How the internet works

[Guest]

Yes, dosxuk. Thank you. I see what you said. Weel all I can add is that if I were designing a computer I would do it differently. Wouldn't you? I'd delete the whole lot of data before using that particular store address again. Oh well. Cheers.

 

We're lucky you're not designing computers, I guess major manufacturers like Intel and AMD don't have a clue about designing technology There's just no reason for it do what you suggest.

 

You need to read up and understand how RAM works, how memory controllers work etc before pronouncing that they're built badly.

[woolyhead 15]

We're lucky you're not designing computers, I guess major manufacturers like Intel and AMD don't have a clue about designing technology There's just no reason for it do what you suggest.

 

You need to read up and understand how RAM works, how memory controllers work etc before pronouncing that they're built badly.

No, I don't think I need to read about how RAM works, just to understand what the result of failing to clear the memory is. Heartbleed gives me a perfect example, whether I understand RAM or not. By the sound of it you seem to think these manufacturer's designers and computer systems designers do a perfect job, in a field I couldn't even begin to understand. As it happens I designed testers for one or two RAM stores for BT to use in the TXE4 digital telephone exchange. I suppose you think I didn't do the job properly? OK, so how did we manage to test them adequately? What you need to do is read a book on logic and realise the difference between undestanding what the results of something are and how the something works.

[Ghozer 112]

No, I don't think I need to read about how RAM works, just to understand what the result of failing to clear the memory is. Heartbleed gives me a perfect example, whether I understand RAM or not. By the sound of it you seem to think these manufacturer's designers and computer systems designers do a perfect job, in a field I couldn't even begin to understand. As it happens I designed testers for one or two RAM stores for BT to use in the TXE4 digital telephone exchange. I suppose you think I didn't do the job properly? OK, so how did we manage to test them adequately? What you need to do is read a book on logic and realise the difference between undestanding what the results of something are and how the something works.

 

Indeed, you could understand how it SHOULD be, and what things to check if the results aren't as expected...

 

like, a car should move, and you need fuel, and oil etc to make that happen..

But could you build said car?

[Dave h-j 10]

No, I don't think I need to read about how RAM works, just to understand what the result of failing to clear the memory is. Heartbleed gives me a perfect example, whether I understand RAM or not.

 

You are assuming that the adjacent memory isn't in use and should be cleared. Heartbleed is a buffer over-read which makes the mistake of not checking what it's being asked to do correctly. So even if memory was blanked after usage, heartbleed would still exist.

[woolyhead 15]

You are assuming that the adjacent memory isn't in use and should be cleared. Heartbleed is a buffer over-read which makes the mistake of not checking what it's being asked to do correctly. So even if memory was blanked after usage, heartbleed would still exist.

Wow, is that what I'm doing? OK, what's an adjacent memory used for then? I'm sorry, I dont understand. It seems to a computer illiterate like me me that if the Heartbeat RAM was cleared after use by person A then subsequent caller B (the hacker) would find the memory empty of data. Why is that wrong? How would B still get data out?

[AlexAtkin 10]

Well you DID say

No, I don't think I need to read about how RAM works, just to understand what the result of failing to clear the memory is.

 

I do not believe its standard procedure to clear memory that is no longer in use anyway as its a HUGE waste of memory bandwidth. The same way that if you delete a file off your HDD it is not actually wiped, the file system is just changed to report that as usable space because it would block IO for much longer otherwise.

 

Plus as pointed out, the memory may still be in use anyway.

 

The issue is that no process running on your PC should ever be able to access memory outside its allocated pool in the first place. I thought this had been prevented with changes on modern CPUs, I was obviously mistaken.

[Guest]

No, I don't think I need to read about how RAM works, just to understand what the result of failing to clear the memory is. Heartbleed gives me a perfect example, whether I understand RAM or not. By the sound of it you seem to think these manufacturer's designers and computer systems designers do a perfect job, in a field I couldn't even begin to understand. As it happens I designed testers for one or two RAM stores for BT to use in the TXE4 digital telephone exchange. I suppose you think I didn't do the job properly? OK, so how did we manage to test them adequately? What you need to do is read a book on logic and realise the difference between undestanding what the results of something are and how the something works.

 

No nothing is perfect, but you don't understand any of this nor are willing to learn about it so I'm not sure why you think your suggestions are of any use to anyone? If you made the effort to learn about why these problems exist by understanding the technologies then maybe, just maybe, you'd understand why things like Heartbleed happen.

 

And here we go, "back in day I did job X which makes me qualified to talk about a subject about which I know nothing and am not interested in learning about yet am quite willing to tell everyone they're wrong about".

 

I'm pretty sure I gave up on one of your previous posts asking for an explanation and then disagreeing with the people trying to help you. So I'll do the same now.

[Dave h-j 10]

And here we go, "back in day I did job X which makes me qualified to talk about a subject about which I know nothing and am not interested in learning about yet am quite willing to tell everyone they're wrong about".

 

yeah, I'm not sure if this is just a wind up now.

 

To the OP - seeing as you profess to having a background in electronic design and, i assume, you do actually know the meaning of the word "adjacent" - then have a look through these.

 

http://www.tutorialspoint.com/computer_fundamentals/index.htm

http://www.tutorialspoint.com/operating_system/index.htm

 

ta ta

[Cyclone 10]

It's like copying a song from the radio... without the volume turned OFF....

 

You are the 'hacker', copying that song from the radio, it won't work until you turn the radio on (make the connection) even though it's still there (the radio is broadcasting)..

 

Once turned on (connection made), you know it's likely to work as the power light is on, and the frequency is correct (you're connected to the a server IP that has the bug)

 

When you press record, you start recording that radio signal (the hacker takes a copy of the data that is being sent between you and the server)

 

When you play it back, you find out what you have recorded, it could be a song, an advert, or the DJ talking... (then hacker won't know what they have, until the session is complete (all the data has been passed, complete or in ENOUGH bits -data packets))

 

Once you play the tape, you know what you have recorded, you can choose to erase it, or keep it..

 

Once the hacker has whatever information was passed between (copying it from the connection) they can then open this information (using various tools to read data) and see what they have, it could be a a simple website connection to something like FaceBook, where you are posting a status, and liking a couple of posts...

 

It could be as you log into FaceBook (as it doesn't send your user/pass EVERY refresh)

 

It could be as you submit your bank details...

 

It's totally pot luck who is affected, and what the hacker gets...

 

As said above, there's nothing you as an individual can really do, other than avoiding affected sites, however there is no definitive list of sites that are affected... (that I know of)

 

--------------------

 

Another explanation, imagine a house with two phones, you phone them, and they answer it down stairs, you and your friend are talking...

 

Another family member picks up the other phone in the other house, but doesn't say anything, they simply listen... they hear both sides of the conversation, and they could record or remember important bits..

 

the "bug" is the house that has two phones...

 

now imagine that, but with you visiting a website, the website is the house with two phones...

 

This is not a description of the heartbleed bug. It's a description of somehow intercepting (which is actually very easy) the exchange between server/client (which presumably isn't encrypted), this is why encryption exists.

 

---------- Post added 15-05-2014 at 15:42 ----------

 

Thanks for all your efforts to explain it. However I am thick and I fail to understand, despite reading what's been said over and over again, how a hacker can (a) "listen in" to what's being said between you and your server or (b) get hold of your data after you close the conection. Why and how does the server allow a second user to share your information?

It's a bug. A mistake in the code.

Please don't quote analogies 'cos they just confuse me more. You may recall that my original question asked whereabouts in the process of describing how the internet works as shown in the address given, does the hacker's begin to gain access to your data. So far everyone has shied away from telling me this. Can anyone please look this description up and tell me the answer, please?

 

No they haven't. It's been described, they send the appropriate packet to the server asking for more data to be returned than they have sent. Due to the mistake (bug), the server returns data from in memory, this is not deliberate, it's a bug. The data is effectively random, ie the hacker has no idea what will be returned each time they make the request, they keep doing it though, hoping to get something useful, like your username and password.

 

---------- Post added 15-05-2014 at 15:51 ----------

 

Yes, dosxuk. Thank you. I see what you said. Weel all I can add is that if I were designing a computer I would do it differently. Wouldn't you? I'd delete the whole lot of data before using that particular store address again. Oh well. Cheers.

 

You wouldn't if you wanted to sell your computer memory controller to anyone

More likely you'd make it impossible to read memory that wasn't allocated to the process, but that's not foolproof either.

 

---------- Post added 15-05-2014 at 16:24 ----------

 

Wow, is that what I'm doing? OK, what's an adjacent memory used for then? I'm sorry, I dont understand. It seems to a computer illiterate like me me that if the Heartbeat RAM was cleared after use by person A then subsequent caller B (the hacker) would find the memory empty of data. Why is that wrong? How would B still get data out?

 

It's used for other data belonging to the server process.

 

It's not 'heartbleed ram', it's just ram, heartbleed takes advantage of a buffer overread bug to expose data that shouldn't be exposed. This can be data that is still allocated, or data that has been deallocated. Wiping the deallocated data would massively impact performance and only partially mitigate the bug.

 

---------- Post added 15-05-2014 at 16:26 ----------

 

Well you DID say

 

I do not believe its standard procedure to clear memory that is no longer in use anyway as its a HUGE waste of memory bandwidth. The same way that if you delete a file off your HDD it is not actually wiped, the file system is just changed to report that as usable space because it would block IO for much longer otherwise.

 

Plus as pointed out, the memory may still be in use anyway.

 

The issue is that no process running on your PC should ever be able to access memory outside its allocated pool in the first place. I thought this had been prevented with changes on modern CPUs, I was obviously mistaken.

 

Is it actually doing that? Server processes can and do get reused multiple times for different clients. In which case the security of the memory between client uses is up to the application developer.

[Obelix 11]

Wow, is that what I'm doing? OK, what's an adjacent memory used for then? I'm sorry, I dont understand. It seems to a computer illiterate like me me that if the Heartbeat RAM was cleared after use by person A then subsequent caller B (the hacker) would find the memory empty of data. Why is that wrong? How would B still get data out?

 

You assume that the memory would remain empty after being cleared...

 

It's not. The memory is in constant use which is why clearing it isn't a big priority. Instead you have proper bounds checking on the buffer size to prevent the problem of inappropriate data getting loose, or you enforce processor rings on sensitive data. This wasn't done in the case of the heartbleed bug - that's the cause of the bug, not of memory not being cleared.

[woolyhead 15]

OK, OK. I give in. It wasn't a windup but from what you said I suppose I must be stupid. I'll send back my two degrees and get a job stacking shelnes. Sorry if I upset you, but thanks to everyone who actually helped me understand.

[Cyclone 10]

If you think you've got it, and you're so sure that you're right, go and get a job designing memory controllers.

Or you could start threads from positions of ignorance, wait until people explain, propose solutions, and when the people who do understand the topic tell you it's not a good solution, spit your dummy as if you've been insulted.