Check Access Control
In Permify, you can perform two different types access checks,
- resource based authorization checks, structured in the following form:
Can user U perform action Y in resource Z ?
- subject based authorization checks, structured in the following form:
Which resources can user U edit ?
In this section we’ll look at the resource based check request of Permify.
You can find subject based access checks in Entity (Data) Filtering section.
Content
- Example Check Requests
- How Access Decisions Evaluated?
- Latency & Performance
- Parameters & Properties
Example Check requests
Resource Based Check (Relationships)
POST /v1/permissions/check
cr, err: = client.Permission.Check(context.Background(), &v1.PermissionCheckRequest {
TenantId: "t1",
Metadata: &v1.PermissionCheckRequestMetadata {
SnapToken: "",
SchemaVersion: "",
Depth: 20,
},
Entity: &v1.Entity {
Type: "repository",
Id: "1",
},
Permission: "edit",
Subject: &v1.Subject {
Type: "user",
Id: "1",
},
if (cr.can === PermissionCheckResponse_Result.RESULT_ALLOWED) {
// RESULT_ALLOWED
} else {
// RESULT_DENIED
}
})
Attribute Based (ABAC) Check With Context Data
client.permission.check({
tenantId: "t1",
metadata: {
snapToken: "",
schemaVersion: "",
depth: 20,
},
entity: {
type: "organization",
id: "1",
},
permission: "hr_manager",
subject: {
type: "user",
id: "1",
},
context: {
data: {
ip_address: "192.158.1.38",
},
},
}).then((response) => {
if (response.can === PermissionCheckResponse_Result.RESULT_ALLOWED) {
console.log("RESULT_ALLOWED");
} else {
console.log("RESULT_DENIED");
}
});
How Access Decisions Evaluated?
Access decisions are evaluated by stored relational tuples and your authorization model, Permify Schema.
In high level, access of an subject related with the relationships or attributes created between the subject and the resource. You can define this data in Permify Schema then create and store them as relational tuples and attributes, which is basically forms your authorization data.
Permify Engine to compute access decision in 2 steps,
- Looking up authorization model for finding the given action’s ( edit, push, delete etc.) relations.
- Walk over a graph of each relation to find whether given subject ( user or user set ) is related with the action.
Let’s turn back to above authorization question ( “Can the user 3 edit document 12 ?” ) to better understand how decision evaluation works.
When Permify Engine receives this question it directly looks up to authorization model to find document edit
action. Let’s say we have a model as follows
entity user {}
entity organization {
// organizational roles
relation admin @user
relation member @user
}
entity document {
// represents documents parent organization
relation parent @organization
// represents owner of this document
relation owner @user
// permissions
action edit = parent.admin or owner
action delete = owner
}
Which has a directed graph as follows:
As we can see above: only users with an admin role in an organization, which document:12
belongs, and owners of the document:12
can edit. Permify runs two concurrent queries for parent.admin and owner:
Q1: Get the owners of the document:12
.
Q2: Get admins of the organization where document:12
belongs to.
Since edit action consist or between owner and parent.admin, if Permify Engine found user:3 in results of one of these queries then it terminates the other ongoing queries and returns authorized true to the client.
Rather than or, if we had an and relation then Permify Engine waits the results of these queries to returning a decision.
Latency & Performance
With the right architecture we expect 7-12 ms latency. Depending on your load, cache usage and architecture you can get up to 30ms.
Permify implements several cache mechanisms in order to achieve low latency in scaled distributed systems. See more on the section Cache Mechanisims
Parameters & Properties
Path Parameters
Identifier of the tenant, if you are not using multi-tenancy (have only one tenant) use pre-inserted tenant <code>t1</code> for this field. Required, and must match the pattern \“[a-zA-Z0-9-,]+\“, max 64 bytes.
Body
PermissionCheckRequestMetadata metadata for the PermissionCheckRequest.
Entity represents an entity with a type and an identifier.
The action the user wants to perform on the resource
Subject represents an entity subject with a type, an identifier, and a relation.
Context encapsulates the information related to a single operation, including the tuples involved and the associated attributes.
Additional arguments associated with this request.
Response
Enumerates results of a check operation.
- CHECK_RESULT_UNSPECIFIED: Not specified check result. This is the default value.
- CHECK_RESULT_ALLOWED: Represents a successful check (the check allowed the operation).
- CHECK_RESULT_DENIED: Represents a failed check (the check denied the operation).
CHECK_RESULT_UNSPECIFIED
, CHECK_RESULT_ALLOWED
, CHECK_RESULT_DENIED
PermissionCheckResponseMetadata metadata for the PermissionCheckResponse.