If you work with the Azure Cloud as I do, keeping up with the service releases or looking at an overview of each service can be challenging. The Azure Periodic Table helps keeping up with these challenges.
A lot of us have been waiting for OWASP to publish a new set of top 10. Now that APIs are key in every solution, OWASP published the Top 10 Most Critical API Security Risks! It is worth the time to read it through. You don’t want to be in the news for the wrong reasons. #infosecurity#applicationsecurity
Yesterday I was wondering if Microsoft support middleware packages for Java to allow the typical resource provider actions in an access_token or id_tokens, similarly to what the OWIN NuGet packages do or the PassportJS libraries for NodeJS. These last two libraries act as middlewares intercepting the HTTP requests. They allow to, programmatically, parse the Authorization headers to extract Bearer tokens, validate the tokens, extract claims from the tokens etc, etc.
The libraries I had found so far, and that I was familiar with, were the MSAL set of libaries and the ADAL set of libraries. These are client side libraries, meant for applications acting or APIs acting as OAuth2 Clients (not as Resource Providers)
Spring Boot is a wrapper of Spring Framework libraries packaged with preconfigured components.
They can also work with AAD and Spring Security, but there aren’t many articles out there related to that framework and how to use it when AAD is the IdP providing JWT tokens except for this one:
There is also this old blog post with some sample code(using spring framework security, however the example is for illustration purposes only, and uses an access_token issued for a SPA client application in order to request access to an API, which is not exaclty the case of the application we’re trying to modernize (multi-page JSP web application)
Microsoft identity providers work with two types of libraries:
Client libraries: Native clients (iOS, Android), JavaScript SPA applications, and servers use client libraries to acquire access tokens for calling a resource such as Microsoft Graph or other APIs.
Server middleware libraries: Web apps use server middleware libraries for user sign-in. Web APIs use server middleware libraries to validate tokens that are sent by native clients or by other servers.
Microsoft developers produce and maintains two client Open Source libraries for Java that work with Microsoft identity providers => ADAL and MSAL
They support the industry standards OAuth2.0 and OpenID
Connect 1.0
There is also an oauth2-oidc-sdk for Java that contain the namespaces needed for token deserialization, token validation(s) and processing of claims, which is typically done server side, when the web app or api receives a bearer token in the HTTP(S) Security Authorization Header.
To my knowledge, this SDK is not maintained by Microsoft.
Note: Server side validation of the token, specifically, the decryption of the token digital signature and the comparison of the decrypted hash vs. the calculated hash is critical to ensure the token claims weren’t tampered in transit and that the IdP wasn’t spoofed. There are other token validations, but this one in particular guarantees the integrity of the information and the source of the token.
Azure AD B2C is one of the most fast growing Identity Providers in the world. When this type of tenant was created for social identities and digital citizens, the Microsoft Identity team didn’t anticipate its massive growth. Over 1 billion of users authenticate to their apps, and apps to apis, using this type of Azure directory or tenant. When privacy norms such as GDPR in the European Union and CCPA in California, USA, came about, the flexibility provided by custom policies; the white label customization of the html/css and the Identity Experience Framework, allowed solutions where the end users have more control of their data, including the ability to remove the personal data that applications collect from them (profile information, credentials, user attributes and permissions). Microsoft Graph also provides one of the first RESTful APIs that allow application developers to programatically perform CRUD operations on user accounts and principals associated to applications and services.
This month the Microsoft Identity team has published a number of public articles to guide developers on the automation steps with Azure AD B2C:
A colleague of mine that also specializes on Identity Management (IdM) with Identity Providers (IdP) such as AAD, Okta, ADFS, Ping Federate and IdentityServer, published this summary on the 6 different flows encompassed by the OAuth2.0 RFCs.
If you work with the OAuth 2.0 Authorization Framework and the OIDC Authentication Framework, you’ll find the article below very interesting.
One thing to remember when you study the Authorization Grants or Flows in OAuth2.0 is that neither the Extension Grant flow nor the Client Credentials flow require the end user to interact with a browser.
The Client Credentials Flow is designed for headless daemons/services where the is no human interaction. The Extension Grant Flow is sometimes named as OBO (on-Behalf-of) and can be used to exchange SAML2.0 tokens for OAuth2.0 access tokens/bearer tokens. This grant (Extension Grant) doesn’t necessarily require user interaction during the token exchange at the IdP for an authorized application if the token being exchanged is still valid. (OAuth2.0 Extension Grant) In this case the SAML2.0 bearer token with its assertions is exchanged for an access token/bearer token/JWT that represents the client application or service.
Mi niece is in middle school. Over FaceTime today she mentioned she was taking logarithm lessons in math. I was so happy she found them interesting!
She did ask though, “I don’t think I will ever use them though. Why do I need to know what a logarithm is to go buy groceries, for instance?”
I rolled my eyes and told her… well, if you go to the store and find a packaged meal that has a pH of 4 and the next similar meal has a pH of 3, you should know that’s a logarithmic scale with base 10 and the second meal is 10 times more acidic than the first one.
If you go to a rock concert and and the sound engineer is asked to pump the volume up and he increases 10dB, on the scale, that is equivalent to a 10-fold increase in sound intensity (which broadly corresponds with a doubling in loudness). Decibels (sound intensity) is also a logarithmic scale. You ears respond to sound in a logarithmic manner.
If you hear in the news a possible earthquake and they give the intensity in the Richter scale, you should know that, that is also a logarithmic scale. If they announce stronger aftershocks, you should know the Richter scale actually corresponds to the square root of 1000, which is a 31.6 times increase in the energy on the quake. If they announce a Category 5 earthquake and then a 6 one, that one will be 31.6 times stronger… and please, please, please, go to safety…
As a software publisher, there are two reasons to sign your code:
To prove its Integrity
To develop its Reputation
Best Practice: Time-stamping.
When signing code, as a publisher, you have the opportunity to timestamp your code. Time-stamping adds a cryptographically-verifiable timestamp to your signature, proving when the code was signed. If you do not timestamp your code, the signature can be treated as invalid upon the expiration of your digital certificate. Since it would probably be cumbersome to re-sign every package you’ve shipped when your certificate expires, you should take advantage of time-stamping. A signed, time-stamped package remains valid indefinitely, so long as the timestamp marks the package as having been signed during the validity period of the certificate
If this is the case, it should be safe to distribute the file (or execute the binaries) even after the original certificate used to generate the digital signature, expired. If the signature was not timestamped, then there is a risk that the file/executable will not pass verifications.
When a file is digitally signed and timestamped, the signature will not expire when the certificate expires. The public key accompanying the executable file will still be valid after the signing certificate expires.
This is how a file that is digitally signed and timestamped will look like in a Windows Client OS (properties of the file):
Example of the properties of a file that was digitally signed and timestamped.
Signing is there to prove that the assembly or file is created by a known person or company, and not changed by anyone else.
When signing an assembly, a hash is created and signed with the private key of the certificate private/public key pairs. That private key is not distributed with the assembly, it stays with the publisher, in this case Microsoft.
The public key of the certificate is, which is what you showed us, distributed with the executable or assembly. To verify that the assembly has not been tampered, you can calculate the hash using the same hashing algorithm specified on the properties of the assembly, for instance SHA1, and compare this hash with the encrypted hash that the assembly signature has, the latter hash should be decrypted with the certificate public key that is distributed with the assembly. If the hashes match, the assembly is the original one published by the publisher.
If the assembly (executable) is changed, for instance by malware, it is easy to find out by checking the hash of the assembly and its signature using the public key (which is included in the executable or file).
This is a verification that antivirus programs and anti-malware programs do.
Digital signing and timestamping is a layer of protection; however, it doesn’t 100% guarantee security, a bad publisher can still do the same and distribute files that are digitally signed and timestamped. The use of Certificate Authorities provide an additional level of trust on the publishers.
If the private key of the certificate used to digitally sign files, is compromised, other files and executables could, potentially, be signed on behalf of the publisher by a malicious attacker, hoping to distribute malware. This though, won’t affect files or executables signed before the breach happened when there is a timestamp. In this scenario, the Certificate Revocation Lists will also be essential to report to the world the blacklisted-invalid certificate.