physical security
In addition to the general theft, damage, natural disasters, man-made disasters and other physical damage to the information system entities, there are also the following information leakage through physics:
- Electromagnetic leakage: The electromagnetic radiation emitted by the device may be analyzed and restored by the highly sensitive receiving device, resulting in information leakage. The input and output signals and their transformations are called the core red signal, the control signal that may cause the leakage of the core red signal is called the key red signal, and the related unit circuit is called the red zone.
- Eavesdropping: Eavesdropping is often theft of information on a physical level. Including wired, wireless, infrared, laser (via laser detection of vibrations on objects caused by speech), satellites, etc.
Physical isolation:
Generally, logical isolation should be established between the public network and the external network, and physical isolation should be established between the internal network and the external network.
Logical isolation and physical isolation:
Logical isolation is based on the premise of normal use of the network, while physical isolation is based on the premise of absolute security of the information system.
Methods of physical isolation of the network:
- Not connected to extranet
- Time-sharing connection to the intranet and extranet (physical isolation on the client side)
- The server realizes data filtering and isolation, so that the internal and external networks are separated and processed.
Fault tolerance: Do not start with mistakes, but start with the system itself to improve the system's ability to resist errors.
A system with high fault tolerance is required: a system with high cost and high versatility; a long-life system (aerospace system, which cannot be manually repaired); a delayed maintenance system (which can be recovered without repairing if it is broken); a high-performance system (automatic recovery) , increase the trouble-free time); mission-critical systems (don't drop the chain at critical moments);
Fault-tolerant technology: (the first three are based on backup)
- Set idle device, top when error occurs (need to backup at any time)
- Mirroring: run two identical devices at the same time, one fails and the other continues;
- Recurrence: Delay mirroring, receive the calculation results of the source machine, delay updating the data of the auxiliary machine, the source machine fails, and the auxiliary machine is on top of the recording point.
- Load balancing: reduce the overload situation on one machine, improve stability, part error, and another part can do something else.
Authentication
Classification:
- password-based authentication
- Password-Based Authentication
Authentication based on biometric information
static authentication
- Dynamic Authentication (SMS, Voice)
Authentication protocol:
The specific rules recognized and complied with by both parties of authentication, the core of which is cryptography.
(1) Authentication protocol based on symmetric key:
Authentication protocol based on challenge-response method:
both parties to the authentication have their own shared secret key, and each user saves the object to be authenticated and its own shared secret key, which is used in small networks with a small number of people.
Process:
- A request communication
- B wants to verify the identity of A and sends it a piece of data, asking it to encrypt
- After A encrypts with the shared key, it returns to B
- B receives the ciphertext and decrypts it. If the decryption is successfully restored, it means A
(2) Needham-Schroeder authentication protocol:
A reliable third-party authentication server (AS) is added, and the AS stores each user's information and its shared secret key. So the users here only trust each other with the AS.
Process:
1.A向AS申请和B进行通话，发送一个自己设置的用于验证AS的信息S和IDa，IDb。
2.AS用和A的密钥Ka把回复信息加密，信息包括回话密钥K，之前的信息S,以及A之后要递交给B的用B的密钥加密了的信息包。
3.A接到AS的回信之后，解密，得到S，这就完成了AS的认证。之后保留K，把B的信息包发送给B。
4.B接到信息包，用自己和AS的共享秘钥Kb解密，得到回话密钥KS和IDa；B希望验证A的身份，他希望A有同样的回话密钥K，所以发送一个用K加密了的信息SS，挑战A
5.A接到挑战，解密之后对SS做函数处理F(SS)（约定好的）,再加密，这样比直接解密更加隐秘有效。
Only one key distribution center KDC of AS is used here.
Vulnerability: Assuming that the attacker has an old session key, it can be sent to B. B has no way of knowing that this is a retransmission attack (the result without a timestamp), assuming that the attacker then intercepts the handshake in the fourth step (because B Still want to send a message to A), you can use KS to decrypt and verify, and let B trust.
(3) kerberos
Kerberos adds a ticket granting server (TGS) on the basis of AS, and adds a ticket mechanism. AS is responsible for sending TGS tickets, and TGS is responsible for other application servers' tickets. Provides third-party authentication for client and server applications.
- User C indicates to apply for TGS ticket (using IDtgs) and application time TS1 from AS, and provides his own IDc
- The AS returns the encrypted information of the shared password KC with C, including the ticket, the call key Ktgs of C and TGS, as well as the time, the validity period of the ticket, and the ID.
- C applies to TGS for the object V to be communicated, and provides the ticket, ID, and information encrypted with Ktgs, including its own website, identity and TS3.
- TGS obtains Ktgs from the ticket encrypted by the shared key between AS and TGS, and the relevant information of C and the validity of the ticket, so the information provided by C needs to be the same as the information provided in the ticket issued by AS, and TGS , TGS is like a user here, but it will get C's information directly from AS to match, instead of looking for C to challenge again. After TGS verifies the identity of C, it will give C a key for communicating with V and a ticket issued by TGS for V. So the TGS here is equivalent to the function before AS.
- C takes the ticket to V and sends his own information
- V will get the ticket encrypted with the key shared by himself and TGS, and inside there are the information of TGS to C, the information of the ticket and the password for communication with C. V also directly uses the comparison of the two information sources to confirm the identity of C without re-challenge. You only need to send a confirmation of the previous C's TS5, indicating that you have received it, and you have also disproved your identity.
Therefore, kerberos uses the ticket mechanism to replace the challenge-response mechanism, but it is also a third-party authentication mechanism, where AS is the third-party authentication between users and TGS, and TGS is the third-party authentication between users and servers, in which the user and TGS are authenticated. The shared key is dynamically issued by the AS.
Windows 2000 server:
Windows server series is a network operating system, that is, an operating system used to manage the network. User authentication when logging in is accomplished through the AS authentication symmetric key.
Users and controllers are not authenticated by passwords, but a cryptographic authentication method. Although the user uses a password to log in, the controller does not check whether the passwords are consistent, but uses the password as a key, allowing the user to use the password to encrypt the information, and then use the shared key with the user, that is, the password to encrypt the message. , check whether it is consistent, so this will not cause the password to be eavesdropped during the propagation process.
(2) Authentication protocol based on public key
(1) Needham-Schroeder public key authentication protocol
A two-way authentication without third-party notarization. The difference from the symmetric key is that the symmetric key is based on the mutual knowledge of the two parties who master the encryption and decryption keys, while the public key is different. The party holding the private key usually does not know who holds the public key. Therefore, if the information is sent when the former requests, then the recipient does not need to verify the identity. The challenge-response mentioned earlier is the plaintext request sent, but it must also be plaintext, because the recipient has many keys and cannot determine the decryption key to be used from the ciphertext.
There is a further consideration here, that is, the party requesting communication also wants to know that the other party really wants it, so it becomes two-way authentication. Generally it is clear here.
- A applies for the communication, attaches its own information and challenge, and encrypts it with B's public key.
- B decrypts, gets A's information, encrypts A's challenge and its own challenge with A's public key, and sends it back to A
- After A decrypts, sees his challenge and comes back, verifies B's identity, and then sends B's challenge to B, which can be encrypted with B's public key. Here encryption is for security, not authentication.
- After decrypting, B gets his own challenge and authenticates A's identity.
(2) Identity authentication protocol based on CA digital certificate
Previous public keys were distributed by individuals at will and obtained by individuals. A third party is added here, but instead of managing the secret information of the shared key and generating the session key, as before, it manages the public key. The public key managed by it is authoritative, and the public key will be bound to the holder's information as a digital certificate CA.
Certificate Application:
- Users submit their own information to the CA to apply for a certificate.
- After the CA has been reviewed, the CA will be generated and encrypted with the CA's own private key, while the CA's public key is open to anyone.
- A has its own private key and its own digital certificate, but the digital certificate is encrypted by the CA, so A cannot modify the digital certificate by itself, which is also a manifestation of binding and the authority of the CA.
Authentication:
- If A wants to communicate with B, it needs to provide its own digital certificate after sending the encrypted information.
- B gets the relevant information. If he wants to view the information, he first needs to have A's public key. Besides, B also wants to verify the identity of A. These are all in the digital certificate encrypted by CA, so B finds CA to get CA's public key and directly Get the public key and A's information. That is, A is authenticated, and information is also obtained. (Although even if the information is not bound, B uses A's public key to get the information, and A's identity can be determined) This is a one-way verification, and A does not consider whether it is sent to B.
Application example of identity authentication - public key infrastructure PKI:
here is an Internet key management basic platform of CA technology.
X.509 is the ISO certificate format standard, which mainly establishes the benchmark for PKI certificates. The management of its CA is achieved through a hierarchical structure such as a directory. That is, the certification authority is tree-like, and the public key of the root CA can be obtained directly from its CA, while the CA of the certification authority of the intermediate node is issued by the upper-level authority, and it can be issued by the subsequent authority. The leaf node certification authority is the CA that issues the registered user. Therefore, in order to obtain the CA of an end user, the CA of the root authority is required. Of course there can be a caching mechanism.