[
Date Prev][
Date Next][
Thread Prev][
Thread Next][
Date Index][
Thread Index]
[
List Home]
Re: [stellation-res] why doesn't stellationd require a "password" argument?
|
On Mon, Jul 22, 2002 at 09:19:41AM -0400, Mark C. Chu-Carroll wrote:
> We already have a pretty decent authentication framework. I'd be
> willing to support changes to the auth layer if they had a real
> benefit. But I'm not willing to switch just for the sake of switching.
>
> Our current authentication system supports pluggable authentication
> modules, challenge authentications, etc. It's really pretty decent.
>
> The main weakness of it, at the moment, is that we're not doing anything
> clever to protect user passwords in the database - so protecting the
> security of the database becomes important, and we're not doing a
> great job of that yet. But that's not a huge deal to fix: the database
> should be run in a protected mode, so that you need a password
> to establish a database connection; the the password storage
> should be changed to not use plaintext; and the encryption key
> to decode the stored passwords should only be accessible
> using an administators key. None of these are hard to do.
>
> Switching to JAAS might make sense, because it's a standard
> mechanism.
The benefits being that there might be more people understanding it
and we get somebody else to review and maintain it for free.
> If it could be done without adding a huge amount
> of complexity to the system, I'd like to see it happen.
As I see it now it will remove more code than it adds.
For a small setup, sending passwords around is OK. If they want more
security than that, they can set up kerberos and use its ticket-granting
capabilities.
> On the
> other hand, adding another non-standard dependency to the system
> for an auth layer is something that we'd need to carefully consider:
> what benefits does Turbine have over what we're using? If they're
> significant, and much easier than JAAS, then great. If not, then
> then I'd rather not.
florin
--
"If it's not broken, let's fix it till it is."
41A9 2BDE 8E11 F1C5 87A6 03EE 34B3 E075 3B90 DFE4
Attachment:
pgpcogi3iE1Mh.pgp
Description: PGP signature