This application claims the benefit of and is a non-provisional of co-pending: U.S. Provisional Application Ser. No. 60/962,848 filed on Jul. 31, 2007; U.S. Provisional Application Ser. No. 61/026,438 filed on Feb. 5, 2008; U.S. Provisional Application Ser. No. 60/962,821 filed on Jul. 31, 2007; and U.S. Provisional Application Ser. No. 60/962,822 filed on Jul. 31, 2007; which are all hereby expressly incorporated by reference in their entirety for all purposes.
This application expressly incorporates by reference: U.S. Application Ser. No. ______, filed on an even day herewith, entitled “INPUT OUTPUT ACCESS CONTROLLER” (temporarily referenced by Attorney Docket No. 017018-017210US/VS-0245); and, U.S. Application Ser. No. ______, filed on an even day herewith, entitled “TRUSTED LABELER” (temporarily referenced by Attorney Docket No. 017018-014610US/VS-0246); in their entirety for all purposes.
BACKGROUNDThis disclosure relates in general to secure computing systems and, more specifically to high-assurance access to keys at different classification levels amongst other things.
Governments classify information at different levels generally according to their sensitivity, for example, SECRET versus TOP SECRET. Users of the information are also classified by what level they are able to get access to. For example, someone with a SECRET clearance is not given access to TOP SECRET information. Procedures are put in place to avoid exposure to persons without the proper classification level.
In processing systems, physical security is used to prevent information of different classification levels from bleeding over to a different classification level. To process at multiple classification levels, there may be several devices running in parallel for each classification level. Devices that may be capable of running at multiple classification levels are run at one classification level, cleared out and then run at a different classification level. Intermixing of different classified information is generally taboo in these systems.
There are situations that require smaller cryptographic devices that can process different classification levels. Switching between classification levels takes time and slows down processing. Some have proposed trusted operating systems that can process information with more flexibility, but these solutions are avoided due to a lack of trust.
Different keys are required for each classification level to maintain the security of information in each classification level. The keys may be simply different values or could be used with different algorithms. Even if security is breached for one classification level, the unique keys and algorithms can keep information protected in the other classification levels safe.
SUMMARYIn an embodiment, a cryptographic device and method are disclosed for processing different levels of classified information. A memory caches keys for use in a cryptographic processor. The cryptographic processor requests a key associated with a particular classification level when processing a packet of the particular classification level. The cryptographic device confirms that the key and the packet are of the same classification level in a high-assurance manner. Checking header information of the keys one or more times is performed in one embodiment. Some embodiments authenticate the stored key in a high-assurance manner prior to providing the key to the cryptographic device.
In one embodiment, a cryptographic device for processing classified information having a number of different classification levels is disclosed. The cryptographic device includes a memory, a cryptographic processor and a key manager. The memory holds a number of keys outside of an integrated circuit. The plurality of keys are for the plurality of different classification levels. The cryptographic processor is part of the integrated circuit and uses the plurality of keys to process packets of information that are categorized according to the number of different classification levels. The key manager can access a plurality of rules associated with the plurality of different classification levels that regulate interaction with the plurality of keys. A first rule of the number of rules is used by the key manager in a first classification level of the number of different classification levels. A second rule of the number of rules is used by the key manager in a second classification level of the number of different classification levels.
In another embodiment, a method for processing classified information in a high-assurance manner is disclosed. In one step, a request is received for a first key by a cryptographic processor. A first rule from a number of rules is applied to a first sterile key retrieved from a memory. The first sterile key is decrypted with a first protection key to produce the first key. The first key is also checked with the first rule. The first key is provided to the cryptographic processor if the checking the first sterile key step and the checking the first key step are completed successfully. A request is received for a second key by the cryptographic processor. A second rule from the number of rules is applied to a second sterile key retrieved from the memory. The second sterile key is decrypted with a second protection key to produce the second key. The second key is also checked with the second rule. The second key is provided to the cryptographic processor if the checking the second sterile key step and the checking the second key step are completed successfully.
In yet another embodiment, a cryptographic device for processing information with a plurality of classification levels is disclosed. The cryptographic device includes a memory, a cryptographic processor and a key manager. The memory holds a number of keys that are used by a cryptographic processor to process packets of information that are correlated to the plurality of classification levels. The key manager includes a rule enforcement circuit and a key decryption circuit. The key manager retrieves a first key for a first packet being processed by the cryptographic processor. The first packet and the first key are of a first classification level. The rule enforcement circuit checks that the first key is designated for the first classification level before providing the first key to the cryptographic processor for processing the first packet. The key manager retrieves a second key for a second packet being processed by the cryptographic processor. The second packet and the second key are of a second classification level. The rule enforcement circuit checks that the second key is designated for the second classification level before providing the second key to the cryptographic processor for processing the second packet.
Further areas of applicability of the present disclosure will become apparent from the detailed description provided hereinafter. It should be understood that the detailed description and specific examples, while indicating various embodiments, are intended for purposes of illustration only and are not intended to necessarily limit the scope of the disclosure.
BRIEF DESCRIPTION OF THE DRAWINGSThe present disclosure is described in conjunction with the appended figures:
FIG. 1 depicts a block diagram of an embodiment of a cryptographic device;
FIG. 2 depicts a block diagram of an embodiment of a partitioned key cache;
FIG. 3 depicts a block diagram of an embodiment of a key manager;
FIG. 4 depicts a diagram of an embodiment of a key unscrambling process; and
FIG. 5 illustrates a flowchart of an embodiment of a process for operating the cryptographic device.
In the appended figures, similar components and/or features may have the same reference label. Further, various components of the same type may be distinguished by following the reference label by a dash and a second label that distinguishes among the similar components. If only the first reference label is used in the specification, the description is applicable to any one of the similar components having the same first reference label irrespective of the second reference label.
DETAILED DESCRIPTIONThe ensuing description provides preferred exemplary embodiment(s) only, and is not intended to limit the scope, applicability or configuration of the disclosure. Rather, the ensuing description of the preferred exemplary embodiment(s) will provide those skilled in the art with an enabling description for implementing a preferred exemplary embodiment. It being understood that various changes may be made in the function and arrangement of elements without departing from the spirit and scope as set forth in the appended claims.
Referring first toFIG. 1, a block diagram of an embodiment of acryptographic device100 is shown. Thecryptographic device100 processes information of different classifications. Information in each classification level is kept separate or partitioned from information of other classification levels throughout thecryptographic device100. Additionally, each classification level can use different cryptographic algorithms and/or keys. Several integrated circuits could be used to implement thecryptographic device100 where at least acryptographic processor120 is in one integrated circuit and the partitionedkey cache116 is in another. Other embodiments could have the partitionedkey cache116 andcryptographic processor120 as part of the same integrated circuit.
Acryptographic processor120 is the circuit that performs encryption, decryption and/or bypass for the information that passes through it. Information may be a stream or packetized incryptographic processor120. The packets or streams are of different classification levels. Thecryptographic processor120 can reconfigure itself for the appropriate processing on a packet-by-packet basis. Different processing steps are set up for each classification level in a pipeline fashion by thecryptographic processor120. The processing steps perform formatting and cryptographic processing with a number of different algorithms and/or keys. Some of these processing steps can be common to multiple classification levels such that thecryptographic processor120 can potentially reuse the sub-circuits performing processing steps for the multiple classification levels.
Asystem bus124 allows theprocessor108 to communicate with akey manager104, a partitionedkey cache116, an input/output (IO)access controller112, and other peripherals that are not shown in the figure. Theprocessor108 communicates with thekey manager104, via theIO access controller112, to access the partitionedkey cache116. Keys are loaded into the partitionedkey cache116 by theprocessor104 in this embodiment and read by thecryptographic processor120. Other embodiments could load the keys from an external source, for example.
TheIO access controller112 checks that theprocessor108 or anything else using thesystem bus124 is operating as expected. Theprocessor108 writes a state to theIO access controller112. Each state is able to access peripherals defined by addresses or ranges of addresses. TheIO access controller112 checks that only the designated addresses are accessed by theprocessor108 in a given state to assure that the interaction with thekey manager104 and partitionedkey cache116 to write keys is authorized. Each address or range of addresses can be designated for read only, write only or read and write accessible. TheIO access controller112 further understands how states transition through the state machine such that state transitions are also checked when theprocessor108 is using thesystem bus124.
There are states defined exclusive to the various classification levels. Thecryptographic processor120 performs certain key operations in certain states. TheIO access controller112 checks if an address range of a partitionedkey cache116 has been properly written by the processor based upon the current state. Table I gives an example of the states used for various classification levels. Additionally, the algorithm and key address is given in the table. These entries in the table serve as rules. For example, states four, seven and nine operate in a CONFIDENTIAL classification level using a DES cryptographic algorithm and the key at address forty-three in the partitioned key cache. TheIO access controller112 in this example would make sure the current states were one of states four, seven or nine when theprocessor108 writes the key at address forty-three in the partitionedkey cache116. Where a violation were determined theIO access controller112, the key could be zeroized and/or other remedial action could be taken.
| TABLE I |
|
| State Enforcement Rules |
| State(s) | Classification | Algorithm | Key Address | |
| |
| 1 | TS | AES 256 | 11 |
| 2, 19 | S | AES 196 | 5 |
| 3, 21 | NC | Triple DES | 21 |
| 4, 7, 9 | C | DES | 43 |
| |
As a packet passes through thecryptographic processor120 certain algorithms use keys to perform the desired processing for that packet. The partitionedkey cache116 holds keys in a sterile or encrypted form. Sterilization puts the keys in a form that protects the encapsulated key even if recovered improperly. Various encryption algorithms could be used for sterilizing the keys. Akey manager104 is capable of deriving the key from the sterilized version. The partitionedkey cache116 could use dynamic random access memory (DRAM) or static random access memory (SRAM). This embodiment uses volatile RAM for the partitionedkey cache116 that is in a separate integrated circuit, but could be integral with the integrated circuit of thekey manager104 and/orcryptographic processor120. The partitionedkey cache116 could be a segment of a larger memory used for other purposes in other embodiments.
Akey manager104 receives requests from thecryptographic processor120 for keys used to process the various packets. Thekey manager104 checks the requests, retrieves the sterilized key, reconstitutes the key, performs checks, and returns the key to thecryptographic processor120. Thekey manager104 is implemented in logic that is not reprogrammable during normal operation. To load keys into the partitionkey cache116, theprocessor108 interacts with thekey manager104 under the supervision of theIO access controller112.
With reference toFIG. 2, a block diagram of an embodiment of a partitionedkey cache116 is shown. The partitionedkey cache116 has a number of partitions204 defined. Those partitions map to peripherals or address ranges used by theIO access controller112. A given classification level stores its keys in one partition204 and is prevented from accessing other partitions204. Although the partitionedkey cache116 is a single integrated circuit with a common interface in this embodiment, the partitioning enforces a logical separation in a high-assurance manner. Table II gives an example of the mapping between state, classification and partition204. These mappings serve as rules. For example, state one operates at a TOP SECRET classification level and has access to partition B204-2, which includes addresses eleven through twenty.
| TABLE II |
|
| State to Partition Mapping |
| State(s) | Classification | Partition Addresses |
| |
| 1 | TS | 11-20 |
| 2, 19 | S | 1-10 |
| 3, 21 | NC | 21-30 |
| 4, 7, 9 | C | 31-45 |
| |
Referring next toFIG. 3, a block diagram of an embodiment of akey manager104 is shown. In a high-assurance manner, the key manager enforces logical separation of the partitionedkey cache116 where no physical separation of the interface to the partitionedkey cache116 exists. Akey request interface332 is coupled to thecryptographic processor120 to receive requests for one or more particular key located at specified addresses in the partitionedkey cache116.
This embodiment includes arequest validator336 that checks the request to make sure the request is formatted correctly. Other embodiments of therequest validator336 could check that the classification level of the requesting packet matches the classification level of the partition204 of the requested key. With a specified key location, a key memory interface132 couples to the partitionedkey cache116 to retrieve the requested key. The sterile key is returned to akey buffer308.
Arule enforcement circuit336 includes a sterilekey validator304, akey decoder324 and a reconstitutedkey validator328. Multiple levels of checks are performed on the key before the key is provided to thecryptographic processor120.
The sterilekey validator304 checks the classification level of the sterile key against the classification level of the packet that precipitated the request of the key. Essentially, the sterile key is checked to make sure it matches the type of processing being performed in thecryptographic processor120 to provide high-assurance. The sterilekey validator304 could also match the algorithm and/or state as further rule checks in some embodiments.
A CRC, checksum or other validity value is appended to the sterile key in this embodiment when stored in the partitionedkey cache116. The software determines the validity value when writing the sterile key into the partitionedkey cache116. A check of the validity value allows conformation that information stored by theprocessor108 was delivered accurately to the sterilekey validator304.
Akey map database316 holds information to validate the keys in their sterile or reconstituted form. The state, classification, address, algorithm, key length, header information, and/or other information could be stored in thekey map database316. A look-up table is used in one embodiment of thekey map database316. Where several keys are used for a given classification level, the finer granularity of state can confirm that the key is the correct one for a particular situation. For example, there could be a different state for each key. Other embodiments could provide granularity using two or more partitions204 for a particular classification level.
Akey decoder324 converts the sterilized key from the partitionedkey cache116 into a reconstituted key that is ready for use by thecryptographic processor120. A payload of the sterilized key is decrypted to produce a reconstituted key. A cache protectionkey store320 holds a cache protection key for each classification level. Table III shows an example of the information stored in the cache protectionkey store320. Other embodiments could have a different cache protection key for each sterilized key or one cache protection key for all keys.
| TABLE III |
|
| Cache Protection Keys |
| State(s) | Classification | Random Key | |
| |
| 1 | TS | 19A5E9F45609DC90h |
| 2, 19 | S | AA5119A456870190h |
| 3, 21 | N | 78A5E49B56A093D0h |
| 4, 7, 9 | C | 15E456894309AE9F0h |
| |
Within the decrypted payload of the sterilized key is information that is checked by a reconstitutedkey validator328. Additionally, a CRC, checksum or other validity value is embedded in the decrypted payload as a second validity value. The reconstitutedkey validator328 also checks the second validity value. This additional check provides a further layer of high-assurance. Should the process pass all of its checks, the decrypted key is provide to thecryptographic processor120 for use in processing the particular packet that requested the key.
With reference toFIG. 4, a diagram demonstrating an embodiment of akey unscrambling process400 is shown. Asterile key416 includes a sterile key header, a sterile key payload and a sterile key CRC, which serves as a validity value. All this information is stored in the partitionedkey cache116 by theprocessor108. The sterile key header holds the classification level, the applicable encryption algorithm(s), and any additional key identifiers. Some embodiments may also include the memory address of the sterile key. The sterile key CRC is a validity value that is calculated on the whole sterile key header and sterile key payload such that any corruption can be discerned.
The sterile key payload is exclusive-ORed324 with a cache protection key420 from the cache protectionkey store320 to decrypt the reconstituted key header, reconstituted key payload and reconstituted key CRC. Those items along with the sterile key header and sterile key CRC form thereconstituted key404. The reconstituted key payload is the actual key that will be used by thecryptographic processor120. The reconstituted key CRC is a validity value that allows checking that the fields of thereconstituted key404 has not changed. The exclusive-ORkey decoder324 is just one example of a simple decryption function. Other embodiments may use any type of decryption function(s).
Referring next toFIG. 5, a flowchart of an embodiment of aprocess500 for operating thecryptographic device100 is shown. The depicted portion of the process is initiated inblock504 where thecryptographic processor120 has a packet that uses a particular key, which is requested from thekey manager104. Some embodiments check the key request. In any event, the key is requested from the partitionedkey cache116 by thekey manager104. The address of the key requested falls within a particular partition204.
Thesterile key416 is retrieved inblock516 from the partitionedkey cache116 inblock516. The sterile key header is checked inblock520 to determine if the classification matches the classification of the packet requesting the key. Additional checks are possible, for example, the sterile key CRC or validity value could be checked. Presuming the check inblock520 is successful, processing continues to block524 where the sterile key payload is decrypted.
In thereconstituted key404, the reconstituted key header is checked inblock528. Prior to thedecoding block524, the reconstituted key header was scrambled. Additionally, a reconstituted key CRC or validity value can be checked in some embodiments. If the check inblock528 passes, processing continues to block532 where the key is returned thecryptographic processor120.
Should any of the checks fail inblocks520 or528, the partitionedkey cache116 is erased and/or the cache protectionkey store320 inblock536. Without the cache protection keys, the keys remain in a sterile form. Erasure of the cache protection keys can typically be performed much more quickly than the partitionedkey cache116. Further remedial action can be taken inblock540.
The above embodiments discuss processing at different classification levels. These classification levels could be government classification levels, but need not be necessarily so. A classification level is just a logical partition in the information passed. Any information that needs to be kept separate from other information could be in a separate classification level or logical partition.
While the principles of the disclosure have been described above in connection with specific apparatuses and methods, it is to be clearly understood that this description is made only by way of example and not as limitation on the scope of the disclosure.