Amazon Web Services 文档中描述的 Amazon Web Services 服务或功能可能因区域而异。要查看适用于中国区域的差异，请参阅 中国的 Amazon Web Services 服务入门 (PDF)。

动态数据屏蔽（预览版）

为 DDM 创建预览集群或工作组

概览

使用 Amazon Redshift 中的动态数据屏蔽（DDM），您可以保护数据仓库中的敏感数据。您可以操纵 Amazon Redshift 在查询时如何向用户显示敏感数据，而无需在数据库中对其进行转换。您可以通过将自定义模糊处理规则应用于给定用户或角色的屏蔽策略来控制对数据的访问。这样，您无需更改底层数据或编辑 SQL 查询即可响应不断变化的隐私要求。

动态数据屏蔽策略会隐藏、模糊处理或假名化与给定格式匹配的数据。附加到表时，对表的一个或多个列应用屏蔽表达式。您可以进一步修改屏蔽策略，使其仅应用于某些用户，或者应用于您可以使用 基于角色的访问控制（RBAC） 创建的用户定义角色。此外，在创建屏蔽策略时，您可以使用条件列在单元格级别应用 DDM。有关条件屏蔽的更多信息，请参阅 条件动态数据屏蔽。

您可以将具有不同模糊处理级别的多个屏蔽策略应用于表中的同一列，并将它们分配给不同的角色。为避免在不同角色对一列应用不同策略时发生冲突，可以为每个应用程序设置优先级。通过这种方式，您可以控制给定用户或角色可以访问哪些数据。DDM 策略可以部分或完全编辑数据，也可以使用以 SQL、Python 或 Amazon Lambda 编写的用户定义函数对其进行哈希处理。通过使用哈希屏蔽数据，您可以对这些数据应用联接，而无需访问潜在的敏感信息。

端到端示例

下面是一个端到端的示例，显示了如何创建屏蔽策略并将其附加到列。这些策略让用户可以访问列并查看不同的值，具体取决于与其角色相关的策略的混淆程度。您必须是超级用户或具有 sys:secadmin 角色才能运行此示例。

创建屏蔽策略

首先，创建一个表并填入信用卡值。

--create the table         
CREATE TABLE credit_cards (
  customer_id INT,
  credit_card TEXT
);

--populate the table with sample values
INSERT INTO credit_cards
VALUES
  (100, '4532993817514842'),
  (100, '4716002041425888'),
  (102, '5243112427642649'),
  (102, '6011720771834675'),
  (102, '6011378662059710'),
  (103, '373611968625635')
;

--run GRANT to grant permission to use the SELECT statement on the table
GRANT SELECT ON credit_cards TO PUBLIC;

--create two users
CREATE USER regular_user WITH PASSWORD '1234Test!';

CREATE USER analytics_user WITH PASSWORD '1234Test!';

--create the analytics_role role and grant it to analytics_user
--regular_user does not have a role
CREATE ROLE analytics_role;

GRANT ROLE analytics_role TO analytics_user;
      

接下来，创建应用于分析角色的屏蔽策略。

--create a masking policy that fully masks the credit card number
CREATE MASKING POLICY mask_credit_card_full
WITH (credit_card VARCHAR(256))
USING ('000000XXXX0000');

--create a user-defined function that partially obfuscates credit card data
CREATE FUNCTION REDACT_CREDIT_CARD (credit_card TEXT)
RETURNS TEXT IMMUTABLE
AS $$
    import re
    regexp = re.compile("^([0-9]{6})[0-9]{5,6}([0-9]{4})")
 
    match = regexp.search(credit_card)
    if match != None:
        first = match.group(1)
        last = match.group(2)
    else:
        first = "000000"
        last = "0000"
    
    return "{}XXXXX{}".format(first, last)
$$ LANGUAGE plpythonu;

--create a masking policy that applies the REDACT_CREDIT_CARD function
CREATE MASKING POLICY mask_credit_card_partial
WITH (credit_card VARCHAR(256))
USING (REDACT_CREDIT_CARD(credit_card));

--confirm the masking policies using the associated system views
SELECT * FROM svv_masking_policy;

SELECT * FROM svv_attached_masking_policy;         
      

附加屏蔽政策

将屏蔽政策附加到信用卡表中。

--attach mask_credit_card_full to the credit card table as the default policy
--all users will see this masking policy unless a higher priority masking policy is attached to them or their role
ATTACH MASKING POLICY mask_credit_card_full
ON credit_cards(credit_card)
TO PUBLIC;

--attach mask_credit_card_partial to the analytics role
--users with the analytics role can see partial credit card information
ATTACH MASKING POLICY mask_credit_card_partial
ON credit_cards(credit_card)
TO ROLE analytics_role
PRIORITY 10;

--confirm the masking policies are applied to the table and role in the associated system view
SELECT * FROM svv_attached_masking_policy;

--confirm the full masking policy is in place for normal users by selecting from the credit card table as regular_user
SET SESSION AUTHORIZATION regular_user;

SELECT * FROM credit_cards;

--confirm the partial masking policy is in place for users with the analytics role by selecting from the credit card table as analytics_user
SET SESSION AUTHORIZATION analytics_user;

SELECT * FROM credit_cards;
      

分离和删除屏蔽策略

以下部分显示如何通过从表中删除所有动态数据屏蔽策略来分离和删除屏蔽策略。

--reset session authorization to the default
RESET SESSION AUTHORIZATION;

--detach both masking policies from the credit_cards table
DETACH MASKING POLICY mask_credit_card_full 
ON credit_cards(credit_card) 
FROM PUBLIC;

DETACH MASKING POLICY mask_credit_card_partial 
ON credit_cards(credit_card) 
FROM ROLE analytics_role;

--drop both masking policies
DROP MASKING POLICY mask_credit_card_full;

DROP MASKING POLICY mask_credit_card_partial;         
      

使用动态数据屏蔽时的注意事项

使用动态数据屏蔽时，请考虑以下事项：