Hi Manikumar,
I am interested to know the reason for exposing this config, given a
user has access to PrincipalBuilder interface to build their interpretation of
an identity from the X509 certificates. Is this to simplify extraction of
identity? and also there are other use cases where user's will extract
SubjectAltName to construct the identity I guess thats not going to supported
by this method.
Thanks,
Harsha
On Tue, Sep 18, 2018, at 8:25 AM, Manikumar wrote:
> Hi Rajini,
>
> I don't have strong reasons for rejecting Option 2. I just felt Option 1 is
> sufficient for
> the common use-cases (extracting single field, like CN etc..).
>
> We are open to go with Option 2, for more flexible mapping mechanism.
> Let us know, your preference.
>
> Thanks,
>
>
> On Tue, Sep 18, 2018 at 8:05 PM Rajini Sivaram <rajinisiva...@gmail.com>
> wrote:
>
> > Hi Manikumar,
> >
> > It wasn't entirely clear to me why Option 2 was rejected.
> >
> > On Tue, Sep 18, 2018 at 7:47 AM, Manikumar <manikumar.re...@gmail.com>
> > wrote:
> >
> > > Hi All,
> > >
> > > We would like to go with Option 1, which adds a new configuration
> > parameter
> > > pair of the form:
> > > ssl.principal.mapping.pattern, ssl.principal.mapping.value. This will
> > > fulfill the requirement for most of the common use cases.
> > >
> > > We would like to include the KIP in the upcoming release. If there no
> > > concerns, would like to start vote on this KIP.
> > >
> > > Thanks,
> > >
> > > On Fri, Sep 14, 2018 at 11:32 PM Priyank Shah <ps...@hortonworks.com>
> > > wrote:
> > >
> > > > Definitely a helpful change. +1 for Option 2.
> > > >
> > > > On 9/14/18, 10:52 AM, "Manikumar" <manikumar.re...@gmail.com> wrote:
> > > >
> > > > Hi Eno,
> > > >
> > > > Thanks for the review.
> > > >
> > > > Most often users want to extract one of the field (eg. CN). CN is
> > the
> > > > commonly used field.
> > > > For this simple change, users need to build and maintain the custom
> > > > principal builder class
> > > > and also package and deploy to the all brokers. Having configurable
> > > > rules
> > > > will be useful.
> > > >
> > > > Proposed mapping rules works on string representation of the X.500
> > > > distinguished name(RFC2253 format) [1].
> > > > Mapping rules can use the attribute types keywords defined in RFC
> > > 2253
> > > > (CN,
> > > > L, ST, O, OU, C, STREET, DC, UID).
> > > >
> > > > Any additional/custom attribute types are emitted as OIDs. To emit
> > > > additional attribute type keys,
> > > > we need to have OID -> attribute type keyword String mapping.[2]
> > > >
> > > > For example, String representation of X500Principal("CN=Duke,
> > > > OU=JavaSoft,
> > > > O=Sun Microsystems, C=US, EMAILADDRESS=t...@test.com")
> > > > will be "CN=Duke,OU=JavaSoft,O=Sun
> > > > Microsystems,C=US,1.2.840.113549.1.9.1=#
> > > 160d7465737440746573742e636f6d"
> > > >
> > > > If we have the OID - key mapping ("1.2.840.113549.1.9.1",
> > > > "emailAddress"),
> > > > the string will be
> > > > "CN=Duke,OU=JavaSoft,O=Sun Microsystems,C=US,emailAddress=
> > > > t...@test.com"
> > > >
> > > > Since we are not passing this mapping, we can not extarct using
> > > > additional
> > > > attribute type keyword string.
> > > > If the user want to extract additional attribute keys, we need to
> > > write
> > > > custom principal builder class.
> > > >
> > > > Hope the above helps. Update the KIP.
> > > >
> > > > [1]
> > > >
> > > > https://docs.oracle.com/javase/7/docs/api/javax/security/auth/x500/
> > > X500Principal.html#getName(java.lang.String)
> > > >
> > > > [2]
> > > >
> > > > https://docs.oracle.com/javase/7/docs/api/javax/security/auth/x500/
> > > X500Principal.html#getName(java.lang.String,%20java.util.Map)
> > > >
> > > > Thanks
> > > >
> > > > On Fri, Sep 14, 2018 at 7:44 PM Eno Thereska <
> > eno.there...@gmail.com
> > > >
> > > > wrote:
> > > >
> > > > > Manikumar, thanks. If I understand the KIP motivation right, this
> > > is
> > > > > currently already possible, but you want to have an easier way of
> > > > doing it,
> > > > > right? The motivation would be stronger if we had 2-3 common
> > cases
> > > > (from
> > > > > experience) where the suggested pattern would solve them, and
> > > > perhaps 1-2
> > > > > cases where the pattern would not be adequate and we'd have to
> > fall
> > > > back to
> > > > > the existing builder class.
> > > > >
> > > > > Thanks
> > > > > Eno
> > > > >
> > > > > On Fri, Sep 14, 2018 at 12:36 PM, Manikumar <
> > > > manikumar.re...@gmail.com>
> > > > > wrote:
> > > > >
> > > > > > Hi All,
> > > > > >
> > > > > > We'd appreciate any thoughts / comments on the proposed options
> > > for
> > > > > > customizing SSL principal names.
> > > > > > We are happy to discuss any alternative approaches/suggestions.
> > > > > >
> > > > > > Thanks,
> > > > > >
> > > > > > On Sat, Sep 8, 2018 at 12:45 AM Manikumar <
> > > > manikumar.re...@gmail.com>
> > > > > > wrote:
> > > > > >
> > > > > > > Hi all,
> > > > > > >
> > > > > > > I have created a KIP that proposes couple of options for
> > > building
> > > > > custom
> > > > > > > SSL principal names.
> > > > > > >
> > > > > > >
> > > > > > > https://cwiki.apache.org/confluence/display/KAFKA/KIP-
> > > > > > 371%3A+Add+a+configuration+to+build+custom+SSL+principal+name
> > > > > > >
> > > > > > > Please take a look.
> > > > > > >
> > > > > > > Thanks,
> > > > > > > Manikumar
> > > > > > >
> > > > > >
> > > > >
> > > >
> > > >
> > > >
> > >
> >
