[mdlug] Realtek silent data corruption

[Raymond McLaughlin]

bob dion wrote:
> Jeff Hanson wrote:
>> On 7/31/07, Ingles, Raymond <Raymond.Ingles at compuware.com> wrote:
>>>> From: Mathew Enders
>>>> I think that after reading it that the data inside the packets was
>>>> corrupted not the packets themselves and that is how it passed the
>>>> TCP/IP check.  Some else asked the same question in the comments.
>>>  With TCP connections, the data has a checksum, too. But it's possible
>>> for certain patterns of data corruption to still pass the check. And
>>> if the driver at the sending end corrupts the data (reading it in from
>>> the sending application) before it puts it on the wire, there's nothing
>>> the receiver can do to detect that.
>> Maybe I'm missing something but logical deduction seems to conclude
>> that this is an improvement:
>>
>> 1.  It's a Windows-only issue.
>> 2.  The corruption could make transferred applications inoperable.
>> 3.  The data from Windows PCs includes a much higher percentage of
>> malware than non-Windows PCs.
>> 4.  Corruption of transferred applications from the affected systems
>> would cause a reduction of malware in general.
>>
>> Logical?
> 
> One could only hope. Unfortunately, the corruption only seem to happened 
> with large 1-2+ gig files.
> 
> BD

Also it seems that the corruption could, theoretically, affect any large
file that passed one hop over an affected windows machine along its
network path.