Open Stack

On Tue, Apr 29, 2014 at 12:42 PM, Nachi Ueno <nachi at ntti3.com> wrote:
> Hi Zang
>
> Thank you for your contribution on this!
> The private key management is what I want to discuss in the summit.
>
Has the idea of using Barbican been discussed before? There are many
reasons why using Barbican for this may be better than developing key
management ourselves.

Thanks,
Kyle

> [1] We are depending DB security, anyway
> When we get stolen the private key in the DB, it means we are also
> stolen ID/PW for DB.
> If we stolen the key, even if we keep the private key secret, the
> attacker can connect the NW for anywhere.
>
> [2] How we manage a passcode for encrypting private key?
> so even if openvpn supports encripted keys, when we input the passcode?
> Vpn process will be launched automatically by neutron-server, so we
> need to store it in the memory.
> This is same security with plain private key.
> For example, most of apache servers using plain private key, I guess.
>
> so the security of ssl-vpn impl depends on db, rpc trasport, file
> system security even if we encrypt the private key or not.
> may be, we have better way, but I think current design isn't so bad to
> prevent get merged.
>
> Best
> Nachi
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
> 2014-04-28 23:02 GMT-07:00 Zang MingJie <zealot0630 at gmail.com>:
>> Hi all:
>>
>> Currently I'm working on ssl vpn, based on patchsets by Nachi[1] and Rajesh[2]
>>
>> There are secure issues pointed by mark, that ssl private keys are
>> stored plain in database and in config files of vpn-agents. As
>> Barbican is incubated, we can store certs and their private keys in
>> Barbican. But after checking openvpn configurations, I don't think
>> there is any way to prevent storing private key in openvpn config
>> files without modify the openvpn implementation.
>>
>> I have also made several changes, added a optional port field to
>> sslvpn-connection table, integrated with service plugin framework
>> (I'll follow service flavor framework when it is ready), and completed
>> the neutronclient part. It is already developed in our testing
>> environment, I'll upload my patch sooner or later.
>>
>> [1] https://review.openstack.org/#/c/58897/
>> [2] https://review.openstack.org/#/c/70274/
>>
>> _______________________________________________
>> OpenStack-dev mailing list
>> OpenStack-dev at lists.openstack.org
>> http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev
>
> _______________________________________________
> OpenStack-dev mailing list
> OpenStack-dev at lists.openstack.org
> http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev