en-US/about_ProtectedData.help.txt
|
TOPIC
about_ProtectedData SHORT DESCRIPTION Provides background information about the ProtectedData module. About ProtectedData When you need to store secret data, such as a set of credentials, in a PowerShell script, you would typically use the Export-Clixml or ConvertFrom-SecureString cmdlets to accomplish this. These commands leverage the Windows Data Protection API (DPAPI) to perform the encryption. The DPAPI is extremely convenient, but it has a limitation: the data can only be decrypted by the user who originally encrypted it, and in many cases, this decryption can only happen on the same computer where the encryption took place (unless you have Credential Roaming or Roaming Profiles enabled in an Active Directory environment.) The ProtectedData module exists to overcome this limitation, while still allowing the convenience of not having to worry about managing or protecting encryption keys. It does this, primarily, by leveraging digital certificates. How It Works When you send a piece of data to the Protect-Data command, it is encrypted using a randomly-generated AES key and initialization vector (IV). Copies of this key and IV are then encrypted using either the public keys of RSA or ECDH certificates, or using a password-derived AES key (more on that later.) The resulting object can be persisted to disk with Export-Clixml, and can later be read back in with Import-Clixml and then passed to the Unprotect-Data command. When you call Unprotect-Data, you must pass in either one of the passwords that was to protect the data, or the thumbprint of one of the certificates that was used to protect the data. If you use a certificate when calling Unprotect-Data, you must have the certificate's private key. Regarding Password-derived Keys This module's intended use is to leverage certificate-based encryption wherever possible. This is what provides security, without the need for the user to worry about key protection or key management; the operating system takes care of this for you when you install a certificate (with or without its private key.) The various -Password parameters to the ProtectedData module's commands are intended as a backup mechanism. If you are unable to decrypt the data with a certificate for some reason, you'd be able to enter the correct password to retrieve it or to add a new certificate-encrypted copy of the keys. If you do use the Password functionality of the module, you're encouraged to always enter these passwords interactively. If you try to persist the passwords in some way, you're back to the original problem: you can either use DPAPI and accept its limitations, or you have to manage and protect encryption keys yourself. All passwords are passed to the ProtectedData commands in the form of SecureString objects. Supported Data Types All data must be serialized to a byte array before it can be encrypted. The ProtectedData module supports automatic serialization / deserialization of PSCredential, SecureString, and String objects. If you want to encrypt another data type instead, you're responsible for converting it to a byte array yourself first, and passing the resulting byte array to Protect-Data's -InputObject parameter. The ProtectedData object which is returned from the Protect-Data command includes a Type property. When you pass the object to Unprotect-Data, it uses this information to build and return an object of the original type for you (PSCredential, SecureString, String or Byte[] .) Regarding In-Memory Security The commands in the ProtectedData module make an effort to minimize the amount of time that any sensitive, unencrypted data is left in memory as well, but this is a tricky topic in a .NET application. The Garbage Collector can sometimes create copies of unencrypted byte arrays before the module has had a chance to pin them. This in-memory security is provided on a "best effort" basis. Certificate requirements The RSA certificates used with this module must allow Key Encipherment in their Key Usage extension, if present. ECDH certificate must allow the Key Agreement Key Usage extension. All certificates must also be issued by a trusted certificate authority and be currently valid, unless you pass the -SkipCertificateVerification switch parameter when calling the various commands in the module. This switch allows you to leverage self-signed certificates, etc, if you don't care about validating a trust chain and just want to get at the certificate's key pair. You can verify which of your certificates are usable for both encryption and decryption ahead of time by running the following command: Get-KeyEncryptionCertificate -RequirePrivateKey -SkipCertificateVerification (With this set of parameters, the command searches the entire Cert: drive, including both CurrentUser and LocalMachine stores, and as mentioned earlier, the -SkipCertificateVerification switch ignores certificate validity periods, trust chains and revocation status.) SEE ALSO Protect-Data Unprotect-Data Add-ProtectedDataCredential Remove-ProtectedDataCredential Get-ProtectedDataSupportedTypes Get-KeyEncryptionCertificate |