I want the account to be able to use one app that requires administrative privileges. I have contacted the support team of the app to find out why it needs these privileges, but I didn’t receive any helpful information.

The app is for viewing surveillance footage, but it requires admin privileges to open. I don’t want to make every employee an administrator just for this one use case. It might be better to switch to a FOSS app that doesn’t require administrative privileges by default.

The cameras we currently use are made by the vendor of the app, so maybe we’re locked in somehow? The NVR is also made by them, so it might be possible, but I don’t know for sure. I need to look into it more.

  • derek@infosec.pub
    link
    fedilink
    arrow-up
    4
    ·
    edit-2
    6 days ago

    There’s some good advice in the comments already and I think you’re on the right track. I’d like to add a few suggestions and outline how I think about the problem.

    Ask if the vendor has installation administrator guides, whitepaper, training material, etc. If yes: ask that they send it to you. You may also be able to find these on the vendor’s website, customer portal, or a public knowledgebase / PDF repo.

    I would want to know three things.

    1. How do users authenticate through the application?
    2. What are all of the ways users may access the application (local only, remote desktop, LAN only, full server/client model)?
    3. Does the vendor have any prescribed solutions for defining who has access to the application, at what privilege level, with access to what features?

    i.e. What parts of the user access, authenticate, authorize pipeline do application admins or system admins have control over and how can we exercise that control?

    Based on some context I assume that the app is reading from Active Directory using RADIUS or LDAP for user auth and that people are physically logging into the machine.

    If this is the only method of authentication then I would gate the application with a second account for each employee who requires access for business reasons defined in their job description (or as close as you can get to that level of justification - some orgs never get there). You can then control who has access to the machine via group policy. Once logged in the user can launch the application with their second account (which would have the required admin access) via “Run as…” or whatever other methods you’d prefer. No local admins logging in directly and yet an application which users can launch as admin. Goal achieved.

    This paradigm lets us attempt balancing security concerns with user pain. The technically literate and daringly curious will either already know or soon discover they can leverage this privilege to install software and make some changes to the system. The additional friction, logging, and 1:1 nature of the account structure makes abusing this privilege less attractive and more easily auditable if someone does choose the fool’s path.

    I can imagine more complex set ups within these constraints but they require more work for the same or worse result.

    Ideally you run the app with a service account and user permissions are defined via Security Groups whose level of access is tied to application features instead of system privs. There are other reasonable schemes. This one is box standard and a decent default sans other pressures.

    If other methods of auth are available (like local, social, cloud, etc) then you’ll have more decent options. I would define the security objectives for application access, define the user access objectives from the Organization’s perspective, and then plot each solution against those two axes (napkin graphs - nothing serious). Whichever of the top three is the least administratively burdensome is then selected as my first choice for implementation with the other two as alternatives.

    An aside: unless there is only one reasonable choice most folks find one option insufficient, two options difficult to decide between, and four options as having one option too many - whenever possible, if another party’s buy-in is desired, present either three options or three variations on one option. This succeeds even when the differences are superficial, especially when the subject is technical, and 2x if the project lead is ignorant of the particulars. People like participating.

    I’d then propose these options to my team/direct report/client, decide on a path forward together, and plan the rest from there. There’s more to consider (again dependent on org maturity) but this is enough to get the project oriented and off the ground.

    Regarding FOSS alternatives: you’re likely locked in with the vendor’s proprietary software for monitoring the cameras. There are exceptions but most commercial security system companies don’t consider interoperability when designing their service offerings. It might be worth investigating but I’d be surprised if you find any third party solutions for monitoring the vendor’s cameras which doesn’t require either a forklift replacement of hardware, flashing all of the existing hardware, or getting hacky with the gear/software.

    I hope this helps! <3