On Wed, Jun 6, 2018 at 1:05 PM, Shazni Nazeer <[email protected]> wrote:

> Currently we persist oidc related scopes and claims in the registry. With
>> this approach we need to access the registry in run time, which is an
>> anti-pattern. So going forward we have decided to persist oidc scopes and
>> claims in the db and remove from the registry.
>
>
> I'm sure there are good reasons to move the claim mappings from the
> registry to plain DB. But isn't this essentially the same as ultimately
> registry stores values in the underlying DB. I'm sure the change may make
> the implementation straightforward to use the DB query instead of the
> registry API and also makes way to move away from registry. But why using
> registry is an anti-pattern?
>

1. Registry is not designed to cater runtime traffic, it was originally
designed to cater non-runtime (design time) traffic, so by design there is
a scalability issue when we use registry in run-time use cases. Few years
back we have experienced this when trying to persist AS and ESB runtime
metafiles in the registry.

2. We are in a process  of moving out registry dependencies from all
products except G-Reg because most of the cases it's unnecessary , API-M 3
did this up to some extend. In this case we have noticed a scalability
issue in userinfo calls and as a part of the fix we are trying to remove
one registry dependency from IS.

Thanks !

>
> On Wed, Jun 6, 2018 at 11:48 AM, Hasanthi Purnima Dissanayake <
> [email protected]> wrote:
>
>> Hi All,
>>
>> Currently we persist oidc related scopes and claims in the registry. With
>> this approach we need to access the registry in run time, which is an
>> anti-pattern. So going forward we have decided to persist oidc scopes and
>> claims in the db and remove from the registry.
>>
>> *With the new implementation:*
>> 1.  In the management console 'Resident Identity Provider > Inbound
>> Authentication Configuration > OAuth2/OpenidConnect configuration'  will be
>> divided in to two sections.
>>    a. First section will include the existing EP URLs
>>    b. Second section will include scope claim table which have the
>> ability to add and delete scope claim mapping.
>> 2. In the first server start up the scopes and claims defined
>> in oidc-scope-config.xml will be stored in the db and a caching layer.
>> 3. So when the UI is loading the scopes and claims that are stored in the
>> table will be populated to the UI as well.
>>
>> I will update the thread with the screen shots of the new UI and the
>> design of the new table soon. Highly appreciate any suggestions or
>> feedbacks on this.
>>
>> Thanks,
>>
>> --
>>
>> Hasanthi Dissanayake
>>
>> Senior Software Engineer | WSO2
>>
>> E: [email protected]
>> M :0718407133| http://wso2.com <http://wso2.com/>
>>
>
>
>
> --
> Shazni Nazeer
>
> Mob : +94 777737331
> LinkedIn : http://lk.linkedin.com/in/shazninazeer
>
> Blogs :
>
> https://medium.com/@mshazninazeer
> http://shazninazeer.blogspot.com
>
> <http://wso2.com/signature>
>



-- 
Sagara Gunathunga

Director; WSO2, Inc.;  http://wso2.com
Linkedin; http://www.linkedin.com/in/ssagara
Blog ;  http://ssagara.blogspot.com
Mobile : +9471 <+94%2071%20565%209887>2149951
_______________________________________________
Architecture mailing list
[email protected]
https://mail.wso2.org/cgi-bin/mailman/listinfo/architecture

Reply via email to