Wednesday, December 14, 2011

Oracle Corp reference one of my test results

Oracle refenrce one of my tests...

http://www.oracle.com/us/corporate/press/497230

On-chip Cryptographic Acceleration – New crypto units support over a dozen industry standard ciphers, enabling security conscious organizations in industries including telecommunications, healthcare, financial services and the public sector to keep their data safe with up to 44 percent faster secure queries than the latest generation of x86 systems when encrypted with Oracle's Advanced Security Products(4), 3x faster Oracle Solaris ZFS file system encryption than the latest generation of x86 systems(5), and 4x faster single-thread OpenSSL security than IBM POWER7(6).

Footnotes:
(6)Comparison is based on internal testing of AES-256-CBC encryption at 8K using OpenSSL against published test results for IBM: http://xmlisnotaprotocol.blogspot.com/2010/10/openssl-098-benchmark-on-power7-35ghz.html.

Wednesday, November 2, 2011

Compromising emissions!

Been reading about various side channel attacks, I think my skills (lateral thinking, electronics, 'puters, physics and math) could lend themselves well to this ...where do I sign up!
basically is eavesdropping information leaked via power,RF/EMI,sound,light or other means to gain information

http://en.wikipedia.org/wiki/Power_analysis
http://youtu.be/4L8rnYhnLt8
(RF/EMI demo)

http://syhw.posterous.com/two-amusing-side-channel-attacks
(USB port power and sound)
http://cs.tau.ac.il/~tromer/acoustic/
(PoC for picking up sound from mobo capacitors to break RSA )
http://digitallounge.gatech.edu/digitallife/index.html?nid=71506
(Keyboard taps)
http://lasecwww.epfl.ch/keyboard/
(keyboard EMI)

http://www.wired.com/threatlevel/2007/08/researchers-cra/
(car keys...brute forcing the private key from all challenge/responses)

http://www.pop.is/1eyo
(a physical side channel aka safe cracking)

Padding oracle attacks
http://www.usenix.org/event/woot10/tech/full_papers/Rizzo.pdf
(decrypt ciphertext without knowing the key eg to bypass CAPTCHA )
http://hal.inria.fr/docs/00/70/47/90/PDF/RR-7944.pdf
( Efficient Padding Oracle Attacks on Cryptographic Hardware )


RSA FOB tokens
http://www.geekosystem.com/broken-tokens/


http://www.newscientist.com/blogs/onepercent/2011/11/encryption-for-transit-cards-h.html
(Crack 3DES smart cards with an RFID reader and an oscilloscope, via power analysis of the chip in the card while de/encrypting)

http://www.techwarelabs.com/rfid-hacking-is-it-a-threat

http://www.cl.cam.ac.uk/~mgk25/ieee02-optical.pdf
(Read displays at a distance)
http://applied-math.org/acm_optical_tempest.pdf
http://dl.acm.org/citation.cfm?doid=545186.545189
(Flashing lights on your network kit may be a spanned port!)


TEMPEST backronyms from wikipedia

Tiny ElectroMagnetic Particles Emitting Secret Things
Transmitted Electro-Magnetic Pulse / Energy Standards & Testing
Telecommunications ElectroMagnetic Protection, Equipment, Standards & Techniques
Transient ElectroMagnetic Pulse Emanation STandard
Telecommunications Electronics Material Protected from Emanating Spurious Transmissions

Tuesday, October 25, 2011

Russian roulette

#cat clint-eastwood.sh
while :
do
echo "are you feeling lucky punk (yes/N)?"
read a ; [[ "$a" = "yes" ]] && kill -9 $(echo $RANDOM)
done
#./clint-eastwood.sh
are you feeling lucky punk (yes/N)?
n
are you feeling lucky punk (yes/N)?
yes
kill: 8051: 0403-003 The specified process does not exist.
are you feeling lucky punk (yes/N)?
yes
kill: 18234: 0403-003 The specified process does not exist.
are you feeling lucky punk (y/N)?
yes
kill: 16882: 0403-003 The specified process does not exist.
are you feeling lucky punk (yes/N)?
yes
kill: 13023: 0403-003 The specified process does not exist.
are you feeling lucky punk (yes/N)?
n
#yes|./clint-eastwood.sh

Cisco IOS on Juniper hardware?

This FOSS networking looks interesting e.g. OpenflowSwitch movement http://www.openflow.org/ http://mvdirona.com/jrh/TalksAndPapers/JamesHamilton_CleanSlateCTO2009.pdf

Trying to open the network stack -networking is where mainframe was years ago (proprietary everything

SANs are just fighting the ASIC vs. x86 battle now too e.g. http://blog.nigelpoulton.com/3par-v-class-the-best-just-got-better/#comments
http://opennetsummit.org/ perhaps when its all converged it will be all commodity hardware running FOSS?

Saturday, October 22, 2011

p7 vs GPU vs intel for HPC

http://domino.watson.ibm.com/library/CyberDig.nsf/papers/EFE521AB23A0D28B85257784004DC9DD/$File/rc25033.pdf

Just cant beat GPUs 4 bang for buck eh


...ask this guy http://youtu.be/hSVo4ejZ7rc

158000 GFLOPS for $420k

Wednesday, October 5, 2011

Perl program to work out what the golden screwdriver has done

Here is a perl script Ive written to go out to the IBM POD/MOD DB web site and get back the activations for your iSeries(AS400) or pSeries(RS6000) systems - just feed it the S/n like below
(you can then compare this to what you get in the HMC to see if ones missed)

$ ./activations.pl 9117 10 06ebb
02 processors activated on 05/23/2011
024 GB memory activated on 05/23/2011
02 processors activated on 03/11/2011
016 GB memory activated on 03/11/2011
055 GB memory activated on 03/02/2011
06 processors activated on 02/22/2011
096 GB memory activated on 02/22/2011
01 processors activated on 01/12/2011
016 GB memory activated on 01/12/2011
06 processors activated on 03/17/2010
072 GB memory activated on 03/17/2010
19 processors activated on 03/16/2010
228 GB memory activated on 03/16/2010
================
TOTAL CPU=36 RAM=519168MB (507GB)

$cat activations.pl
#!/usr/bin/perl
#
# PoC Script to go out on the net to the IBM POD site and tally up activations for a given i or p Series machine
#rutherfc-gmail.c
# v1.0 9/9/2011
#
# todo - add logic for processor deactivation - who uses that? IBM dont even publish the code on the pod site ;)
#
use LWP::UserAgent;
$ua = LWP::UserAgent->new;
$ua->agent("mozilla 8.0");
# $ua->proxy(['http'], 'http://proxy:8080/');
$ua->timeout(10);
use HTTP::Request::Common qw(POST);
if ($#ARGV != 2) {
print "usage: $0 MODEL XX XXXXX e.g $0 9119 83 9f6bf\n";
exit;
}
($model, $serial1, $serial2) = @ARGV;

##### main #####
get('http://www-912.ibm.com/pod/pod',"$serial2.htm");
html2txt("$serial2.htm","$serial2.txt");
total("$serial2.txt");
exit;

sub get # fakes a mozilla browser, fills in the CGI form and snags the returned page to a local html file
{
my $req = (POST $_[0],
["system_type" => $model,
"system_serial_number1" => $serial1,
"system_serial_number2" => $serial2 ]);
$request = $ua->request($req);
$activations = $request->content;
open(POD,">$_[1]");
print POD $activations;
close(POD);
}

sub html2txt # strips out the crap and converts the table to a local txt file to parse
{
open(HTML,"<$_[0]");
open(TXT,">$_[1]");
while () {
if (/<\/table>/) {$f = 0;};
if (/Posted Date \(MM/) {$f = 1;};
if ($f == 1) {
# poor mans HTML::TableExtract - excuse my sed like perl....
s//:/g;
s/<[^>][^>]*>//g;
s/ //g;
s/\n//g;
s/:/\n/g;
print TXT $_;
};
};
close(TXT);
close(HTML);
}
sub total # totals up the de/activations to get totals
{
open(TXT,"<$_[0]");
$[ = 1; $\ = "\n";# set array base & output record separator
while () {
($code,$hex,$date) = split(' ', $_, -1);
if (/POD/) {
$p = substr($hex, 27, 2);
print $p . ' processors activated on ' . $date;
$pt = $pt + $p;
};
if (/MOD/) {
$r = substr($hex, 26, 3);
print $r . ' GB memory activated on ' . $date;
$rt = $rt + $r;
};
if (/RMEM/) {
$r = substr($hex, 27, 2);
print $r . ' GB memory activated on ' . $date;
$rt = $rt - $r;
};
};
print '================';
print 'TOTAL CPU=' . $pt . ' RAM=' . $rt*1024 . 'MB (' . $rt . 'GB)';
close(TXT);
}

SAP on AIX6/AIX7/POWERHA/Oracle

Im just collecting these in one spot so I dont need to trawl SAP OSS/SDN again...

http://www.redbooks.ibm.com/Redbooks.nsf/RedpieceAbstracts/sg247564.html (SAP on per redbook)

SAP Notes regarding an AIX based SOE
http://service.sap.com/sap/support/notes/1048686 Recommended AIX settings for SAP
http://service.sap.com/sap/support/notes/1121904 SAP on AIX: Recommendations for paging space
http://service.sap.com/sap/support/notes/1278944 AIX C/C++ runtime requirements
http://service.sap.com/sap/support/notes/1131691 CPU Utilisation metrics of IBM System p
http://service.sap.com/sap/support/notes/1087498 Support for AIX 6.1
http://service.sap.com/sap/support/notes/973227 AIX Virtual Memory Management: Tuning

https://websmp130.sap-ag.de/sap/support/notes/1522360 FAQ for SAP HA installations using PowerHA (HACMP)
https://websmp130.sap-ag.de/sap/support/notes/1541935 Using SAP systems with AIX 7.1
https://websmp130.sap-ag.de/sap/support/notes/1458918 Support for AIX 7.1

Oracle RDBMS and IBM related

http://www.ibm.com/developerworks/aix/library/au-aix6tuning/index.html
http://www-03.ibm.com/support/techdocs/atsmastr.nsf/WebIndex/WP101621 IBM System Storage and SAP High Availability Installations on AIX
http://www.ibm.com/developerworks/wikis/display/WikiPtype/SAP+and+PowerHA
https://websmp130.sap-ag.de/sap/support/notes/1578479 SAP Support of Power Hardware with AIX
http://www-03.ibm.com/support/techdocs/atsmastr.nsf/WebIndex/TD101347 PowerHA compatibility Matrix
http://www-03.ibm.com/support/techdocs/atsmastr.nsf/WebIndex/WP101616 IBM Storage and SAP - evaluation of HA and DR concepts


AIO new method (AIX 6.1)
lpar9[/] > ioo -a | grep active
aio_active = 0
posix_aio_active =

there are no more AIO devices in the ODM.

two new parameters have been added to ioo: aio_active and posix_aix_active.
These can only be changed by AIX, and they are set to 1 only when AIO kernel extensors are used and pinned - you won't find any more AIO servers.
there are now aioLpools and aioPpools thesekernel processes manage AIO


and dont forget....
Language Translation to install →ISO8859-1 German [de_DE]
This installs the required bos.loc and bos.iconv file sets