Please create an account to participate in the Slashdot moderation system


Forgot your password?
Check out the new SourceForge HTML5 internet speed test! No Flash necessary and runs on all devices. ×

Comment Re:Flood defenses? (Score 1) 135

That's a great idea, because it will kill the whole 'Internet of Things' idiocy overnight. No-one will risk attaching anything to their network if they can't verify it's secure.

Well, that's one potential side-effect - and not necessarily a bad one, in my opinion. Either they learn how to manage their devices, or don't connect them to the Internet.

Comment Re:Flood defenses? (Score 3, Insightful) 135

but it isn't as simple as just blocking a few IPs.

And this is why people need to be fined, if a device on their home network is found to be part of a botnet. Individuals need to be responsible for their networks, because the authorities are virtually powerless against botnets, Unless it costs them money, people just won't care.

Comment Re:How does it work? (Score 1) 109

I'm not 100% sure, either, but based on what I'm reading, this exploit requires some type of local access to use directly. While it's not as bad as all the hype, it's still not great, and can still be exploited remotely; it just takes an extra step.

Say you're running a web server, and Apache has a buffer overflow vulnerability. A hacker can break in and, normally, only has access to whatever the "apache" user has access to. If the hacker knows about dirty cow, he can now give himself root access.

Comment Re: The only way this will get fixed (Score 4, Interesting) 164

Fines wont work unless they are income based

So fine the people who own the devices. Start with a small fine, like $10, then double it for each repeat offense. Eventually, the word will get out, people will stop buying products from that vendor, and sales will suffer. They won't have any choice but to make their products secure.

Comment Re:Can Slashdot stop post IBM advertisements? (Score 1) 61

Meanwhile, in the real world, most systems are not CPU bound but IO bound.

Seriously. According to IBM's literature, entry-level pSeries systems do 96GBps per socket. I don't know of any Intel-based systems that can even touch that, in the price range they're talking about.

I wonder if their process synchronization/IPC is faster, too.

Comment Re:massive parallel processing=limited application (Score 1) 114

With a multiuser, multitasking OS you can have 25 different unrelated processes running on something with 25 cores.

In practice, most jobs running on a computer have some relation to each other, and the more jobs you have - and this CPU clearly expects to be able to run a lot of jobs - the more likely that will be. (Where I work, we actually have an application that gets slower when you add more cores.) Like most CPUs with high core counts, this one looks like it'll be great at compute-intensive tasks, but as soon as you try to do I/O, it'll slow to a crawl. Given the number of terabytes people are trying to process these days, I'm thinking this CPU's applications are limited

Slashdot Top Deals

Enzymes are things invented by biologists that explain things which otherwise require harder thinking. -- Jerome Lettvin