| Author |
Message
|
| PeterPotkay |
 Posted: Thu Oct 23, 2014 3:13 am Post subject: |
 |
|
Poobah
Joined: 15 May 2001
Posts: 7722
|
| tczielke wrote: |
So it does look like there are MQ SSL v3 ciphers (i.e RC4_MD5_US) that are not susceptible to POODLE.
|
I would not use any Cipher based on MD5.
http://en.wikipedia.org/wiki/MD5
| Quote: |
| In 2004 it was shown that MD5 is not collision resistant.[5] As such, MD5 is not suitable for applications like SSL certificates or digital signatures that rely on this property for digital security. Also in 2004 more serious flaws were discovered in MD5, making further use of the algorithm for security purposes questionable; |
_________________
Peter Potkay
Keep Calm and MQ On |
|
| Back to top |
|
 |
| tczielke |
| Posted: Thu Oct 23, 2014 4:50 am Post subject: |
|
|
Guardian
Joined: 08 Jul 2010
Posts: 941
Location: Illinois, USA
|
The two articles I included also talked disparagingly about RC4.
I am definitely not trying to endorse RC4_MD5_US, or refute the overall recommendation of the IBM MQ security bulletin on POODLE to move from SSL v3 to TLS ciphers.
However, it does seem that POODLE itself is a vulnerability against SSL v3 CBC based ciphers. So POODLE itself would not require a remediation of moving from RC4_MD5_US. If my understanding is incorrect, I would appreciate if someone would correct it. |
|
| Back to top |
|
|
| mqjeff |
| Posted: Thu Oct 23, 2014 7:13 am Post subject: |
|
|
Grand Master
Joined: 25 Jun 2008
Posts: 17447
|
My understanding is that POODLE addresses a vulnerability in SSLv3 itself, not necessarily one with CBCs.
If that's the case, the cipherspec wouldn't matter.
But I could be wrong on that. |
|
| Back to top |
|
|
| tczielke |
| Posted: Thu Oct 23, 2014 7:19 am Post subject: |
|
|
Guardian
Joined: 08 Jul 2010
Posts: 941
Location: Illinois, USA
|
| mqjeff wrote: |
My understanding is that POODLE addresses a vulnerability in SSLv3 itself, not necessarily one with CBCs.
If that's the case, the cipherspec wouldn't matter.
But I could be wrong on that. |
That was my original understanding based on the MQ POODLE security bulletin, too. However, after talking to one of our web server admins, he mentioned that he had heard that POODLE was actually a vulnerability against CBC based SSL v3 ciphers. I did a little research on the web to get other sources that seemed to support for that assertion. So just curious if anyone had a definitive answer on that topic. |
|
| Back to top |
|
|
| tczielke |
| Posted: Fri Oct 24, 2014 4:39 am Post subject: |
|
|
Guardian
Joined: 08 Jul 2010
Posts: 941
Location: Illinois, USA
|
I did open a PMR to get some clarity to my question on if a SSL v3 cipher like RC4_MD5_US (which I think is non-CBC) is susceptible to POODLE. This was the response.
“The vulnerability is specifically within the SSLv3 protocol, thus that protocol must not be used if the customer wants to close their exposure to the vulnerability.” |
|
| Back to top |
|
|
| PeterPotkay |
| Posted: Fri Oct 24, 2014 5:41 am Post subject: |
|
|
Poobah
Joined: 15 May 2001
Posts: 7722
|
| tczielke wrote: |
I did open a PMR to get some clarity to my question on if a SSL v3 cipher like RC4_MD5_US (which I think is non-CBC) is susceptible to POODLE. This was the response.
“The vulnerability is specifically within the SSLv3 protocol, thus that protocol must not be used if the customer wants to close their exposure to the vulnerability.” |
I would be more confident in that PMR response if they went on to clarify whether non-CBC is susceptible or not. To be fair, that is probably more detailed then they need or want to be on this topic - just get off any and all SSL and let's move on is the thinking of most. The question would be better directed at an SSL expert...curious minds want to know.
_________________
Peter Potkay
Keep Calm and MQ On |
|
| Back to top |
|
|
| mqjeff |
| Posted: Fri Oct 24, 2014 5:45 am Post subject: |
|
|
Grand Master
Joined: 25 Jun 2008
Posts: 17447
|
| PeterPotkay wrote: |
| tczielke wrote: |
I did open a PMR to get some clarity to my question on if a SSL v3 cipher like RC4_MD5_US (which I think is non-CBC) is susceptible to POODLE. This was the response.
“The vulnerability is specifically within the SSLv3 protocol, thus that protocol must not be used if the customer wants to close their exposure to the vulnerability.” |
I would be more confident in that PMR response if they went on to clarify whether non-CBC is susceptible or not. To be fair, that is probably more detailed then they need or want to be on this topic - just get off any and all SSL and let's move on is the thinking of most. The question would be better directed at an SSL expert...curious minds want to know. |
"Specifically within the SSLv3 Protocol" means "it doesn't matter what cipher you use, the exchange of ciphers and the securing of the channel itself is vulnerable".
That is, that some part of the handshaking/identifying process is vulnerable to hacking. |
|
| Back to top |
|
|
| PeterPotkay |
| Posted: Mon Oct 27, 2014 3:57 am Post subject: |
|
|
Poobah
Joined: 15 May 2001
Posts: 7722
|
|
| Back to top |
|
|
| samsansam |
| Posted: Thu Oct 30, 2014 8:08 pm Post subject: |
|
|
Apprentice
Joined: 19 Mar 2014
Posts: 41
|
so if we are not using SSL at all , do we have to care about POODLE?
We have MQ 7.0.1 |
|
| Back to top |
|
|
| fjb_saper |
| Posted: Thu Oct 30, 2014 9:01 pm Post subject: |
|
|
Grand High Poobah
Joined: 18 Nov 2003
Posts: 20756
Location: LI,NY
|
| samsansam wrote: |
so if we are not using SSL at all , do we have to care about POODLE?
We have MQ 7.0.1 |
So I figure you're not using any server connection channels without security exits either right? 
_________________
MQ & Broker admin |
|
| Back to top |
|
|
| samsansam |
| Posted: Thu Oct 30, 2014 9:47 pm Post subject: |
|
|
Apprentice
Joined: 19 Mar 2014
Posts: 41
|
Sound not secure , but we do use server connection channels without security exits.
I know how sound stupid that , but I just join the company |
|
| Back to top |
|
|
| mqjeff |
| Posted: Fri Oct 31, 2014 4:44 am Post subject: |
|
|
Grand Master
Joined: 25 Jun 2008
Posts: 17447
|
SSL in MQ serves two purposes. To allow you to verify the identity of the remote end, and to ensure that traffic moving across the wire can not be read.
If you only need to do the first thing, and not the second thing, then channel exits are just fine and count as "secure" under your local criteria.
If you are not using SSL at all, I am not sure if your queue managers are vulnerable to POODLE or not. I don't know what the queue manager would do if it has SSL as OPTIONAL and someone makes an SSLV3 connection. |
|
| Back to top |
|
|
| smdavies99 |
| Posted: Fri Oct 31, 2014 10:20 am Post subject: |
|
|
Jedi Council
Joined: 10 Feb 2003
Posts: 6076
Location: Somewhere over the Rainbow this side of Never-never land.
|
| samsansam wrote: |
Sound not secure , but we do use server connection channels without security exits.
I know how sound stupid that , but I just join the company |
What if your network is triple firewalled from the Internet plus each subnet is firewalled from every other subnet AND only the required ports are open AND the queue managers are configured to accept incoming SVRCONN connetions from know IP Addresses. ***
Plus NO personal or Financial data is carried.
Do you use SSL(TLS) + Exits or leave the data open and unencrypted?
Is the QMGR Secure or insecure?
What I'm trying to say is that each and every system is different from every other and needs to be judged accordingly.
***
That said we have a customer asking for hardware encryption everywhere in the plant including every sensor device. They even want to encrypt SNMP messages. It will add at least $5M to the project price (if it could be done at all). So far they have not baulked at the cost because they have oodles of oil money to spend.
_________________
WMQ User since 1999
MQSI/WBI/WMB/'Thingy' User since 2002
Linux user since 1995
Every time you reinvent the wheel the more square it gets (anon). If in doubt think and investigate before you ask silly questions. |
|
| Back to top |
|
|
|
|
