@ipbyrne/mongo-encrypted-query

Mongo Encrypted Query

This is a simple SDK that can be used to encrypt all of your data in MongoDB in a way that makes it searchable with any MongoDB equality operator.

How To Use

This package is intended to be layed over any implementation using MongoDB by utilizing 3 functions whenever you are interacting into the database.

Encrypt Data (encryptData)

This function is to be used whenever you are saving data into the database. You are expected to pass into this functio the data to be saved, the public key you are going to use to encrypt the data, and the salt you want to use to make the data queryable.

Encrypt Query (encryptQuery)

This function is to be used to format any query you want to use when querying the database. This function takes in the traditional MongoDB query, along with the salt used when saving the data you are trying to query.

Decrypt Data (decryptData)

This function is used to decrypt the data returned from MongoDB. This function is expecting the data returned from MongoDB, the private key you will use to decrypt the data, and the salt used when saving the data.

This will then return the decrypted data.

Generate Encryption Key (generateEncryptionKeyPair)

This function is used to generate a key pair that can be used for encrypting and decrypting the data. These keys should never be saved in source or the database and should be stored in a key service where they are accessed whenever you are encrypting/decrypting data.

This will then return the decrypted data.

Working In The Repo

To run the test to see how this works in practice you can run:

npm install
npm run test

Out of the box, this will support all MongoDB equality query operators.

How Data is Stored

Once data is added to the DB this will be it's shape:

{
  "7818": "640e9625f2d4c216e3786e36ad726f8e4445086202b8d0cda065b56c51704cfd9da8c7d75500cb88cbfd8ae328",
  "511f9771f3c4d306": [
    "79088c6ff49b845dae6a2b2ab92222d8040f15330ab8c5d6f226b625552608f2d1a1839e1b02",
    "79088c6ff49b845dae2e3560e07e7ed0591c486059ec9c9bf838ba3549670aaa"
  ],
  "6505887a": [
    "47198a76e1c8ca10b5781f76ab7569d902015b6d"
  ],
  "7f1d957a": "45198b6ba792",
  "75198b7cf5c8db06b07232": "54049972f7cdce529a6f3960ab7f78de17041a32",
  "780f8b6ae6cfc8179d7c2861": "234cc927aa91995feb290834fb2b3e8f4c580e5b",
  "720e9d7be2cfdf1bb8710f71ac7b69d402": {
    "7818": "75159c25e2d9ca1fa971393efc2238",
    "70189c6de2d2d8": {
      "6b15885ce8c5ce": "264bcb28be"
    },
    "7f198b6be2c5ea00ab7c2540af656d": [
      {
        "78129c7aff": "24"
      },
      {
        "78129c7aff": "27"
      },
      {
        "78129c7aff": "26"
      }
    ]
  },
  "780f8b6ae2d3": {
    "7818": "75159c25f0c4c948bc653d69be7d6999150757"
  }
}

In practice, each user or organization should use their own private key for encryption so no users will have matching encrypted values even if they are storing the same values.

These keys should never be comitted to source or saved in the database directly and ideally would be kept in a service like Google Secret Manager or another providers equivelant service.

You would then retrieve the keys when encrypting and decrypting data.

The ones in the 'key.js' file are there so you can easily test and play with the repository to understand how everything works together.

Optional Configurations

Encrypting Keys

If you are not worried about the keys being readable in the database you can opt out of encrypting them. One benefit of doing this would be it will allow you to build indexes on fields to speed up queries.