fudge the stats and account for the overages in case of an audit.
He had that shit locked down tight. It was only by mistake he got caught. If he hadn't accidentally blown up a chemical plant he would have walked away with an estimated $15M. Not gone to prison and prohibited from ever working as an engineer of record any US company, agency or majority US owned overseas projects.
Things are obviously a little more regulated today and everybody from theoretical physicists to microwave popcorn manufacturers use supercomputers so costs are better monitored, because governments aren't the only people with supercomputers, so somebody other than the taxpayer has to fund them. But at the same time, 'common' HPC systems are so large the entire Manhattan Project could run on the interconnects alone. You've got a different type of computer savvy people crawling everywhere on the planet, and you can use those systems for direct returns (virtual money) without needing millions of dollars of other equipment to make use of the computer output (exploiting metal alloy chemistry requires you to have the stuff to exploit it with, even after the computer told you how to do it). I fully expect more of this sort of thing and would not be remotely surprised if there aren't a few organized groups of people out there right now making a lot of money just hiding stuff in the excess capacity that rattles around in all HPC systems.
I read a news item a few weeks back about some guy who was using uni resources to mine crypto currency. He obviously was interviewed anonymously but I wonder if it could be the same person. Perhaps Harvard saw the news item and decided to dig around in their systems to make sure it wasn't them being exploited and caught him. It may just be coincidence though.
Today Asok Learns that life is not like Star Trek:
[ Posted URL shortened by Mod to http://reg.cx/29NT to avoid forum layout breakage - I want to sit down with the person who came up with that URL scheme ]
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2024
