Abstract: Fault-tolerant storage of cryptographic information maintained on a fleet of HSMs may be provided by dividing the cryptographic information into a number of stripes which are distributed and stored on individual HSMs in the HSM fleet. Parity information is generated which allows one or more stripes to be regenerated if one or more stripes becomes corrupt or is lost. The parity information may be stored on an HSM in the HSM fleet, or outside the fleet on a storage service, HSM management hub, tangible computer-readable media, or other device. If an HSM in the HSM fleet fails, resulting in the loss of a stripe, an HSM in the fleet can recover the missing stripe by re-creating the missing stripe from the remaining stripes combined with the parity information. In some examples, stripes are mirrored within the fleet of HSMs.

Abstract: A service provider provides virtual computing services using a fleet of one or more host computer systems. Each of the host computer systems may be equipped with a trusted platform module (“TPM”). The service provider, the host computer systems, and the virtual computing environments generate attestations that prove the integrity of the system. The attestations are signed with a one-time-use cryptographic key that is verifiable against the public keys of the service provider, a host computer system, and a virtual computing environment. The public key of the host computer system is integrated into a hash tree that links the public key of the host computer system to the public key of the service provider. The public key of the virtual computing environment is signed using a one-time-use graphic key issued to the host computer system that hosts the virtual computing environment.

Abstract: Techniques described herein enhance information security in contexts that utilize key management systems and cryptographic keys. A cryptographic structure is utilized to maintain cryptographic keys with associated expiration times such that after an expiration time associated with a cryptographic key has passed, the cryptographic key is no longer accessible.

Abstract: A cryptographic key management service receives a request to import a first cryptographic key. In response to the request, the service creates a public cryptographic key and a private cryptographic key. The private cryptographic key is encrypted using a second cryptographic key to create an import key token. The import key token and the public cryptographic key are provided in response to the request. The service receives an encrypted first cryptographic key, which the service decrypts using the private cryptographic key to obtain the first cryptographic key. The service stores the first cryptographic key and enables its use for the performance of cryptographic operations.

Abstract: A service provider provides virtual computing services using a fleet of one or more host computer systems. Each of the host computer systems may be equipped with a trusted platform module (“TPM”). The service provider, the host computer systems, and the virtual computing environments generate attestations that prove the integrity of the system. The attestations are signed with a one-time-use cryptographic key that is verifiable against the public keys of the service provider, a host computer system, and a virtual computing environment. The public key of the host computer system is integrated into a hash tree that links the public key of the host computer system to the public key of the service provider. The public key of the virtual computing environment is signed using a one-time-use graphic key issued to the host computer system that hosts the virtual computing environment.

Abstract: A trusted co-processor can provide a hardware-based observation point into the operation of a host machine owned by a resource provider or other such entity. The co-processor can be installed via a peripheral card on a fast bus, such as a PCI bus, on the host machine. The provider can provide the customer with expected information that the customer can verify through a request to an application programming interface (API) of the card, and after the customer verifies the information the customer can take logical ownership of the card and lock out the provider. The card can then function as a trusted but limited environment that is programmable by the customer. The customer can subsequently submit verification requests to the API to ensure that the host has not been unexpectedly modified or is otherwise operating as expected.

Abstract: A service provider provides virtual computing services using a fleet of one or more host computer systems. Each of the host computer systems may be equipped with a trusted platform module (“TPM”). The service provider, the host computer systems, and the virtual computing environments generate attestations that prove the integrity of the system. The attestations are signed with a one-time-use cryptographic key that is verifiable against the public keys of the service provider, a host computer system, and a virtual computing environment. The public key of the host computer system is integrated into a hash tree that links the public key of the host computer system to the public key of the service provider. The public key of the virtual computing environment is signed using a one-time-use graphic key issued to the host computer system that hosts the virtual computing environment.

Abstract: A digital signature over a message may be compressed by determining a plurality of values based at least in part on the message. A mapping of the plurality of values over a digital signature scheme may be used to determine a value from which a portion of the compressed digital signature is decompressible by cryptographically deriving one or more components of the uncompressed digital signature. A public key may be used to verify the authenticity of the compressed digital signature and message.

Abstract: A data storage service redundantly stores data and keys used to encrypt the data. Data objects are encrypted with first cryptographic keys. The first cryptographic keys are encrypted by second cryptographic keys. The first cryptographic keys and second cryptographic keys are redundantly stored in a data storage system to enable access of the data objects, such as to respond to requests to retrieve the data objects. The second cryptographic keys may be encrypted by third keys and redundantly stored in the event access to a second cryptographic key is lost.

Abstract: Cryptographic keys can include logging properties that enable those keys to be used only if the properties can be enforced by the cryptographic system requested to perform one or more actions using the keys. The logging property can specify how to log use of a respective key. A key can also include a mutability property for specifying whether the logging property can be changed, and if so under what circumstances or in which way(s). The ability to specify and automatically enforce logging can be important for environments where audit logs are essential. These can include, for example, public certificate authorities that must provide accurate and complete audit trails. In cases where the data is not to be provided outside a determined secure environment, the key can be generated with a property indicating not to log any of the usage.

Abstract: A host machine operated for a specific purpose can have restricted access to other components in a multi-tenant environment in order to provide for the security of the host machine. The access restriction can prevent the host machine from obtaining updates to critical system-level configurations, but such information can be obtained through a signed command received to an API for the host machine. The command can be signed by a quorum of operators, and the host machine can be configured to verify the signatures and the quorum before processing the command. The host machine can store the updates to ephemeral storage as well as persistent storage, such that upon a reboot or power cycle the host machine can operate with current configuration data.

Abstract: Fault-tolerant storage of cryptographic information maintained on a fleet of HSMs may be provided by dividing the cryptographic information into a number of stripes which are distributed and stored on individual HSMs in the HSM fleet. Parity information is generated which allows one or more stripes to be regenerated if one or more stripes becomes corrupt or is lost. The parity information may be stored on an HSM in the HSM fleet, or outside the fleet on a storage service, HSM management hub, tangible computer-readable media, or other device. If an HSM in the HSM fleet fails, resulting in the loss of a stripe, an HSM in the fleet can recover the missing stripe by re-creating the missing stripe from the remaining stripes combined with the parity information. In some examples, stripes are mirrored within the fleet of HSMs.

Abstract: A secret is stored in a computing device. The device generates a value determined based at least in part on a substantially random process. As a result of the value satisfying a condition, the device causes the secret to be unusable to perform cryptographic operations such that the device is unable to cause the secret to be restored. The secret may be programmatically unexportable from the device.

Abstract: A client establishes a network session with a server. The network session is used to establish an encrypted communications session. The client establishes another network session with another server, such as after terminating the first network session. The client resumes the encrypted communications session over the network session with the other server. The other server is configured to receive encrypted communications from the client and forward them to the appropriate server.

Abstract: Fault-tolerant storage of cryptographic information maintained on a fleet of HSMs may be provided by dividing the cryptographic information into a number of stripes which are distributed and stored on individual HSMs in the HSM fleet. Parity information is generated which allows one or more stripes to be regenerated if one or more stripes becomes corrupt or is lost. The parity information may be stored on an HSM in the HSM fleet, or outside the fleet on a storage service, HSM management hub, tangible computer-readable media, or other device. If an HSM in the HSM fleet fails, resulting in the loss of a stripe, an HSM in the fleet can recover the missing stripe by re-creating the missing stripe from the remaining stripes combined with the parity information. In some examples, stripes are mirrored within the fleet of HSMs.

Abstract: Cryptographic keys are durably stored for an amount of time. A cryptographic key is encrypted so as to be decryptable using another cryptographic key that has a limited lifetime. The other cryptographic key can be used to decrypt the encrypted cryptographic key to restore the cryptographic key during the lifetime of the other cryptographic key. After the lifetime of the other cryptographic key, if a copy of the cryptographic key is lost (e.g., inadvertently and unrecoverably deleted from memory), the cryptographic key becomes irrecoverable.

Abstract: A proof-of-work system where a first party (e.g., a client computer system) may request access to a computing resource. A second party (e.g., a service provider) may determine a challenge that may be provided to the first party. A valid solution to the challenge may be generated and provided for the request to be fulfilled. The challenge may include a message and a seed, such that the seed may be used at least in part to cryptographically derive information that may be used to generate a solution to the challenge. A hash tree may be generated as of generating the solution.

Abstract: A trusted co-processor can provide a hardware-based observation point into the operation of a host machine owned by a resource provider or other such entity. The co-processor can be installed via a peripheral card on a fast bus, such as a PCI bus, on the host machine. The co-processor can execute malware detection software, and can use this software to analyze data and/or code obtained from the relevant resources of the host machine. The trusted co-processor can notify the customer or another appropriate entity of the results of the scan, such that an appropriate action can be taken if malware is detected. The results of the scan can be trusted, as malware will be unable to falsify such a notification or modify the operation of the trusted co-processor.

Abstract: A computing resource service receives a request. In response to the request, the computing resource service queries a probabilistic data structure for an entry corresponding to the request. The computing resource service obtains, from the probabilistic data structure, a value that corresponds to the entry. Based at least in part on this value, the computing resource service determines whether the entry has expired. If the entry is expired, the request is fulfilled. However, if the entry has not expired, the request is denied.

Abstract: A concordance service receives a probabilistic data structure query generated based at least in part on a set of query parameters for a search of a plurality of resources. In response to receiving the query, the concordance service uses the probabilistic data structure query and a probabilistic data structure tree to determine a set of nodes of the tree that individually satisfy the set of query parameters. The concordance service verifies that the resources corresponding to the set of nodes satisfy the query parameters. Based at least in part on this verification, the concordance service provides a response to the query.