#Access control

Introduction

In today's rapidly evolving threat landscape, organizations deploying mission-critical applications must implement robust security architectures that protect sensitive data while maintaining high availability and performance. This is especially crucial for enterprises utilizing advanced database management systems like InterSystems IRIS, which often powers applications handling highly sensitive healthcare, financial, or personal data.

In the previous article, I had just started working with Arduino, and got a meteorological station to show as a result. In this article, let's go further: we will set up authentication via RFID cards and Arduino against the InterSystems Caché application.

This post is meant to provide a quick possible explanation for a very perplexing problem.

Scenario: You’ve just created your own administrative user in your 2014.1 (or later) instance of Caché. You gave it every possible security role (including %All), so it should in theory be able to do anything within the instance.

You’ve written a very advanced routine with a break command in it for debugging:

Why I decided to write this

Recently I had the challenge to create a secure authentication method to authorize access to some data, but unfortunately I had zero experience with those security configurations and I felt that I was missing some basic concepts to have a better understanding of the official documentation.

After studying and managing to deliver the classes that I was asked to develop, I'd like to share a little bit of my new knowledge, which helped me follow the topics in the documentation.

Have you ever thought of creating your own systems for editing users or, perhaps, even an API that you can call?

It's almost a year since I have published a series of articles explaining how to configure Cache instance as a client / resource server / authorization server. By that time, the implementation of OAuth 2.0 was still a pre-release software.

When a company is quite large and many different applications used by employees. But while those applications are mostly completely different, how to make it possible to not force users to enter credentials as many times as many applications they would like to use. The best way is to use SSO, so, it will be possible to have a portal, where users could launch any application used in a company. There are many different ways how to give access to your application by using the SSO mechanism, and some of them are:

• OAuth2
• Kerberos
• SAML

InterSystems already supports OAuth2 and can be quite easily deal with Kerberos. But I would like to discuss about using SAML (Security Assertion Markup Language).

The ideal number of table permissions to assign for your users is zero. Permissions should be granted upon sign-in based on the application used for access. For web applications, we have a simple way of doing this by appointing application roles, matching roles, and required resources in the System Management Portal.

ODBC and JDBC connections present a different problem, however, especially when third-party applications are involved. As providers of an ERP system, our customers often wish to be able to employ various software packages to integrate with or report on their data. Many of these programs are capable of running any kind of query. Yet, letting them do that can be devastating to a customer’s data.

We are building a bunch of rest based services using Ens 2016.2 to serve our browser based application (Angular 4).

Two questions:

1. The initial authentication seems only work if credentials are placed in the url parameters. Trying to use the Authorization header instead, the client code immediately complains about Access-Control-Allow-Origin. How can I resolve this?

2. After initial authentication, what is the proper way to send subsequent rest calls without having to include credential every time?

Hello community!

I am trying to set up Startup Tag^Routine field for the UnknownUser as follows:

And my simple routine is the next:

I'm trying to setup a new accesss role for the company support team to use the message viewer and production config page to trace the errors that eventually could occurr on the production integrations.

I've already assign the following privileges:

OAuth server to be deployed on the IRIS learning cloud platform. Clients - one on the other instance of the learning IRIS server, the other client locally on my computer in the container docker.

Both clients get a seemingly correct link (through ##class(%SYS.OAuth2.Authorization).GetAuthorizationCodeEndpoint()) to the login request form:

Does anyone have any sample code for updating Caché users (and specifically Full Name and Comments) via SQL?

Bonus points for knowing specifically what security resource needs to be assigned to allow someone to run the SQL.

Thanks in Advance!

Ben

This is more a WARNING than a question