cloudstack-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Chiradeep Vittal <Chiradeep.Vit...@citrix.com>
Subject Re: API refactoring work for 4.1.0
Date Tue, 04 Dec 2012 18:39:00 GMT
There should be a facility to locate the API under a specific namespace.
E.g., 
compute: <server>:8080/compute/
loadbalancer: <server>:8080/lb/
volume: <server>:8080/volume/


On 12/4/12 2:19 AM, "Rohit Yadav" <rohit.yadav@citrix.com> wrote:

>Hi everyone,
>
>The API refactoring work for 4.1.0 is going on in api_refactoring branch:
>https://git-wip-us.apache.org/repos/asf?p=incubator-cloudstack.git;a=short
>log;h=refs/heads/api_refactoring
>So far I've broken the user level APIs (build system is broken on the
>branch), admin level API is wip.
>
>API refactoring plan and fs:
>https://cwiki.apache.org/confluence/display/CLOUDSTACK/CloudStack+API+refa
>ctoring
>
>Annotation refactoring:
>https://issues.apache.org/jira/browse/CLOUDSTACK-518
>
>The aim is to fix api classification, grouping and eventually run as a
>separate server which would communicate with the data center abstraction
>engine called cloud-engine over its REST end point.
>
>For those who've attended Alex's talk on architectural rework, this
>rework would make it possible to have:
>
>- Separate end points for sysadmins and role based users.
>- Blacklisting of apis etc, creation of different kinds of roles (system
>admins could run an api-server with a set of apis they wish to offer over
>an endpoint by adding/removing from commands.properties.in).
>- Separate out api server from mgmt server, so we move towards our goal
>to have a cloud core; everything would talk to the scalable cloud-engine.
>- Grouping as per security roles (org.apache.cloudstack.api.{user, admin})
>- Multiple api servers, with load balancing etc. which would make it very
>scalable.
>- Renaming package names to org.apache.cloudstack.* namespace.
>- A scalable queue (considering apache compliant Quartz)
>- Annotations to have automated docs with better notion of linking
>entities an arg to an API, separate out security layer etc. Right now the
>API layer and mgmt server (cloud-server, cloud-core) is tightly coupled.
>- Auto-discovery and caching of APIs for a users using a helper plugin
>for cloudmonkey cli, right now the API info is precached in cloudmonkey
>during build time.
>
>Please go through the api fs proposal. Suggestions, proposals, feedback,
>flames?
>
>Regards.
>
>Ref. rfc email:
>http://mail-archives.apache.org/mod_mbox/incubator-cloudstack-dev/201211.m
>box/%3CB1DF26ECC0458748AC97CECE2DA98D41012FA1E1CAFD@SJCPMAILBOX01.citrite.
>net%3E


Mime
View raw message