Re: Algorithm suggestions


"Geoffrey Summerhayes" <sumrnot@xxxxxxxxxxx> wrote in
message
news:1159294685.907516.118210@xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx
I'm looking for a relatively simple algorithm to add a
second layer of
protection to a transmitted packet.

The data to be encrypted is less than 256 bytes per
packet, with a
fair amount of similarity in the data, multiple senders,
each can have
their own key, the final packet is sent using SSL.

The main reason for the additional layer is to deter
packet sniffing
between the sender and the SSL device.

The units are fairly limited, it would be nice if the same
function
worked for the encode and decode. Any suggestions?


Yes. You need to think about what you really want from this
security and what you are willing to give up to get it. For
example:

1. Does your attack model envision anything except
snooping? For example, will your attacker try to become an
active man-in-the-middle, seeking to replace all or part of
one device's data with data of his own choosing? Will your
attacker seek to corrupt portions of your data packet
(without necessarily knowing what the original content was
or what the decrypted content will be)?

2. How valuable is the data that the attacker could obtain
by snooping? For how long will it maintain its value? How
much could it cost if an attacker could substitute
meaningful data into the data packet undetected?

3. How limited is "fairly limited"? Do the devices support
TCP/IP stacks? Do they run embedded Windows or Linux? How
much CPU power do they have and how much of that is
available? What cryptographic support is available from the
O/S or your programming language? Will your app be
implemented in Java using a current Sun JDK?

Rolling your own "simple algorithm to add a second layer of
protection" will be, at best, a waste of time. At worst it
will engender a false sense of security in those who are
depending on true security and set the stage for a rude
awakening in the future.

If you must implement at the nuts-and-bolts level in C, use
a portable, well-tested library like libtomcrypt. The
advantages of this are too numerous to list, the most
important being that you can get support through this
newsgroup if you need it. If you are working in Java, use
the Sun JDK's Crypto API. Either way, you can at least be
sure you're using a reputable cipher, even if you might not
be using it in the most secure way possible.



.

Relevant Pages

  • Re: Smurf ,land attacks
    ... Using libnet/libpcap you can write directly to the ... > What the attacker does is not allowing the Kernel to ... > from the packet he's spoofing, ... >> wire? ...
    (Security-Basics)
  • Re: Algorithm suggestions
    ... protection to a transmitted packet. ... the final packet is sent using SSL. ... This is not a "second layer of protection", ...
    (sci.crypt)
  • [UNIX] Security Analysis of VTun
    ... The following security advisory is sent to the securiteam mailing list, and can be found at the SecuriTeam web site: http://www.securiteam.com ... An attacker can modify ... Packet forwarding: ... password) as encryption key. ...
    (Securiteam)
  • [Full-disclosure] Making unidirectional VLAN and PVLAN jumping bidirectional
    ... Wepwedgie, a tool by Anton Rager for traffic injection on 802.11 networks protected by WEP, solves the problem of unidirectional communication by bouncing packets from the target host to a third external host under the attackers control. ... We employ exactly the same principle to bypass both VLAN and PVLAN network segmentation. ... The attacker tags his malicious data with two 802.1q tags and sends the packet with a spoofed source IP of a host under his or her control. ...
    (Full-Disclosure)
  • Making unidirectional VLAN and PVLAN jumping bidirectional
    ... Wepwedgie, a tool by Anton Rager for traffic injection on 802.11 networks protected by WEP, solves the problem of unidirectional communication by bouncing packets from the target host to a third external host under the attackers control. ... We employ exactly the same principle to bypass both VLAN and PVLAN network segmentation. ... The attacker tags his malicious data with two 802.1q tags and sends the packet with a spoofed source IP of a host under his or her control. ...
    (Bugtraq)