Gwen,
There is one product called ElasticSearch which has been quite
successful. They recently added security, what they actually did is quite
nice. They really separated Authentication and Authorization which many
people get really confused about and often mix them up. I looked through
what they did and quite impressed by it, I think there are many things we
can borrow from. Here is a link to it.
http://www.elastic.co/guide/en/shield/current/architecture.html. The
product name is called "shield" which is implemented as an ElasticSearch
plugin. The promise here is that you can have a running ElasticSearch, then
you install this plugin, configure it, then your ElasticSearch service is
secured. The goal should be really the same for Kafka, you have a Kafka
service running, you install a new plugin (in this case security plugin),
configure it, then your Kafka service is secured. I think that the key here
is that we should introduce a true pluggable framework in Kafka which
allows security, quota, encryption, compression,
serialization/deserialization all being developed as plugins which can be
all easily added and configured onto a running Kafka service, then the
functions/features provided by the plugins will start working. Once we have
this framework in, how a security plugin works internally becomes the
really the concern of that plugin, for example, how a new user gets
registered, permission granted, revoked, all these will be the concern of
that plugin, rest of the Kafka components should not really be concerned
about them. This way we are really following the design principal
(Separation of concerns). With all that, what I am proposing is a true
pluggable framework introduction into Kafka which I have also talked about
in a previous email. For security we can implement a simple file based
security plugin, other plugins such as LDAP, AD for authentication can come
later, plugin for authorization such as RBAC can also come later if people
care so much about using them.
Thanks.
Tong Li
OpenStack & Kafka Community Development
Building 501/B205
liton...@us.ibm.com
From: Gwen Shapira <gshap...@cloudera.com>
To: "dev@kafka.apache.org" <dev@kafka.apache.org>
Date: 04/16/2015 12:44 PM
Subject: [DISCUSSION] KIP-11: ACL Management
Hi Kafka Authorization Fans,
I'm starting a new thread on a specific sub-topic of KIP-11, since
this is a bit long :)
Currently KIP-11, as I understand it, proposes:
* Authorizers are pluggable, with Kafka providing DefaultAuthorizer.
* Kafka tools allow adding / managing ACLs.
* Those ACLs are stored in ZK and cached in a new TopicCache
* Authorizers can either use the ACLs defined and stored in Kafka, or
define and use their own.
I am concerned of two possible issues with this design:
1. Separation of concerns - only authorizers should worry about ACLs,
and therefore the less code for ACLs that exist in Kafka core, the
better.
2. User confusion - It sounded like we can define ACLs in Kafka itself
but authorizers can also define their own, so "kafka-topics
--describe" may show an ACL different than the one in use. This can be
super confusing for admins.
My alternative suggestion:
* Authorizer API will include:
grantPrivilege(List<Principals>, List<Privilege>)
revokePrivilege(List<Principals>, List<Privilege>),
getPrivilegesByPrincipal(Principal, Resource)
....
(The exact API can be discussed in detail, but you get the idea)
* Kafka tools will simply invoke these APIs when topics are added /
modified / described.
* Each authorizer (including the default one) will be responsible for
storing, caching and using those ACLs.
This way, we keep almost all ACL code with the Authorizer, where it
belongs and users get a nice unified interface that reflects what is
actually getting used in the system.
This is pretty much how Sqoop and Hive implement their authorization APIs.
What do you think?
Gwen
