Hey,

Yes, the communication is using UDP port 1514 (by default). This is
how it works:

-Network level:
1-Agent connects to the manager port 1514

2-Manager acknowledges the connection by replying back to the agent.

31-Agent sends the events to the manager as they are read locally.

32-Agent sends a keep alive every 5 minutes and waits for the manager to reply.
321-Manager acknowledges the keep alive.
322-If the manager is offline (and don't answer the keep alive), the
agent will retry it a few
times. If no reply is received, it will enter the offline mode, stop
processing events and keep
trying to re-connect to the manager.

4- If the manager doesn't receive the keep alive from the agent, it
will alert that the
agent is offline (after a wait period to give the agent a chance to reconnect).


-Low level

1- Agent reads the key from client.keys and the message count from the
rids directory.

2- When sending the event, the agent increases the rids count, adds a random pad
and prepends the event to the final string. It encrypts it all using
blowfish and sends to
the manager.

3- The manager checks if the source IP address matches the configured IP address
for that agent and decrypts the message using the symmetric key.

4- If the rids count is repeated, it drops the message (injection
attack or network problem).


I hope that helps to summarize how it works...

Thanks,

--
Daniel B. Cid
dcid ( at ) ossec.net

On Fri, Feb 26, 2010 at 2:01 PM, Wim Remes <[email protected]> wrote:
> Hi,
>
> I might be off but if I remember correctly OSSEC uses blowfish crypto.  The 
> key is based (amongst others) on the ip address of the client.
>
> Cheers,
>
> Wim
>
> On 26 Feb 2010, at 16:07, HotteFred wrote:
>
>>
>>
>> Greetings everyone,
>>
>> Could someone tell me what kind of crypto and authentication ossec
>> uses for thesecureconnectionfrom clients to the server?  We're
>> considering making some adjustments to our network for reliability
>> purposes and we need to know how exactly ossec handles the
>> communication channel.  I haven't seen any documentation on this,
>> although maybe I missed it.
>>
>> Full disclosure: No, I admit I haven't read the source - C and I are
>> not on good terms.
>>
>> --cryogen
>>
>> ------------------------------------------------------------------------
>>
>> Im interested too. Is there anyone, who can tell me more about the
>> secure connection between client server? How does it works?
>>
>> Thanks
>> Christian
>
>

Reply via email to