Custom ACL for Row-Level Security - acl

I have seen a few implementation of Row-Level ACL using a Permission table having a structure such as
User_Id
Subject_Class
Subject_Id
Permission_Id
where Permission_Id is (Read, Write, Update, Delete, Approve, etc.)
I was wondering if there is any benefit to describe a relationship (Relationship_Id) with the data instead of describing a permission.
The idea we would to describe that a user is an "Owner", "Approver", "Reviewer", "Public Viewer", etc.
This relationship would then define a set of permissions. This could reduce the size of the Permission.
Any thoughts on this methodology?

Roles, as you note, are combinations of permissions. If you have a permission set of size n, you can have 2^n combinations of all of these permissions. This means that if you have 5 permissions, you'd need 32 roles. If you have 10 permissions, you'd have 1024 roles!
Using the permission route also appears to be easier for maintainability - say that you want to remove access permissions from a group of users. You'd just do a simple query to remove the rows that give access privileges to those users. But if you had roles, you'd first need to find all the roles that include access, remove those from the specific users, and then find all the roles that are like the previous role of each user minus the access permission, and assign those roles to the users. Seems like a lot of work.

Related

User roles schema design comparison

I'm learning how to design a DB structure for assigning users permission to access certain pages
if the user is an admin that user would have access to crud operations
if the user is an editor that user would have access to only edit
user can have custom permission then access it would vary depending on the config
I have two schema designs and both seems good, one requires simple queries and the other can hold more description about each role and permission.
Design 1
role id is stored in a table called user and i will need to lookup role_has_permission table get all the permission ids then lookup permission table to get the permission_name column. comparatively longer query with more data being fetched, but i can have description column in permission table
Design 2
role id stored in table user, i can simply make a single query and check for permission. eg: role.canEdit is set to true user is allowed to edit. smaller and faster query.
why cant i go with the second design? and why do many articles go with the first design?
Design 1 lets you add permissions dynamically without changing the software. If you need a new permission, say can order lunch for entire team, you just add a record in the permission table and as many in the role_has_permission as needed, and you're done. In design 2 you'd have to add an operation canOrderLunchForEntireTeam. So design 1 is more flexible.
However, the flexibility of design 1 has a price. It's not enough to define and assign these permissions, but the software shall probably also check them when a function is performed. Adding a function for ordering lunch is a software change anyway, so adding an attribute to your design 2 class might be tolerable. The generic way of defining permissions in design 1 will therefore only pay out if you implement a similarly generic way of applying them.

How to correctly handle multiple privileged MySQL connection accounts

I have read many posts and articles that state quite clearly that for the best in worst-case damage limitation, MySQL accounts should only be able to do what they're intended to do, and nothing else. This makes perfect sense and so there would be a connection account to my Database that does all the SELECTing of data, and another account that does UPDATE and INSERT activities, so that if someone does get the compromise by SELECTing data they shouldn't, then they can't quite as easily then UPDATE that data or INSERT etc.
You get the idea.
But, I have various Databases and use accounts to read/SELECT data and the this is output to the client page, but often these accounts will need to be updated, small things such as updating when a (website) user logs into their account, or updating some sort of hitcounter or other minor feature. Given the ringfencing of concerns outlined above, I feel it's a bit like using a flood to put out a campfire, to allow UPDATEing (etc.) to a privileged MySQL connection simply to say that user Bob logged in last at 4:10pm.
I have been digging on the web for suitable guides, blog posts and articles about how to best structure using multiple MySQL privileged accounts to complete the nessecary work with as minimum a risk of excess privilege as possible, but I have found nothing that has been much use, (mostly because of my wording seems to be attracting articles about setting up website users, or other topics associated with these keywords :-/ )
I have a few ideas on current approach and wanted a bit of feedback on the best method for doing activities as described in paragraph 2, typically 95% SELECTing, and a few specific instances of UPDATEing, or if any of the following are possible (or on flipside, are very bad ideas)? :
I currently have seperate PHP connection objects for each connection privilege user. Is this the best approach?
Could I somehow giving a privilege user access to only update a certain table (or even a certain table column?), as well as SELECT from any table? This would be perfect.
Are using TRIGGERs a common approach and would this have any down sides if I created a Trigger (with a privileged user) and then let a SELECT user account access triggers?
Could I set certain users can only use certain triggers?
Is there another way of doing this?
MySQL allows for users to have different privileges set both at database and individual table levels. The documentation on the GRANT (http://dev.mysql.com/doc/refman/5.7/en/grant.html) syntax gives an example of setting all privileges to a user on one database while only select access to a table in another database.
Privileges can even be set for specific columns (http://dev.mysql.com/doc/refman/5.7/en/grant.html#grant-column-privileges) in a table & also for stored procedures (http://dev.mysql.com/doc/refman/5.7/en/grant.html#grant-routine-privileges).

Granting permission to individual fields in MySQL

I have MySQL Database and have several tables in it. One, of course, is the users table for storing the username and passwords (which also has the rest of their information). For all the tables in the schema, I want to be able to grant users (or groups) permission to individual fields. Maybe an example will help:
There;'s a table called ACCOUNTS. In this table, the user fills out (and keeps up to date) all the data about their company (name, address, POC, etc). But I also want to have fields attached to this table that are read only for those users, such as LastPaymentDate (can't have them changing that!). Further, among those users, permissions differ. For example, the admin/superuser can change the name and address of the company, but standard users should not.
I'm thinking this might need to be done by making several Views for the table, one for each level of permission (group). I'm relatively new to MySQL, so I don't know if this is the best way. I can also see a lookup table that says which fields is allowed to view/edit.
My initial thought was to include in the comments (or the name of the field) a value from 0-5, and then the user would have a permission level (0-can't see; 1-Read only; 2-Read-write; 3-(not used); 4-(not used); 5-Edit/Delete the field itself.
Any suggestions? Views? Lookup table to determine which fields to display? Again, it'd not for the whole table, for each column within a table.
You can GRANT the rights to individual columns to a user, using this code:
GRANT SELECT (col1), INSERT (col1,col2) ON mydb.mytbl TO 'someuser'#'somehost';
Example taken from here:
http://dev.mysql.com/doc/refman/5.1/en/grant.html
Also there is no support for groups of users or SQL ROLES (which are groups of privileges) in MySQL.

Database schema for ACL

I want to create a schema for a ACL; however, I'm torn between a couple of ways of implementing it.
I am pretty sure I don't want to deal with cascading permissions as that leads to a lot of confusion on the backend and for site administrators.
I think I can also live with users only being in one role at a time. A setup like this will allow roles and permissions to be added as needed as the site grows without affecting existing roles/rules.
At first I was going to normalize the data and have three tables to represent the relations.
ROLES { id, name }
RESOURCES { id, name }
PERMISSIONS { id, role_id, resource_id }
A query to figure out whether a user was allowed somewhere would look like this:
SELECT id FROM resources WHERE name = ?
SELECT * FROM permissions WHERE role_id = ? AND resource_id = ? ($user_role_id, $resource->id)
Then I realized that I will only have about 20 resources, each with up to 5 actions (create, update, view, etc..) and perhaps another 8 roles. This means that I can exercise blatant disregard for data normalization as I will never have more than a couple of hundred possible records.
So perhaps a schema like this would make more sense.
ROLES { id, name }
PERMISSIONS { id, role_id, resource_name }
which would allow me to lookup records in a single query
SELECT * FROM permissions WHERE role_id = ? AND permission = ? ($user_role_id, 'post.update')
So which of these is more correct? Are there other schema layouts for ACL?
In my experience, the real question mostly breaks down to whether or not any amount of user-specific access-restriction is going to occur.
Suppose, for instance, that you're designing the schema of a community and that you allow users to toggle the visibility of their profile.
One option is to stick to a public/private profile flag and stick to broad, pre-emptive permission checks: 'users.view' (views public users) vs, say, 'users.view_all' (views all users, for moderators).
Another involves more refined permissions, you might want them to be able to configure things so they can make themselves (a) viewable by all, (b) viewable by their hand-picked buddies, (c) kept private entirely, and perhaps (d) viewable by all except their hand-picked bozos. In this case you need to store owner/access-related data for individual rows, and you'll need to heavily abstract some of these things in order to avoid materializing the transitive closure of a dense, oriented graph.
With either approach, I've found that added complexity in role editing/assignment is offset by the resulting ease/flexibility in assigning permissions to individual pieces of data, and that the following to worked best:
Users can have multiple roles
Roles and permissions merged in the same table with a flag to distinguish the two (useful when editing roles/perms)
Roles can assign other roles, and roles and perms can assign permissions (but permissions cannot assign roles), from within the same table.
The resulting oriented graph can then be pulled in two queries, built once and for all in a reasonable amount of time using whichever language you're using, and cached into Memcache or similar for subsequent use.
From there, pulling a user's permissions is a matter of checking which roles he has, and processing them using the permission graph to get the final permissions. Check permissions by verifying that a user has the specified role/permission or not. And then run your query/issue an error based on that permission check.
You can extend the check for individual nodes (i.e. check_perms($user, 'users.edit', $node) for "can edit this node" vs check_perms($user, 'users.edit') for "may edit a node") if you need to, and you'll have something very flexible/easy to use for end users.
As the opening example should illustrate, be wary of steering too much towards row-level permissions. The performance bottleneck is less in checking an individual node's permissions than it is in pulling a list of valid nodes (i.e. only those that the user can view or edit). I'd advise against anything beyond flags and user_id fields within the rows themselves if you're not (very) well versed in query optimization.
This means that I can exercise blatant
disregard for data normalization as I
will never have more than a couple
hundred possible records.
The number of rows you expect isn't a criterion for choosing which normal form to aim for.
Normalization is concerned with data integrity. It generally increases data integrity by reducing redundancy.
The real question to ask isn't "How many rows will I have?", but "How important is it for the database to always give me the right answers?" For a database that will be used to implement an ACL, I'd say "Pretty danged important."
If anything, a low number of rows suggests you don't need to be concerned with performance, so 5NF should be an easy choice to make. You'll want to hit 5NF before you add any id numbers.
A query to figure out if a user was
allowed somewhere would look like
this:
SELECT id FROM resources WHERE name = ?
SELECT * FROM permissions
WHERE role_id = ? AND resource_id = ? ($user_role_id, $resource->id)
That you wrote that as two queries instead of using an inner join suggests that you might be in over your head. (That's an observation, not a criticism.)
SELECT p.*
FROM permissions p
INNER JOIN resources r ON (r.id = p.resource_id AND
r.name = ?)
You can use a SET to assign the roles.
CREATE TABLE permission (
id integer primary key autoincrement
,name varchar
,perm SET('create', 'edit', 'delete', 'view')
,resource_id integer );

Permissions for web site users

I'm working on a web site where each user can have multiple roles/permissions such as basic logging in, ordering products, administrating other users, and so on. On top of this, there are stores, and each store can have multiple users administrating it. Each store also has it's own set of permissions.
I've confused myself and am not sure how best to represent this in a db. Right now I'm thinking:
users
roles
users_roles
stores
stores_users
But, should I also have stores_roles and stores_users_roles tables to keep track of separate permissions for the stores or should I keep the roles limited to a single 'roles' table?
I originally thought of having only a single roles table, but then what about users who have roles in multiple stores? I.e., if a user is given a role of let's say 'store product updating' there would need to be some method of determining which store this is referring to. A stores_users_roles table could fix this by having a store_id field, thus a user could have 'store product updating' and 'store product deletion' for store #42 and only 'store product updating' for store #84.
I hope I'm making sense here.
Edit
Thanks for the info everyone. Apparently I have some thinking to do. This is simply a fun project I'm working on, but RBAC has always been something that I wanted to understand better.
This is probably obvious to you by now, but role based access control is hard. My suggestion is, don't try to write your own unless you want that one part to take up all the time you were hoping to spend on the 'cool stuff'.
There are plenty of flexible, thoroughly-tested authorization libraries out there implementing RBAC (sometimes mislabeled as ACL), and my suggestion would be to find one that suits your needs and use it. Don't reinvent the wheel unless you are a wheel geek.
It seems likely to me that if I have permission to do certain roles in a set of stores, then I would probably have the same permissions in each store. So having a single roles table would probably be sufficient. So "joe" can do "store product updating" and "store product deletion", then have a user_stores table to list which stores he has access to. The assumption is for that entire list, he would have the same permissions in all stores.
If the business rules are such that he could update and delete in one store, but only update, no delete, in another store, well then you'll have to get more complex.
In my experience you'll usually be told that you need a lot of flexibility, then once implemented, no one uses it. And the GUI gets very complex and makes it hard to administer.
If the GUI does get complex, I suggest you look at it from the point of view of the store as well as the point of view of the user. In other words, instead of selecting a user, then selecting what permissions they have, and what stores they can access, it may be simpler to first select a store, then select which users have access to which roles in that store. Depends I guess on how many users and how many stores. In a past project I found it far easier to do it one way than the other.
Your model looks ok to me. The only modification I think you need is as to the granularity of the Role. Right now, your role is just an operation.
But first, you need a store_role table, a joint table resolving the Many-to-many relationship b/w a role and a store. ie, one store can have many roles and one role can be done in many stores.
Eg: StoreA can CREATE, UPDATE, DELETE customer. and DELETE customer can be done in StoreA, StoreB and StoreC.
Next, you can freely associate users to store_role_id in the user_store_roles table.
Now, a user_store_role record will have a user_id and a store_role_id:
A collection of
SELECT * FROM USER_STORE_ROLE WHERE user_id = #userID
returns all permitted operations of the user in all the stores.
For a collection of users's roles in a particular store, do an inner join of the above to user_store table adding a WHERE part of like
where STORE_ROLE.store_id = #storeID
Put a store_id in the user_roles table.
If this is Rails, the user model would have_many :stores, :through => :roles