2047 bit keys in PGP

What is the deal with the 2047 bit keys? How do you produce one? IS it
compatible with international versions?
--
==================Douglas Elznic===================
                 delznic@storm.net
           http://www.vcomm.net/~delznic/
            (315)682-5489 (315)682-1647
               4877 Firethorn Circle
                 Manlius, NY 13104
    "Challenge the system, question the rules."
===================================================
PGP key available:
http://www.vcomm.net/~delznic/pgpkey.asc
PGP Fingerprint:
 68 6F 89 F6 F0 58 AE 22  14 8A 31 2A E5 5C FD A5 
===================================================

--Boundary..3937.1071713532.multipart/signed
Content-Type: text/plain
Content-Transfer-Encoding: 7bit

"Douglas F. Elznic" <delznic@storm.net> said:

DE> What is the deal with the 2047 bit keys? How do you produce one? IS
DE> it compatible with international versions?

	When you do 'pgp -kg', you are asked to pick a PGP key size, and
given 3 preselected sizes.  You can select one of the sizes, or enter
your own choice.  Actually, the pgp source will allow 2048, but there is
a bug in the DOS version (from the compiler) that limits that one to
2047.  The international version is identical to the domestic one,
except for the RSA code, so everything is interoperable.  There is a
hacked version that allows 4096 bit keys, and the supersized keys are
incompatible.

-- 
#include <disclaimer.h>				/* Sten Drescher */
To get my PGP public key, send me email with your public key and
	Subject: PGP key exchange
Key fingerprint =  90 5F 1D FD A6 7C 84 5E  A9 D3 90 16 B2 44 C4 F3
Junk email is NOT appreciated.  If I want to buy something, I'll find
you.


--Boundary..3937.1071713532.multipart/signed
Content-Type: application/octet-stream; name="pgp00000.pgp"
Content-Transfer-Encoding: base64
Content-Disposition: attachment; filename="pgp00000.pgp"
Content-Description: "PGP signature"

LS0tLS1CRUdJTiBQR1AgTUVTU0FHRS0tLS0tClZlcnNpb246IDIuNi4yCgpp
UUNWQXdVQk1PckUyRnZCZ1JjZGtHd0ZBUUhCZkFRQWxlaXdBaURKbFhsbGlj
ckk1UmtmQ1VQd0gxT2xvYUFmCnRQRTRLRVNlL09UamE0VUo0SVVEQ3hhbkRv
TzZ2aDkraitacVNaR0l0TDdkaXUvbWtkMnZGMnVEQmpPay9ZS24KbXBZN0Zt
OHIzWTZjTFk0Y2NoOXJZV0YwUTE5UEptdEtpc1N4Z1RGcjZlRllSbGh5MDdV
SXVYRjZXTUNNZVgvcgp1eGdzZm9sVURJRT0KPVdIRGQKLS0tLS1FTkQgUEdQ
IE1FU1NBR0UtLS0tLQo=
--Boundary..3937.1071713532.multipart/signed--

> "Douglas F. Elznic" <delznic@storm.net> said:
> 
> DE> What is the deal with the 2047 bit keys? How do you produce one? IS
> DE> it compatible with international versions?
> 
> =09When you do 'pgp -kg', you are asked to pick a PGP key size, and
> given 3 preselected sizes.  You can select one of the sizes, or enter
> your own choice.  Actually, the pgp source will allow 2048, but there is
> a bug in the DOS version (from the compiler) that limits that one to
> 2047.  The international version is identical to the domestic one,
> except for the RSA code, so everything is interoperable.  There is a
> hacked version that allows 4096 bit keys, and the supersized keys are
> incompatible.

Are you sure it's a bug in the DOS version? When I did a pgp -kg in my 
UNIX shell (US version 2.6.2) I also entered 2048 bits and it too 
created a 2047 bit key instead.

Why is there a limit to the size of the key anyway? It's too bad PGP 
doesn't support any size key (within reason).

Laszlo Vecsey <master@internexus.net> said:

LV> Are you sure it's a bug in the DOS version? When I did a pgp -kg in
LV> my UNIX shell (US version 2.6.2) I also entered 2048 bits and it too
LV> created a 2047 bit key instead.

	I had heard elsewhere that there was such a bug.  My mistake,
then.

LV> Why is there a limit to the size of the key anyway? It's too bad PGP
LV> doesn't support any size key (within reason).

	As I understand it (which, given my previous error, is in
serious doubt), after a point the IDEA session keys become far easier to
use a brute force attack on than the RSA keypair.  Since I think that
increasing the RSA keysize is supposed to double the attack time, if a
RSA key size of N takes as much time to break as 1 IDEA key, making the
RSA key N+8 bits makes it better to break the IDEA keys of 200 messages
rather than the RSA key.

	Does anyone know if there are comparisons of estimates of the
time to break the IDEA session keys used in PGP vs time to break RSA
keys of various sizes?

-- 
#include <disclaimer.h>				/* Sten Drescher */
To get my PGP public key, send me email with your public key and
	Subject: PGP key exchange
Key fingerprint =  90 5F 1D FD A6 7C 84 5E  A9 D3 90 16 B2 44 C4 F3
Junk email is NOT appreciated.  If I want to buy something, I'll find
you.

--Boundary..3937.1071713532.multipart/signed
Content-Type: text/plain
Content-Transfer-Encoding: 7bit


(sorry if this is a duplicate)

Laszlo Vecsey <master@internexus.net> said:

LV> Are you sure it's a bug in the DOS version? When I did a pgp -kg in
LV> my UNIX shell (US version 2.6.2) I also entered 2048 bits and it too
LV> created a 2047 bit key instead.

	I had heard elsewhere that there was such a bug.  My mistake,
then.

LV> Why is there a limit to the size of the key anyway? It's too bad PGP
LV> doesn't support any size key (within reason).

	As I understand it (which, given my previous error, is in
serious doubt), after a point the IDEA session keys become far easier to
use a brute force attack on than the RSA keypair.  Since I think that
increasing the RSA keysize is supposed to double the attack time, if a
RSA key size of N takes as much time to break as 1 IDEA key, making the
RSA key N+8 bits makes it better to break the IDEA keys of 200 messages
rather than the RSA key.

	Does anyone know if there are comparisons of estimates of the
time to break the IDEA session keys used in PGP vs time to break RSA
keys of various sizes?

-- 
#include <disclaimer.h>				/* Sten Drescher */
To get my PGP public key, send me email with your public key and
	Subject: PGP key exchange
Key fingerprint =  90 5F 1D FD A6 7C 84 5E  A9 D3 90 16 B2 44 C4 F3
Junk email is NOT appreciated.  If I want to buy something, I'll find
you.


--Boundary..3937.1071713532.multipart/signed
Content-Type: application/octet-stream; name="pgp00001.pgp"
Content-Transfer-Encoding: base64
Content-Disposition: attachment; filename="pgp00001.pgp"
Content-Description: "PGP signature"

LS0tLS1CRUdJTiBQR1AgTUVTU0FHRS0tLS0tClZlcnNpb246IDIuNi4yCgpp
UUNWQXdVQk1PcnBZbHZCZ1JjZGtHd0ZBUUgvb0FRQXNyeTdPdjYyRzkzR1pC
UEU3Z3huSUw3VjBZeUxwbzFaCm5mTldobGlIelUyUzh5UTA0U0NaVjJ0dFFt
QjlWY0RROGFHOWVzMEZLeTVZQ2plTUZxdnE2Q2Z1OVQrV0ZFMDQKT1RQR2dI
dlVEdElZem9vU3NscjVsV1NhK0tsSU1GbHV4VWY2Zkk3TG9leThrdm1NY1RQ
elRjTDdzTU5LRTRMSQpzSi9tdEpxUUFmbz0KPTF2WkUKLS0tLS1FTkQgUEdQ
IE1FU1NBR0UtLS0tLQo=
--Boundary..3937.1071713532.multipart/signed--

-----BEGIN PGP SIGNED MESSAGE-----

Sten Drescher writes:
> Since I think that
> increasing the RSA keysize is supposed to double the attack time, if a
> RSA key size of N takes as much time to break as 1 IDEA key, making the
> RSA key N+8 bits makes it better to break the IDEA keys of 200 messages
> rather than the RSA key.
> 
> Does anyone know if there are comparisons of estimates of the
> time to break the IDEA session keys used in PGP vs time to break RSA
> keys of various sizes?

Off the top of my head, the figure I have usually heard quoted puts RSA at
about 100 times slower than your average symmetric key algorithm. So ignoring
key setup, I would expect an extra factor of 100 in the brute forcing time
for RSA over IDEA. 

I don't believe it's worth spending much time worrying about your RSA
key size. If you pick some decent size (1-2k), it's likely that RSA itself
will have been broken, or your key compromised by some other means, before
any direct brute force attack will succeed. 

Futplex <futplex@pseudonym.com>

-----BEGIN PGP SIGNATURE-----
Version: 2.6.2

iQEVAwUBMOsKNCnaAKQPVHDZAQHIRgf/ZS29BnGaZ60PeMlhIoniETAtI2VYNstM
yFV6tl5w1Kzu9Q2TcJk/tdpW9QVbWOrB2IMdELBrk1urcYBS6YUBXcAlI7UhinA9
sapoZpz3WUCnRdb/64HkGFsOYgEVyVjsrrmu+M2RUUNRnOwWSS0KFAz8GYqj83ry
xSpvrRNJPqCNARBsh9VPKgrRS1qNH5Zc1Tyu5Dr/E3OiQkzVCqHhQYYDj/PCESLL
Y1Sly6n133Jq8J3TWoXAzeNKAOwy4tLz6TFn63OgbfcnTp1hndsMlIwCN3tzn9el
T7b4LBMeVq2hXVkmotE0BURW7Phuckpmk1Xiow3vBXFMRxWPFz6lOg==
=Njig
-----END PGP SIGNATURE-----

-----BEGIN PGP SIGNED MESSAGE-----


In article <Pine.LNX.3.91.960103133137.3229F-100000@micro.internexus.net>
 master@internexus.net (Laszlo Vecsey) writes:

   Are you sure it's a bug in the DOS version? When I did a pgp -kg in my 
   UNIX shell (US version 2.6.2) I also entered 2048 bits and it too 
   created a 2047 bit key instead.

   Why is there a limit to the size of the key anyway? It's too bad PGP 
   doesn't support any size key (within reason).

Regarding the unix version, what sort of processor is the machine
running?  We noticed on a 486 running linux, with a vanilla MIT
release pgp, that it made 2047 bit keys when prompted for 2048.
However, after compiling the same code on a SunOS 4.1.3 it had no
problems making a 2048 key, though it took significantly longer due to
differences in the how the operating systems function.  Also, most of
the unix machines I work on now are running hacked versions which will
handle up to 4096 bit keys, so I could handle all of my keys
relatively painlessly (and because I was curious about the code in pgp
itself). 

In any case, that's about all I know on the topic, check the
architecture of the machine your shell account is on, and if you have
access to a sun, you might try it there.  If you really want larger
keys just poke around in the code, it's not that hard of a feat to
accomplish.  
- ----BEGIN PGP SIGNED HEADERS----
From: "Marc Martinez"   <lastxit@alphachannel.com>
To: cypherpunks@toad.com
Subject: Re: 2047 bit keys in PGP
Date: 03 Jan 1996 18:49:44 -0600
Message-ID: <wtbuokpyvr.fsf@arrakis.alphachannel.com>

-----BEGIN PGP SIGNATURE-----
Version: 2.6.2
Comment: Processed by Mailcrypt 3.4, an Emacs/PGP interface

iQDVAwUBMOskKHutHIVnrGOxAQGvIAYAwU7RJkBu33HXd7g1V1DtH7p8cXTwpral
lrYDCQDwRflxJUeNRAGUvL5cnfCGP0SGLcolkw4bsia92JtooyBrPhzNkAvh56O7
r9cXNb7EVnZIhEbgc5aVwa2BBSNgsbXNMYKhXmknrCkIUdBvIAf539xzkq5CXQQS
4ht8zhNku9UhAtuwNKa85zxUW+xmGdHX5kVn+aVAWUByxw5ndXq0aQkGFU7W9PVq
Vr/qLVrMheMMgIWw9w86ZQnz7UmWbWht
=vl/U
-----END PGP SIGNATURE-----

-- 
 Marc Martinez        "Sleep is unnecessary in the presence of more espresso."
 lastxit@mindport.net         Key fingerprint:   
 PGP public key available      47 AD 25 FF C2 B7 F8 57 C5 B6 2E B3 5E 98 A5 DE 
  by finger or keyserver

-----BEGIN PGP SIGNED MESSAGE-----

On Wed, 3 Jan 1996, Laszlo Vecsey wrote:

> > "Douglas F. Elznic" <delznic@storm.net> said:
> > 
> 
> Are you sure it's a bug in the DOS version? When I did a pgp -kg in my 
> UNIX shell (US version 2.6.2) I also entered 2048 bits and it too 
> created a 2047 bit key instead.

This is correct.  I believe there are some UNIX flavors under which U.S. PGP
can generate 2048 bit keys.  However, most only allow 2047 bit keys.  The
international version does not have this bug.

> 
> Why is there a limit to the size of the key anyway? It's too bad PGP 
> doesn't support any size key (within reason).

I really don't see the point of using a key larger than 2048 bits.  Any larger
key would actually be harder to factor than brute forcing the IDEA keyspace.
Very little security would be gained from using a key larger than 3000 bits.
Of course, one can always argue that improved factoring methods would require
that an RSA public key be longer than 3000 bits to have equal security to
IDEA.  However, I doubt that factoring methods will improve that much.  A
2048 bit key should be more than enough security for most applications.



-----BEGIN PGP SIGNATURE-----
Version: 2.6.2

iQCVAwUBMOrkUbZc+sv5siulAQFWJgP+IlKURN3TtdXrqzLf3vCrva1tYkYC/lZU
fIOlk5Cvnt9wpm/huZKu/nESvFmJutoTbZVvJz1EPglLc1YrAlo4xyWTJZgwMpgv
khXzkEMaPludU1qfKowaM0qqeSHv80zSB97Mq0SbqNEPyM2K0r+gDobSjUgwKQCQ
Mb5D9L3hTLA=
=CDHg
-----END PGP SIGNATURE-----


finger -l markm@voicenet.com for PGP key  http://www.voicenet.com/~markm/
Fingerprint: bd24d08e3cbb53472054fa56002258d5  Key-ID: 0xF9B22BA5
-----BEGIN GEEK CODE BLOCK-----
Version: 3.1
GAT d- s:- a? C++++ U+++>$ P+++ L++(+++) E--- W++(--) N+++ o- K
w--- O- M- V-- PS+++>$ PE-(++) Y++ PGP+(++) t-@ 5? X++ R-- tv+
b+++ DI+ D++ G+++ e! h* r! y?
------END GEEK CODE BLOCK------