For the mainframe, one of the words I like to use is "Integrity". The
hardware designers and programmers did not want to let any error, no
matter how small, go unnoticed. A quick example: On a real 3278
terminal if you tried to type on top of a protected field, the box made
a clicky noise and locked the keyboard from further input until you
pressed Reset. This is Integrity, protecting the data entry clerk from
missing even a single character on that COBOL coding form.
Same with the OS of course - with all the logging and SMF and error
recovery and message manuals a mile long. You don't see anything close
to this in Windows and Linux/Unix. In fact, on those platforms it's
sometimes the opposite... Call a subroutine, forget to check the return
code, and the program continues on as if nothing happened - good luck.
On the mainframe, if I call an SVC and it fails I get an abend by
default with a decent message. That's just one small part of the
built-in integrity the mainframe has over the other platforms.
One of my old jokes working with Unix/Linux was: If you got an error
that makes no sense, check to see if the disk ran out of space. I was
surprised how often this turned out to be true.
Over the years one of the bigger Windows/Unix hacks has been buffer
overflows, I believe based on string copies with no supplied length (you
keep copying bytes until you hit a zero). But variable fields on the
mainframe typically had a length byte or halfword - so we avoided one of
the bigger hacking issues. Well, until USS came along and C started
being used for mainframe coding.
But this all makes no difference when you find passwords.txt on a
sysprog's laptop because the security folks decided the password needs
to be changed each month to something impossible to remember. That's
more the point I'm trying to make.
On 5/6/2019 6:52 PM, Bill Johnson wrote:
A plethora of reasons.
Lack of emphasis on security by MSFT. More interest in selling the next release
than securing each release.
Buggy code. Went to a security seminar once where it was stated that MSFT code
had one bug for every 25 lines of code. IBM was around one bug every 250 lines
and NASA was one bug every 10,000 lines. Called KLOC.
MSFT is also more available. There are millions of possible targets. That’s why
Safari was less exposed to hack than IE. IE had a much larger install base so
was more targeted.
Many other reasons.
Hackers tend to target the easier platform.
Sent from Yahoo Mail for iPhone
On Monday, May 6, 2019, 9:27 PM, Tom Brennan <[email protected]>
wrote:
Ok, but why is Windows easier to hack than the mainframe?
Personally, I'd find a mainframe far easier to hack because I know a
little bit about control blocks, APF auth, SVC's, subsystems, address
spaces, RACF, etc., and I know far less about the equivalents on
Windows. But of course the first step is to get any kind of userid, and
that's done by pretty-much the same methods - regardless of platform.
On 5/6/2019 1:18 PM, Bill Johnson wrote:
It’s why banks stay on the mainframe. Security.
Sent from Yahoo Mail for iPhone
On Monday, May 6, 2019, 4:09 PM, Bigendian Smalls
<[email protected]> wrote:
Bill, would you care to back that sweeping generalization up with some detail?
On May 6, 2019, at 22:06, Bill Johnson
<[email protected]> wrote:
Completely different. Hacking Microsoft is way easier.
Sent from Yahoo Mail for iPhone
On Monday, May 6, 2019, 3:53 PM, Bigendian Smalls
<[email protected]> wrote:
Which is how 80% of all the hacks today start. Find purchase and advance your
position. This is how the game is played. It was as classic of a hack as
anything today.
On May 6, 2019, at 21:43, Bill Johnson
<[email protected]> wrote:
Still never would have occurred without a valid userid.
Sent from Yahoo Mail for iPhone
On Monday, May 6, 2019, 3:18 PM, Charles Mills <[email protected]> wrote:
No.
From the link you cite:
"According to various sources, the hackers succeeded in finding (and exploiting) at
least 2 previously unknown errors enabling them to raise their authorisations in the
system. One of them was an error in an IBM HTTP server and the other one was an error in
the CNMEUNIX file, which in the default configuration has SUID 0 authorisations (which
means that by leveraging on the errors it contains, one is able to execute commands with
the system administrator’s authorisations)."
His "user" access to InfoTorg was not a problem for the mainframe. (It was a
problem for the MPAA lawyer whose account he accessed, but not for the mainframe in
general.) The above mainframe security vulnerability was.
Charles
-----Original Message-----
From: IBM Mainframe Discussion List [mailto:[email protected]] On Behalf
Of Bill Johnson
Sent: Monday, May 6, 2019 11:17 AM
To: [email protected]
Subject: Re: mainframe hacking "success stories"?
The Pirate Bay hack acquired a valid mainframe userid and password off of a
Microsoft laptop. In effect, not really a mainframe hack. He just logged on.
https://badcyber.com/a-history-of-a-hacking/
Sent from Yahoo Mail for iPhone
On Monday, May 6, 2019, 1:21 PM, Charles Mills <[email protected]> wrote:
#1: Noooooo. It was a legitimate mainframe hack (assuming you consider USS a
legitimate part of the mainframe, which it has been for 20 years or so). It was
an exploit of CGI buffer overrun.
#2: It drives me nuts to hear mainframers explain away mainframe breaches. "It wasn't really a mainframe
hack, they got in through USS." "It wasn't really a mainframe hack, they re-used a Windows
password." "It wasn't really a mainframe hack ... whatever." If your CEO was standing in front of
the press explaining how your company let x million credit card numbers go astray, would it matter HOW they got
into your mainframe, or only that they DID?" If your mainframe is vulnerable to a USS hack, or a shared
Windows password, or whatever, you need to fix THAT, or risk having to explain to your CEO why he got fired (like
Target's) for letting all those credit card numbers go astray.
Charles
-----Original Message-----
From: IBM Mainframe Discussion List [mailto:[email protected]] On Behalf
Of Bill Johnson
Sent: Sunday, May 5, 2019 10:00 AM
To: [email protected]
Subject: Re: mainframe hacking "success stories"?
Wasn’t really a mainframe hack. It was a laptop hack that acquired legitimate
mainframe credentials.
----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [email protected] with the message: INFO IBM-MAIN
----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [email protected] with the message: INFO IBM-MAIN
----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [email protected] with the message: INFO IBM-MAIN
----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [email protected] with the message: INFO IBM-MAIN
----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [email protected] with the message: INFO IBM-MAIN
----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [email protected] with the message: INFO IBM-MAIN
----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [email protected] with the message: INFO IBM-MAIN
----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [email protected] with the message: INFO IBM-MAIN
----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [email protected] with the message: INFO IBM-MAIN
----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [email protected] with the message: INFO IBM-MAIN
----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [email protected] with the message: INFO IBM-MAIN
