Features
On this page
Overview
On this page, you can learn about the security benefits of Queryable Encryption, how it works, and how it compares to other security mechanisms supported by MongoDB. You can also view a fictional scenario that demonstrates the value of Queryable Encryption in securing your data.
Queryable Encryption
Queryable Encryption enables a client application to encrypt data before transporting it over the network using fully randomized encryption, while maintaining queryability. Sensitive data is transparently encrypted and decrypted by the client and only communicated to and from the server in encrypted form.
Unlike Client-Side Field Level Encryption that can use Deterministic Encryption, Queryable Encryption uses fast, searchable encryption schemes based on structured encryption. These schemes produce different encrypted output values even when given the same cleartext input.
How Queryable Encryption Works
The diagram below shows the process and architecture of how Queryable Encryption is used in a customer environment.
In this diagram, the user is able to query on fully randomly encrypted data such as SSN number.
The process and mechanisms that make this possible within Queryable Encryption are as follows:
When the application submits the query, MongoDB drivers first analyze the query.
The driver recognizes the query is against an encrypted field and requests the encryption keys from the customer-provisioned key provider such as:
AWS Key Management Service (AWS KMS)
Google Cloud KMS
Azure Key Vault
Any KMIP-compliant key provider
The driver submits the query to the MongoDB server with the encrypted fields rendered as ciphertext.
Queryable Encryption implements a fast, searchable scheme that allows the server to process queries on fully encrypted data, without knowing anything about the data. The data and the query itself remain encrypted at all times on the server.
The MongoDB server returns the encrypted results of the query to the driver.
The query results are decrypted with the keys held by the driver and returned to the client and shown as plaintext.
Queryable Encryption functions with the help of the following data structures. It is critical that these are not modified or deleted, or query results will be incorrect.
Queryable Encryption adds a
__safeContent__
field to documents in any collection where there's a Queryable Encryption encrypted field.Queryable Encryption creates two internal metadata collections in the same database as the collection where there's a Queryable Encryption encrypted field. These are named as follows:
enxcol_.<collectionName>.esc
enxcol_.<collectionName>.ecoc
Warning
Do not modify these data structures or query results will be incorrect and security could be impacted.
Queryable Encryption keeps encrypted fields secure in the following scenarios:
Direct access to encrypted fields by a database superuser
Access to encrypted fields by reading the server's memory
Capture of encrypted fields over an insecure network
Access to on-disk encrypted fields by reading database or backup files
Frequency analysis attacks by identifying patterns in documents with encrypted fields
While all clients have access to the non-sensitive data fields, only appropriately-configured Queryable Encryption clients are able to run read and write queries using the encrypted data fields.
Important
Remote Key Management System
When you use Queryable Encryption in production, you must use a remote Key Management System (KMS) to store your encryption key.
To view a step-by-step guide demonstrating how to use a remote KMS with Queryable Encryption, see Tutorials.
To view a list of all supported KMS providers, see KMS Providers.
To learn more about why you should use a remote KMS, see Reasons to Use a Remote Key Management System.
Other Security Mechanisms
This section describes the following security mechanisms supported by MongoDB and explains their use cases and limitations:
Role-Based Access Control
Role-Based Access Control is a security mechanism that allows administrators to grant and restrict collection-level permissions for users. With the appropriate role definition and assignment, this solution prevents accidental disclosure of data and access.
Role-Based Access control cannot protect against the following scenarios:
Capture of the data over an insecure network
Access to on-disk data by reading database or backup files
Access to data by reading the server's memory
Direct access to data by a database superuser
To learn more, see Role-Based Access Control.
Encryption at Rest
Encryption at Rest is a mechanism that encrypts database files on disk. This mechanism prevents a person who lacks database credentials, but has access to the computer hosting your database, from viewing your data.
This mechanism does not protect your data against the following scenarios:
Capture of the data over an insecure network
Access to data by reading the server's memory
Direct access to data by a database superuser
To learn more, see Encryption at Rest.
Transport Encryption (TLS/SSL)
Transport Encryption using TLS/SSL encrypts your data over the network. TLS/SSL protects your data as it travels over an insecure network, but cannot protect your data from a privileged user or as it sits on disk.
To learn more, see Transport Encryption using TLS/SSL
Comparison of Features
To learn more about Client-Side Field Level Encryption, see Client-Side Field Level Encryption Features.
The following table describes potential security threats and how MongoDB encryption features address them. Use these mechanisms together: Role-Based Access Control, Encryption at Rest, Transport Encryption, and In-Use Encryption. Note that you can't use both Client-Side Field Level Encryption and Queryable Encryption in the same collection.
Important
This is a high-level summary meant for general comparison. For detailed information, see the Overview of Queryable Encryption and Design and Analysis of a Stateless Document Database Encryption Scheme whitepapers.
Threat | TLS/SSL Transport Encryption | Encryption at Rest (EaR) | Queryable Encryption (Equality) + TLS/SSL + EaR | CSFLE + TLS/SSL + EaR |
---|---|---|---|---|
Network Snooping (attacker has access to network traffic) | Reveals operation metadata | Reveals operation metadata | Reveals operation metadata | Reveals operation metadata |
Database Recoveries from Disk (attacker has physical disk access) | Reveals database | Reveals size of database and operation metadata | Reveals size of database and operation metadata | Reveals size of database and operation metadata |
Database exfiltration from Disk and
Memory (attacker has physical disk access and multiple database
snapshots) [1] | Reveals database | Reveals database | Reveals size of database and operation metadata | Reveals frequency of values and operation metadata |
Advanced Persistent Threat (attacker has long-term, continuous
access to network, disk, and memory while remaining undetected) | Reveals database | Reveals database | Queryable Encryption is not designed to protect against an ATP. See whitepaper for details. | CSFLE is not designed to protect against an ATP. See whitepaper for details. |
[1] | This assumes exfiltration occurs between completed operations. See whitepaper for details. |
Warning
Queryable Encryption defends against data exfiltration, not against adversaries with persistent access to an environment, or those who can retrieve both database snapshots and accompanying query transcripts/logs.
When using Queryable Encryption, equality and range queries offer similar security against attackers with database snapshots. However, an attacker with access to both database snapshots and query information is beyond the scope of Queryable Encryption's security guarantees. This is especially true for range queries, even if only a small number of query transcripts or logs are retrieved. See 6.1: Range Queries in the Persistent Model in the overview whitepaper for details.
Scenario
The following fictional scenario demonstrates the value of Queryable Encryption in securing your application's data, and how Queryable Encryption interacts with the other security mechanism discussed in this guide.
In this scenario, we secure sensitive data on a medical care management system that stores patients' personal information, billing information, and medical records for a fictional company, MedcoMD. None of the patient data is public, and specific data such as their social security number (SSN, a US government-issued ID number), patient ID number, billing information, and medication information are particularly sensitive and subject to privacy compliance. It is important for the company and the patient that the data is kept private and secure.
MedcoMD needs this system to satisfy the following use cases:
Doctors use the system to access patients' medical records, billing information, and update medications.
Receptionists use the system to verify patients' identities using their contact information.
Receptionists can view a patient's billing information, but not their patient ID number.
Receptionists cannot access a patient's medical records.
MedcoMD is also concerned with the disclosure of sensitive data through any of the following methods:
Accidental disclosure of data on a receptionist's publicly-viewable screen.
Direct access to the database by a superuser such as a database administrator.
Capture of data over an insecure network.
Access to data by reading the database server's memory.
Access to data by reading database or backup files.
What can MedcoMD do to balance the functionality and access restrictions of their medical care management system?
Solution
MedcoMD uses the following security mechanisms to satisfy their use cases and protect against the disclosure of sensitive medical data:
Transport Encryption (TLS/SSL) to secure data as it travels over the network.
Encryption at Rest to protect against disclosure of data by reading database or backup files.
Role-Based Access Control to limit the access of database users to the collections necessary for them to perform their tasks.
Encrypting sensitive fields with Queryable Encryption to satisfy the following use cases and constraints:
Prevent reading data from server memory as the Queryable Encryption encrypted data is never on the database server in an unencrypted form.
Allow receptionists to verify patients' identities and prevent accidental disclosure of sensitive data on a receptionist's publicly viewable screen by providing receptionists with a client that is not Queryable Encryption enabled.
Allow doctors to view sensitive data privately in their offices by providing doctors with a Queryable Encryption enabled client.
Learn More
To view a list of security measures you should implement to protect your MongoDB deployment, see the Security Checklist.
To start using Queryable Encryption, see the Quick Start.