Security

Core Concepts

API authentication is largely a solved problem and generally outside the scope of Elide.Authorization - the act of verifying data and operation access for an already authenticated user in the Elide framework involves a few core concepts:

  • User - Each API request is associated with a user principal. The user is opaque to the Elide framework but is passed to developer-defined check functions that evaluate arbitrary logic or build filter expressions.
  • Checks - a function or filter expression that grants or denies a user permission to perform a particular action.
  • Permissions - a set of annotations (read, update, delete, create, and share) that correspond to actions on the data model’s entities and fields. Each permission is decorated with one or more checks that are evaluated when a user attempts to perform that action.

    Security Evaluation

Security is applied hierarchically with three goals:

  • Granting or denying access. When a model or field is accessed, a set of checks are evaluated to determine if the access will be denied (i.e. 403 HTTP status code (JSON-API) or GraphQL error object) or permitted. If a user has explicitly requested access to part of the data model they should not see, the request will be rejected.
  • Filtering Collections. If a model has read permissions defined, these checks are evaluated against each model that is a member of the collection. Only the models the user has access to (by virtue of being of being able to read at least one of the model’s fields) are returned in the response.
  • Filtering a model. If a user has read access to a model, but only for a subset of a model’s fields, the disallowed fields are excluded from the output (rather than denying the request). However, when the user explicitly requests a field-set that contains a restricted field, the request is rejected rather than filtered.

    Hierarchical Security

Both JSON-API and GraphQL define mechanisms to fetch and manipulate entities defined by the data model schema. Some (rootable) entities can be reached directly by providing their data type and unique identifier in the query. Other entities can only be reached through relationships to other entities– by traversing the entity relationship graph. The Elide framework supports both methods of access. This is beneficial because it alleviates the need for all models to be accessible at the root of the graph. When everything is exposed at the root, the developer needs to enumerate all of the valid access patterns for all of the data models which quickly becomes unmanageable. In addition to eliminating redundancy in security declaration, this form of security can have significant performance benefits for enforcing security on large collections stored in key-value stores that have limited ability for the underlying persistence layer to directly apply security filters. It is often possible to deny access to an entire collection (i.e. hierarchical relationship) before attempting to verify access to each individual member within that collection. Typically, security rules only need to be defined for a subset of models and relationships– often near the roots of the graph. Applying security rules to the relationships to prune the graph can eliminate invalid access patterns. To better understand the sequence of how security is applied, consider the data model depicted in Figure 1 consisting of articles where each contains zero or more comments.

Security Article Comment UML

The request to update a specific comment of a particular article involves the following permission checks:

  • Read permission check on the Article’s comments field.
  • Update permission check on the Comment’s title field.
    When a client modifies one side of a bidirectional relationship, Elide will automatically update the opposite side of the relationship. This implies that the client must have permission to write both sides of the relationship.

Checks

Checks are simply functions that either return:

  • whether or not access should be granted to the requesting user.

  • a filter expression that can be used to filter a collection to what is visible to a given user.
    Checks can either be invoked:

  • immediately prior to the (create, read, update, and delete) action being performed.

  • immediately before committing the transaction that wraps the entire API request.
    The former is useful for checks that depend on data that is already persisted. Read, delete, and most update checks fall into this category. The latter is useful for checks involving complex mutations to the object graph for newly created data. These checks cannot execute until the data model has been fully updated to reflect the entire set of modifications. Corresponding to these categories, there are two types of checks in Elide: InlineCheck and CommitCheck. Checks must be implemented by extending one of these abstract classes. The class you choose to extend impacts how Elide will evaluate your the check. There are specific types of check described by the following hierarchy:

InlineCheck is the abstract super class of the three specific variants:

Operation Checks

Operation checks are inline checks whose evaluation requires the entity instance being read from or written to. They operate in memory of the process executing the Elide library.

Operation checks are expected to implement the following Check interface:

  1.     /**
  2.      * Determines whether the user can access the resource.
  3.      *
  4.      * @param object Fully modified object
  5.      * @param requestScope Request scope object
  6.      * @param changeSpec Summary of modifications
  7.      * @return true if security check passed
  8.      */
  9.     boolean ok(T object, RequestScope requestScope, Optional<ChangeSpec> changeSpec);

User Checks

User checks depend strictly on the user principal. These are inline checks (i.e. they run as operations occur rather than deferring until commit time) and only take a User object as input. Because these checks only depend on who is performing the operation and not on what has changed, these checks are only evaluated once per request - an optimization that accelerates the filtering of large collections.

User checks are expected to implement the following Check interface:

  1.     /**
  2.      * Method reserved for user checks.
  3.      *
  4.      * @param user User to check
  5.      * @return True if user check passes, false otherwise
  6.      */
  7.     boolean ok(User user);

Filter Expression Checks

In some cases, the check logic can be pushed down to the data store itself. For example, a filter can be added to a database query to remove elements from a collection where access is disallowed. These checks return a FilterExpression predicate that your data store can use to limit the queries that it uses to marshal the data. Checks which extend the FilterExpessionCheck must conform to the interface:

  2. /**
  3.  * Check for FilterExpression. This is a super class for user defined FilterExpression check. The subclass should
  4.  * override getFilterExpression function and return a FilterExpression which will be passed down to datastore.
  5.  * @param <T> Type of class
  6.  */
  7. public abstract class FilterExpressionCheck<T> extends InlineCheck<T> {
  9.     /**
  10.      * Returns a FilterExpression from FilterExpressionCheck.
  11.      * @param entityClass The entity collection to filter
  12.      * @param requestScope Request scope object
  13.      * @return FilterExpression for FilterExpressionCheck.
  14.      */
  15.     public abstract FilterExpression getFilterExpression(Class<?> entityClass, RequestScope requestScope);
  16. }

Most FilterExpressionCheck functions construct a FilterPredicate which is a concrete implementation of the FilterExpression interface:

  1. /**
  2.  * Constructs a filter predicate
  3.  * @param path The path through the entity relationship graph to a particular attribute to filter on.
  4.  * @param op The filter comparison operator to evaluate.
  5.  * @param values The list of values to compare the attribute against.
  6.  */
  7. public FilterPredicate(Path path, Operator op, List<Object> values) {
  8.  ...
  9. }

Here is an example to filter the Author model by book titles:

  1.    //Construct a filter for the Author model for books.title == 'Harry Potter'
  2.    Path.PathElement authorPath = new Path.PathElement(Author.class, Book.class, "books");
  3.    Path.PathElement bookPath = new Path.PathElement(Book.class, String.class, "title");
  4.    List<Path.PathElement> pathList = Arrays.asList(authorPath, bookPath);
  5.    Path path = new Path(pathList);
  7.    return new FilterPredicate(path, Operator.IN, Collections.singletonList("Harry Potter"));

Filter expression checks are most important when a security rule is tied in some way to the data itself. By pushing the security rule down to the datastore, the data can be more efficiently queried which vastly improves performance. Moreover, this feature is critical for implementing a service that requires complex security rules (i.e. anything more than role-based access) on large collections.

User

Each request is associated with a user. The user is computed by a function that you provide conforming to the interface:

  1. Function<SecurityContext, Object>

Permission Annotations

The permission annotations include ReadPermission, UpdatePermission, CreatePermission, DeletePermission, and SharePermission. Permissions are annotations which can be applied to a model at the package, entity, or field-level. The most specific annotation always take precedence (package < entity < field). More specifically, a field annotation overrides the behavior of an entity annotation. An entity annotation overrides the behavior of a package annotation. Entity annotations can be inherited from superclasses. When no annotation is provided at any level, access is implicitly granted for ReadPermission, UpdatePermission, CreatePermission, and DeletePermission and implicitly denied for SharePermission.

The permission annotations wrap a boolean expression composed of the check(s) to be evaluated combined with AND, OR, and NOT operators and grouped using parenthesis. The checks are uniquely identified within the expression by a string - typically a human readable phrase that describes the intent of the check (“principal is admin at company OR principal is super user with write permissions”). These strings are mapped to the explicit Check classes at runtime by registering them with Elide. When no registration is made, the checks can be identified by their fully qualified class names. The complete expression grammar can be found here.

To better understand how permissions work consider the following sample code. (Only the relevant portions are included.)

  1. @ReadPermission(expression = "Prefab.Roll.All")
  2. @UpdatePermission(expression = "user is a superuser OR user is this user")
  3. @DeletePermission(expression = "user is a superuser OR user is this user")
  4. @SharePermission(expression = "Prefab.Role.All")
  5. public class User {
  6.     String name;
  8.     @OneToMany(mappedBy = "author")
  9.     Set<Post> posts;
  10. }
  1. @ReadPermission(expression = "Post.isVisible OR User.ownsPost OR user is a superuser")
  2. @UpdatePermission(expression = "user owns this post now")
  3. @CreatePermission(expression = "user owns this post")
  4. @DeletePermission(expression = "user owns this post now")
  5. @SharePermission(expression = "Prefab.Role.All")
  6. public class Post {
  7.     @ManyToOne
  8.     User author;
  10.     @UpdatePermission(expression = "user owns this post now OR user is a superuser now")
  11.     boolean published;
  13.     @OneToMany(mappedBy = "post")
  14.     Set<Comment> comments;
  15. }
  1. // user has to be able to see the post and to see the comment, or else be a super user
  2. @ReadPermission(expression = "((Post.isVisible OR User.ownsPost) AND (comment is visible OR user made this comment)) OR user is a superuser")
  3. @UpdatePermission(expression = "user made this comment")
  4. @CreatePermission(expression = "post is visible now")
  5. @DeletePermission(expression = "user made this comment")
  6. @SharePermission(expression = "Prefab.Role.All")
  7. public class Comment {
  8.     @ManyToOne
  9.     User author;
  10.     @ManyToOne
  11.     Post post;
  12.     @UpdatePermission(expression = "user owns this post now OR user is a superuser now")
  13.     boolean suppressed;
  14. }
  1. public class IsOwner {
  2.     public static class Inline<Post> extends OperationCheck {
  3.         @Override
  4.         boolean ok(Post post, RequestScope requestScope, Optional<ChangeSpec> changeSpec) {
  5.             return post.author.equals(requestScope.getUser());
  6.         }
  7.     }
  8.     public static class AtCommit<Post> extends CommitCheck {
  9.         @Override
  10.         boolean ok(Post post, RequestScope requestScope, Optional<ChangeSpec> changeSpec) {
  11.             return post.author.equals(requestScope.getUser());
  12.         }
  13.     }
  14. }
  1. public class IsSuperuser {
  2.     public static class Inline<User> extends OperationCheck {
  3.         @Override
  4.         boolean ok(User user, RequestScope requestScope, Optional<ChangeSpec> changeSpec) {
  5.             return user.isSuperuser();
  6.         }
  7.     }
  8. }
  1. HashMap checks = new HashMap();
  2. checks.put("post is visible now", IsPublished.Inline.class);
  3. checks.put("post is visible", IsPublished.AtCommit.class);
  4. checks.put("user owns this post now", IsOwner.Inline.class);
  5. checks.put("user owns this post", IsOwner.AtCommit.class);
  6. checks.put("user is a superuser", IsSuperuser.Inline.class);
  7. //...
  8. EntityDictionary dictionary = new EntityDictionary(checks);

You will notice that IsOwner actually defines two check classes; it does so because we might want to evaluate the same logic at distinct points in processing the request (inline when reading a post and at commit when creating a post). For example, we could not apply IsOwner.Inline when creating a new post because the post’s author has not yet been assigned. Once the post has been created and all fields assigned by Elide, the security check can be evaluated.

Contrast IsOwner to IsSuperuser which only defines one check. IsSuperuser only defines one check because it only depends on who is performing the action and not on the data model being manipulated.

Read

ReadPermission governs whether a model or field can be read by a particular user. If the expression evaluates to true then access is granted. Notably, ReadPermission is evaluated as the user navigates through the entity relationship graph. Elide’s security model is focused on field-level access, with permission annotations applied on an entity or package being shorthand for applying that same security to every field in that scope. For example, if a request is made to GET /users/1/posts/3/comments/99 the permission execution will be as follows:

  • ReadPermission on User<1>#posts
  • ReadPermission on Post<3>#comments
  • ReadPermission on any field on Comment<99>
    If all of these checks succeed, then the response will succeed. The contents of the response are determined by evaluating the ReadPermission on each field. The response will contain the subset of fields where ReadPermission is granted. If a field does not have an annotation, then access defaults to whatever is specified at the entity level. If the entity does not have an annotation, access defaults to whatever is specified at the package. If the package does not have an annotation, access defaults to granted.

Update

UpdatePermission governs whether a model can be updated by a particular user. Update is invoked when an attribute’s value is changed or values are added to or removed from a relationship. Examples of operations that will evaluate UpdatePermission given objects Post and User from the code snippets above:

  • Changing the value of Post.published will evaluate UpdatePermission on published. Because more specific checks override less specific checks, the UpdatePermission on the entity Post will not be evaluated.
  • Setting Post.author = User will evaluate UpdatePermission on Post since author does not have a more specific annotation. Because author is a bidirectional relationship, UpdatePermission will also be evaluated on the User.posts field.
  • Removing Post from User.posts will trigger UpdatePermission on both the Post and User entities.
  • Creating Post will not trigger UpdatePermission checks on any fields that are initialized in the request. However, it will trigger UpdatePermission on any bidirectional relationship fields on preexisting objects.

    Create

CreatePermission governs whether a model can be created or a field can be initialized in a newly created model instance.Whenever a model instance is newly created, initialized fields are evaluated against CreatePermission rather than UpdatePermission.

Delete

DeletePermission governs whether a model can be deleted.

Share

SharePermission governs whether an existing model instance (one created in a prior transaction) can be assigned to another collection other than the one in which it was initially created. Basically, does a collection ‘own’ the model instance in a private sense (composition) or can it be moved or referenced by other collections (aggregation).

Graph APIs generally have two ways to reference an entity for CRUD operations. In the first mechanism, an entity is navigable through the entity relationship graph. An entity can be reached only through other entities. The alternative is to provide a mechanism to directly reference any entity by its data type and an instance identifier. The former approach is especially useful for modeling object composition (as opposed to aggregation). It enables the definition of hierarchical security. The latter approach is especially useful when directly manipulating relationships or links between edges. More specifically, to add an existing entity to a collection, it is simplest to reference this entity by its type and ID in the API request rather than defining a path to it through the entity relationship graph. Elide distinguishes between these two scenarios by tracking an object’s lineage or path through the entity relationship graph. By default, Elide explicitly denies adding an existing (not newly created) entity to a relationship with another entity if it has no lineage. For object composition, this is typically the desired behavior. For aggregation, this default can be overridden by adding an explicit SharePermission. SharePermission is always either identical to ReadPermission for the entity or it is explicitly disallowed.

原文: http://elide.io/pages/guide/03-security.html