[dns-operations] Summary: Anyone still using a Sun/Oracle SCA6000 with OpenSSL?

[Richard Lamb]

Why not the tpm migration method? I. E.


The receiving hsm produces the public half of a master storage key.  
Then the starting hsm "authorizes" the key for use for exporting with pomp and circumstance ;-)
Then the starting hsm encrypts it's keys with this key (rsa) for transfer to the receiving hsm.
Receiving hsm unwraps the key using its private key.
Done


Sent from my iPhone

On Oct 14, 2012, at 18:36, "George Michaelson" <ggm at apnic.net> wrote:

> 
> On 15/10/2012, at 3:10 AM, Ondřej Surý <ondrej.sury at nic.cz> wrote:
> 
>> Just a question - would anyone would be interested in joining a project to build an OpenHardware FPGA-based HSM with focus on DNSSEC?
>> 
>> O.
>> 
> 
> APNIC has no skills in FPGA level design and construction, but I would be very interested in participating in the specification of the user space requirements for this work.
> 
> I'm particularly interested in its ability to support a key migration mechanism which would prevent capture of the signing materials by a single implementation. We're finding that the equipment we have now, supports simple mechanisms within the single vendor chain (so you can upgrade) but has no good inter-provider key exchange mechanism, and we've had a similar experience with other (DNSSEC) solutions.
> 
> A recent conversation suggests that there are no good standards in this space, and that a public-private key exchange between the HSM is the way to go, possibly augmented by a shared-secret generated on-box and shared between them.
> 
> -G
> 
> 
>> On 16. 8. 2012, at 2:24, George Michaelson <ggm at apnic.net> wrote:
>> 
>>> 
>>> I got 8 replies. 2 ccTLD, 2 root Ops, almost everyone in s/w development or operational related roles, and some independent consultants.
>>> 
>>> Only one happy user, and I'd qualify that: they'd want a longterm migration plan off the device. This person is using Solaris.
>>> 
>>> Everyone said avoid more than 255 keys on the device. Several said use the import/export mechanism.
>>> 
>>> Two people explicitly mentioned the bad Linux driver. 
>>> 
>>> The overall tone of the (small sample) responses is: "this is not a good choice right now"
>>> 
>>> 
>>> My context is not DNSSEC, its RPKI, which has a far larger keypair requirement. Noting a suggestion to re-use keypairs, I'd still have to risk-manage future potential for multiple keys per hosted client, and exceed the on-card keystore size, so the suggestion to use the import/export features makes sense. Having said that, documentation on this is really scant, and its hard to confirm how easily you can manage this given there is no explicit OpenSSL PKCS11 support for managing PKCS12 wrapped objects, and you are therefore using a java or shell command to do the key import, followed by OpenSSL engine, followed by shell/java to remove the key. 
>>> 
>>> If you use a pure Java solution its probably more tenable.
>>> 
>>> Thank you to everyone for the response. I hope this summary meets a sense of privacy, and OT posting.
>>> 
>>> -G
>>> _______________________________________________
>>> dns-operations mailing list
>>> dns-operations at lists.dns-oarc.net
>>> https://lists.dns-oarc.net/mailman/listinfo/dns-operations
>>> dns-jobs mailing list
>>> https://lists.dns-oarc.net/mailman/listinfo/dns-jobs
>> 
>> --
>> Ondřej Surý -- Chief Science Officer
>> -------------------------------------------
>> CZ.NIC, z.s.p.o.    --    Laboratoře CZ.NIC
>> Americka 23, 120 00 Praha 2, Czech Republic
>> mailto:ondrej.sury at nic.cz    http://nic.cz/
>> tel:+420.222745110       fax:+420.222745112
>> -------------------------------------------
>> 
> 
> _______________________________________________
> dns-operations mailing list
> dns-operations at lists.dns-oarc.net
> https://lists.dns-oarc.net/mailman/listinfo/dns-operations
> dns-jobs mailing list
> https://lists.dns-oarc.net/mailman/listinfo/dns-jobs