1

UPDATE

I see that lots of people find my question too long (because there is lots to explain), read the first sentence and then just think that I'm going on the worst tangent possible without seeing the entire question. If the question isn't clear enough please let me know. I'm trying to condense it in the simplest way and not to cause any confusion.

The reason for the public key decryption is to achieve a form of digital signing where the recipient decrypts the encrypted content to reveal a hash value. I didn't see the need to mention this in the question as I wanted to find out how to perform this operation in its basic form. However to avoid any further concerns and warnings around what RSA is all about and that public key decryption is bad, I updated my question with that disclaimer.


BACKGROUND

I have written a C# application that uses the Chilkat's RSA library to take content and encrypt it using a personal Private Key.

Then I would like to use a public website to allow someone to decrypt that very content (that's encrypted) by using an associated public key.

Now, I found a 3rd party website (and there are not a lot of them, BTW) that allows you to decrypt content using a RSA public key (https://www.devglan.com/online-tools/rsa-encryption-decryption).

Unfortunately when I try to use it, I get a "Decrypt error".

Here is a sample setup. I have generated my own personal Public & Private Key pairs. In my C# application, I'm taking a string and encrypting it with a private key and encoding it using Base64.

const string originalContent = "This !s original c0nt3nt";

var rsa = new Chilkat.Rsa();
rsa.GenerateKey(2048);

var encryptedBytes = rsa.EncryptBytes(Encoding.UTF8.GetBytes(originalContent), true);
var encryptedEncodedString = Convert.ToBase64String(encryptedBytes);
Console.WriteLine($"Encrypted:{Environment.NewLine}{encryptedEncodedString}");
Console.WriteLine();

var privateKeyBytes = rsa.ExportPrivateKeyObj().GetPkcs8();
var privateKeyEncodedString = Convert.ToBase64String(privateKeyBytes);
Console.WriteLine($"Private Key:{Environment.NewLine}{privateKeyEncodedString}");
Console.WriteLine();

var publicKeyBytes = rsa.ExportPublicKeyObj().GetDer(false);
var publicKeyEncodedString = Convert.ToBase64String(publicKeyBytes);
Console.WriteLine($"Public Key:{Environment.NewLine}{publicKeyEncodedString}");
Console.WriteLine();

var decyptedContentBytes = rsa.DecryptBytes(encryptedBytes, false);
var decryptedContentString = Encoding.UTF8.GetString(decyptedContentBytes);
Console.WriteLine($"Decrypted:{Environment.NewLine}{decryptedContentString}");
Console.WriteLine();

Console.WriteLine("Press ENTER to quit");
Console.ReadLine();

This sample console app will write out all the necessary information needed to use for the next part of the process and to demonstrate that in principle it works as expected.

Example: Console output

Here are the sample values from the console window:

ENCRYPTED CONTENT

H5JTsGhune1n3WWSPjwVJuUwp70Hsh1Ojaa0NFCVyq0qMjVPMxnknexOG/+HZDrIYsZM7EnPulpmihJk4QyLM8T2KNQIhbWuMHvzgHYlcPJdXpGZhAxwfklL4HP0iRUUXJBsJcS/2XoUDZ6elUoMIFY9cDB4O+WFxKS/5vzLEukTLbQ3aEBNg3xaf9fg12F8LcMxZ3GDsk0W9b6oJci09NTxXd6KKes0RM1hnOhw6bu0U33ZLF3sa0nH9Kdf8w23PoKc/tl12Jsa8N1A4OjaT5910UF8FRH6OkAbNKnxqXcL7+V4HVuHchi3ghuFivAW57boLeHr7OG7wOEC/gfPOw==

PRIVATE KEY

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

PUBLIC KEY

MIIBIjANBgkqhkiG9w0BAQEFAAOCAQ8AMIIBCgKCAQEAvMWHDV3JF39V+CneqhOXO6XXX1nSqO9ZnMGVrJLmqvsKtxU3AkwHnaP1XMfFcZ/kTJ+VxO0Ag83AADdCvwc0LKFPNZskc2fmyj/kyeRx9IwpQlDI/onXhJtH2svO9452xtIShDiT1cT1sCqt47wyy5lq3YL3f1Ig0rQ5tvOkhSnzJS1pEsPGCHOZ17LYQID9Ib/DY+1Ayr0xBuaN26E7oRtSUAEQhFz9H03SIr9HBBZUjdypIBeCg399lbdseezf/jyV6UL3BApC8D6QQKv2jNKm7YHoiC8Wr3wSUcEOc1w1JmR3qb0oyuiiGELE2r0ECw15xSrl2tn8TiRyJcrOqQIDAQAB

WEBSITE PUBLIC KEY DECRYPTION ATTEMPT

Now, when I go to the website (mentioned above) I paste in my encrypted content in the encrypted content text block and I paste my public key that I generated in the text block underneath it and set the RSA Key Type to Public Key. But it fails.

My attempt to let the website decrypt my encrypted content


SECOND ATTEMPT

But...

I have done some troubleshooting by taking my personal Private & Public keys that I generated and I use the website to perform the encryption & decryption with my keys and I'm able to encrypt my string and decrypt that encrypted content successfully which leads me to believe that somehow my Chilkat encryption setup is not fully aligned with the one that the website uses.

Have the website test to make sure my keys are working by encrypting and decrypting content


WHAT I NOTICED

So I started reading what the website had to offer and the author of the page posted an explanation on how to accomplish this (https://www.devglan.com/java8/rsa-encryption-decryption-java) which uses the Java RSA libraries under the hood. Apparently, there are two Java RSA ciphers that can be used "RSA" and "RSA/ECB/PKCS1Padding".

I am not so familiar with the Java libs and I know enough of cryptography to know how to get things done but there are lots of technical aspects that are still unclear to me as to help me figure out where to go next.


QUESTION

My question is, is there anything in Chilkat that I need to setup so that it can encrypt content that would allow a Java application (like the website link posted above) to be able to decrypt? (of course Chilkat needs to be able to decrypt it as well)

Dandré
  • 2,053
  • 3
  • 18
  • 38
  • Your question is unclear. First you say decryption returns an error, and you show the image for that. Then you say decryption works, and you show the image for that. So which is it? Also, why not provide your decryption code and example private keys and inputs that exhibit the error? – President James K. Polk Nov 01 '19 at 13:46
  • Apologies for any confusion and for writing the original in a hurry. Thank you for your advise and suggestions. It has been rewritten with lots more information. I hope everything is more clear. – Dandré Nov 01 '19 at 20:45
  • It is crucial to be able to verify a digital signature (decryption the signature with the public key) using external tools. It is okay that Chikat can encrypt and decrypt successfully but that is not enough, what it encrypts another cryptographic service must be able to decrypt to prove the validity of the signature. Interested in the outcome, can someone please assist? – Marius Vorster Nov 06 '19 at 04:05

2 Answers2

0

I'm going to (hopefully) answer this question after only reading the 1st part of it. I got to the point where you wrote "... I'm taking a string and encrypting it with a private key ...", and this raised the red flag.

Public key encryption should be where you encrypt using the recipient's public key. The private key is used to decrypt. The point of encryption is that only the intended recipient can decrypt and view the message. With public/private key pairs, you can provide your public key to anybody, but you are in sole possession of your private key. Therefore, anybody can use your public key to encrypt a message intended for you, but you are the only one who can decrypt. This makes sense.

Signing is the opposite: You use your private key to sign, and anybody can verify using your public key. A signature can optionally contain the signed data, so that the act of verifying the signature also extracts the original data. Thus, you verify that (1) the data could only have been signed by the holder of the private key, (2) the data was not modified, and (3) you recover the original data.

Chilkat's API provides the ability to use the public/private keys in the opposite way, which doesn't make any sense, but was needed because there are systems "out there" that do things that make no sense, and Chilkat was needed to perform the opposite. (It makes no sense to encrypt something that anybody can decrypt.)

I think the code behind the devglan website is not capable of doing the RSA encrypt/decrypt in the opposite way. You would need to encrypt using your public key, and then give your private key to the other person.

Or.. you could instead create an "opaque signature" using Chilkat, which is a signature containing the data, and then find the devglan online tool to verify/extract the data from the PKCS7 signature (if the devglan tool exists). This way you can keep your private key and give the public key to the recipient.

Finally.. it seems to me that you're really treating the public/private key as a shared secret -- i.e. a secret only shared between sender and receiver. In that case, why bother using RSA at all? (Remember, RSA is only for encrypting/decrypting small amounts of data. The max number of bytes you can encrypt is equal to the key size minus some overhead. So if you have a 2048-bit key, then you can maximally encrypt 2048/8 bytes minus the overhead used in padding, which is on the order of 20 bytes or so.) If semantically you just have a shared secret, then you might simplify and use symmetric encryption (AES) where the secret key is just a random bunch of bytes and you have no data size limit.

Chilkat Software
  • 1,405
  • 1
  • 9
  • 8
  • `I think the code behind the devglan website is not capable of doing the RSA encrypt/decrypt in the opposite way. You would need to encrypt using your public key, and then give your private key to the other person.` At the middle hald of my question you would see me achieve it the other way around using the keys that are generated with Chilkat – Dandré Nov 02 '19 at 15:10
  • Does it make more sense now? – Dandré Nov 02 '19 at 20:50
0

A message that has been encrypted with the private key using the Chilkat-library cannot be decrypted with the public key using Java (at least not with the standard SunJCE-provider) or the java-based web-site, since different padding variants are used on both sides.

The prerequisite for a successful decryption is that both encryption and decryption use the same padding variant. The same applies to signing and verification.

There are two variants of the PKCS1-v1.5-padding described in RFC8017: One is RSAES-PKCS1-v1_5, which is used in the context of encryption and decryption, and the other is RSASSA-PKCS1-v1_5, which is used in the context of signing and verifying. RSAES-PKCS1-v1_5 is non-deterministic, i.e. repeated encryption of the same plaintext with the same key always generates different ciphertexts. RSASSA-PKCS1-v1_5 is deterministic, that is, it always generates the same ciphertext under the mentioned conditions.

Since the padding variant depends on the respective platform/library, a general statement is not possible. However, for the Chilkat-library and Java (standard SunJCE provider) the following applies (PKCS1-v1.5-padding assumed):

  • The methods that Chilkat provides in the context of encryption/decryption use RSAES-PKCS1-v1_5 regardless of whether the public or private key is used for encryption. Analog methods also exist in the context of signing/verifying. These use RSASSA-PKCS1-v1_5. To check this, the padding variant can be determined by setting the Chilkat.Rsa#NoUnpad flag to true, so that the padding is not removed during decryption. Another option for a test is to repeatedly encrypt the same plaintext with the same key. Since RSAES-PKCS1-v1_5 is probabilistic, different ciphertexts are generated each time.

  • In Java, the Cipher-class determines which padding variant is used based on the mode (encryption or decryption) and the key type used (private or public). For encryption with the public key and decryption with the private key, RSAES-PKCS1-v1_5 is used. For encryption with the private key and decryption with the public key, RSASSA-PKCS1-v1_5 is used. For signing/verifying, Java provides the Signature-class which uses RSASSA-PKCS1-v1_5. To check this, proceed as described above. In Java, you can prevent the padding from being removed with RSA/ECB/NoPadding during decryption.

Since in the context of encryption/decryption the public key is used for encryption and the private key is used for decryption, and dedicated classes or methods are used in the context of signing/verifying, there are no or few use cases for direct encryption with the private key and decryption with the public key. Furthermore or maybe because of that these processes are not uniformly implemented in the libraries as you can see in the example of the Chilkat-library and Java.

Altogether three cases can be distinguished for the Chilkat-library and Java:

  • Within the same library/language, encryption can be performed with the public or private key and decryption with the respective counterpart. For this reason the encryption and decryption on the web site (using Java) works in the posted example Second Attempt: Both the encryption with the private key and the decryption with the public key use RSASSA-PKCS1-v1_5.
  • If in the Chilkat-code the public key is used for encryption and in Java the private key is used for decryption, RSAES-PKCS1-v1_5 is used for both encryption and decryption, which is why decryption works.
  • However, if in the Chilkat-code the private key is used for encryption and in Java the public key is used for decryption, RSAES-PKCS1-v1_5 is used for encryption and RSASSA-PKCS1-v1_5 is used for decryption. Both padding variants therefore differ and decryption fails. This corresponds to the scenario described in the question.

After this explanation now to your question: My question is, is there anything in Chilkat that I need to setup so that it can encrypt content that would allow a Java application (like the website link posted above) to be able to decrypt? Since the Java-code uses RSASSA-PKCS1-v1_5 for decryption with a public key, it would be necessary for compatibility to change the padding variant in the Chilkat-code from RSAES-PKCS1-v1_5 to RSASSA-PKCS1-v1_5 in the context of encryption/decryption. If you look at Chilkat's RSA-methods, it seems that this is not intended, but that the logic for determining the padding variant is hard coded (as probably with most libraries). You can only choose between PKCS1-v1.5-padding and OAEP for padding. This means that a message encrypted with the private key using the Chilkat-code cannot be decrypted with the public key in Java or on the website.

What are the alternatives? According to the question, the goal is: The reason for the public key decryption is to achieve a form of digital signing where the recipient decrypts the encrypted content to reveal a hash value.

  • Here it would be a good idea to create a standard signature on the Chilkat-side, e.g. with signBytes. The hash of the data is created automatically and RSASSA-PKCS1-v1_5 is used as padding variant (if the data are already hashed, the method signHash can be used). On the Java-side, this signature can be verified. Alternatively, the signature can be decrypted with the public key, which allows the hash value to be determined, since Java uses the padding variant RSASSA-PKCS1-v1_5 in both cases. Decryption is also possible on the web site, but the decrypted data are not displayed properly because they are only given as a string (which does not produce any meaningful output because of the arbitrary byte-sequences in a hash) and the encoding cannot be changed to hexadecimal or Base64.
  • Another possibility might be to use Chilkat on the Java-side as well. Probably Chilkat uses a uniform logic across platforms (which I didn't verify however).
Community
  • 1
  • 1
Topaco
  • 40,594
  • 4
  • 35
  • 62
  • Topaco, thanks so much for your answer. Even though its not what I was hoping, it was very comprehensive and lead me into the correct direction and for that I'm grateful. It explains so much. I actually tried to use the Chilkat Sign operation and on the website it actually decrypts the signature but in some sort of binary representation that doesn't display well (as expected but it was to just test what you've mentioned) but it proves that the Java Cypher class was able to decrypt it. – Dandré Nov 18 '19 at 09:23
  • 1
    So unfortunately this means that using online tools to verify won't work. So what I'll do instead is potentially find an off-line method for signature verification (like OpenSSL commands) and update my application to rather use the standard Sign/Verify operations instead. – Dandré Nov 18 '19 at 09:25
  • 1
    I agree with that, with a little note: The _verification of a signature_ created with Chilkat or Java is no problem, even _online_, e.g. [here](https://8gwifi.org/rsasignverifyfunctions.jsp). What might not (or only with difficulty) be possible is to _decrypt_ the signature _online_ in order to determine the hash, because this requires a website that allows decryption with the _public_ key, uses the correct padding and displays the result in a suitable encoding (such as Base64 or hexadecimal). The devglan-website comes very close to this and fails only because of the last mentioned condition. – Topaco Nov 18 '19 at 12:39
  • I forgot about that website. I didn't know that they had signature verification tools. I've tested it with the same test data that I used to verify with OpenSSL and it is working. You're a star! Thanks for all the help! – Dandré Nov 18 '19 at 12:48