Follow Slashdot blog updates by subscribing to our blog RSS feed


Forgot your password?
Note: You can take 10% off all Slashdot Deals with coupon code "slashdot10off." ×

Comment How are you using the data? (Score 2) 219

What clients will you be exporting it to? Linux, OS X, Windows? All three?

What kind of throughput do you need? Is 10 MB/sec enough? 100 MB/sec? 10 GB/sec?

What kind of IO are you doing? Random or sequential? Are you doing mostly reads, mostly writes, or an even mix?

Is it mission critical? If something goes wrong, do you fix it the next day, or do you need access to a tier 3 help desk at 3 am?

We have a couple of petabytes of CMS-HI data stored on a homegrown object filesystem we developed and exported to the compute nodes via FUSE. Reed-Solomon 6+3 for redundancy. No SAN, no fancy hardware, just a bunch of Linux boxes with lots of hard drives.

There is no "one shoe fits all" filesystem, which is part of the reason we use our own. If you have the ability to run it, I'd suggest looking at Ceph. It only supports Linux, but has Reed-Solomon for redundancy (considered it a higher tier of RAID) and good performance if you need it. If you have to add Windows or OS X clients into the mix, you may need to consider NFS, Samba, WebDAV, or (ugh) OpenAFS.

Comment Fail deadly (Score 1) 256

Venus? A floating colony in Venus's atmosphere is the very definition of "fail deadly". Anything goes wrong you are dead, whether dead quickly or dead slowly. Plus, given the conditions on Venus, if there ever was an ecology, it has long been reduced to ash. It is also not likely we could terraform Venus (reduce the atmosphere and spin it up) given the resources of the entire solar system to do so.

If I were planning humanity's journey to the stars, I'd go with the moon first, followed by Ceres. Resource rich, low gravity, and far enough out of the Earth's (in the moon's case) and the Sun's (in Ceres' case) gravity well to make exploring other places much easier.

Comment Re:When do we get a real boost over 2013 speeds? (Score 1) 126

I do high-performance computing for a living, and Moore's Law has been on its last gasps for a while now.

Until around 2006, the smaller you made a transistor, the faster it could work. This was called Dennard scaling. But once transistors reach a certain size, current leakage and thermal issues prevent you from making the transistors faster.

While they can't drive transistors any faster, smaller processes still allow them to put *more* transistors on a chip. This is why we've gone from single-core to multi-core to multi-core with GPU compute on a die.

Despite all the complaints about "CPU's haven't gotten much faster since Nehalem", they *have* gotten quite a bit faster. You just have to rewrite/optimize/recompile your program to take advantage of multi-core, GPU compute, and SIMD instructions like AVX2.

This is the primary reason programs aren't running much faster than before. Silicon isn't getting any faster, and rewriting programs to scale isn't easy and sometimes isn't worth it so many people don't. Moore's Law no longer results in "free", "easy" speed-ups.

CPU's for the next few years are looking pretty incremental. I'd expect a one-off moderate increase in single-core performance once Intel moves off silicon onto III-V semiconductors (10 or 7 nm?), but past that you will likely be waiting several years for your graphene/nanotube/topological insulator/spintronics overloads to deliver something substantially faster.

Comment Customers... (Score 1) 49

Software has zero intrinsic value. It doesn't generate a single cent (unless you've written a BitCoin miner, I guess).

Customers, on the other hand, can generate lots of value if they use your software. Customers and the potential for more customers are usually the reason small software firms get acquired for Rockefeller money by the Google's and IBM's of the world (the other reasons are usually acquiring patents or the talent of the development team itself.) The software itself is rarely the target.

Open-sourcing the software increases the odds of someone using their software, either because it's "free", or because having the code in hand keeps them out of trouble if the company were to fold. And even if they're using it for free, it increases the odds that they would be willing to use a paid version at a later date, which is valuable.

And companies pay for reliability, both for necessity and so they have someone to pass the blame to if something fails. Even if someone got a copy of their code and decided to try their own business, are you going to trust them over the original creators when it comes to your job security?

Comment I wonder why... (Score 5, Informative) 289

You have to admire the hypocracy of state legislators who argue for "state's rights", who don't care about "city and county rights" to roll out broadband to attract jobs and new people to their area. It's almost like they were hypocrites, ignorant of freshman economics, sold to the highest bidder or something... /Lives in Tennessee, has the same bunch of ignorant cretins passing laws that an 18 year old freshman could easily shoot down as dumb.

Comment Re:Won't save most of the 4000 lives (Score 4, Interesting) 615

To give a counterexample, I was driving down a long hill that I have driven daily for 20+ years. At the bottom of the hill, right before it went around a curve, I saw cars hitting their brakes, and knew there was probably a traffic jam around the corner, so I started slowing down.

There was a truck driver pretty far behind me, and he didn't bother slowing down until he came around the curve, saw the traffic jam, locked his brakes, and ran off the road, and blamed me for the accident.

I'm a physics major, so I measured the location of where he locked his brakes, and the point he came to a stop. A little high school algebra showed he was moving 80-85 MPH in a 70 MPH zone when he hit his brakes.

For that reason, I subsequently installed a dashcam in my car. It pays for itself the first time some idiot lies and tries to pin the blame on you.

Comment I want this to be true, but... (Score 5, Insightful) 480

I want a non-Newtonian drive as much as any other nerd out there, but it's still more probable that (assuming it works) it uses conventional physics, just in ways they haven't figured out yet.

That said, I think this result is the point where NASA, DOD, Lockheed Martin, Boeing et al should turn on the money spigot for research. There's obviously something going on, even if it's just conventional physics in unexpected ways. And on the odd chance it *is* new physics, the results could change the world.

Comment Don't... (Score 2) 315

I'm as geeky as they come. Most of a Ph.D. in theoretical physics, and have spent over a decade working in high-performance computing.

When I was 7 years old, I was wandering through the woods, looking under rocks for creepy crawlies, playing hide-and-seek, and playing baseball with my brother and cousins. Not only did it *not* set me back in anyway, but it is some of my fondest memories of being a child.

Let kids be kids for goodness sake. Take him to a science museum, and let *him* tell *you* what interests you. When I was a kid and hyped about computers, my dad thought computers were a fad only used to play Pac-Man. Not only do I have a good-paying career, but any time dad can't connect to the internet, I get an emergency telephone call.

Let your child steer his future. He's the one who has to live it.

Comment Re:Users are *bad* at choosing passwords (Score 1) 159

Passphrases *can* be done securely; most people won't. They will concatenate simple words, which means if I have a dictionary of, say, the top 1,000 words, it's still reasonably feasible to crack.

For instance, here are some long passphrase-like passwords that I cracked from the LinkedIn debacle. They used plain MD5 as the hash, which admittedly helps cracking a lot. I haven't tried the depleted hash list in a long time, but I'm willing to bet with advances in both OCLHashcat and my own skills, I could get quite a bit more.

24 sociological imagination
24 linkedinlinkedinlinkedin
23 newlinkedinpassword1234
22 harekrishnaharekrishna
21 networknetworknetwork
21 managerialeconomics23
20 vaffanculovaffanculo
20 serafimovaserafimova
20 Restoration Hardware
20 powerpowerpowerpower
20 keepitrealkeepitreal
20 kazakhstankazakhstan
20 internationalnetwork
20 crisscrossapplesauce

At the end of the day, there's just no substitute for a long random password.

Comment Users are *bad* at choosing passwords (Score 5, Insightful) 159

I run a GPU cracker on my user's password hashes to preemptively weed out weak passwords. Several times I have seen them try to change it from (for example) "password" to P@ssw0rd99", which in a certain sense is significantly more complex, but OCLHashCat has rules for capitalization, leet-speak, appending/prepending numbers. You've only changed the time it takes to crack that hash from fractions of a second to a few minutes.

The only highly secure password requires long, random characters. Given a choice, users will always prefer an easy-to-remember password because it makes their life easier. Unfortunately, it also makes the bad guy's life easier, and the sysadmin's life harder.

Websites should be required to disclose the hash format they are storing user's passwords in, to hopefully prevent another Linkedin plain-md5 type debacle.

The only difference between a car salesman and a computer salesman is that the car salesman knows he's lying.